USF Libraries
USF Digital Collections

Interim evaluation report : task order number 1 : Automatic Vehicle Classification (AVC) systems

MISSING IMAGE

Material Information

Title:
Interim evaluation report : task order number 1 : Automatic Vehicle Classification (AVC) systems
Physical Description:
Book
Language:
English
Creator:
University of South Florida. Center for Urban Transportation Research
Publisher:
Center for Urban Transportation Research (CUTR)
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Traffic surveys--Equipment and supplies   ( lcsh )
Intelligent Vehicle Highway Systems--Florida   ( lcsh )
Automatic Vehicle Classification (AVC) (TRIS)   ( trt )
Genre:
letter   ( marcgt )

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
usfldc doi - C01-00191
usfldc handle - c1.191
System ID:
SFS0032291:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Interim Evaluation Report Task Order #1-Automatic Vehicle Classificaton (AVC) Systems prepared by Center for Urban Transportation Research College of Engineering University of South Florida July 1997 The opinions, fmdings, and recommendations expressed in this report are those of the Center for Urban Transportation Research and the University of South Florida and not necessarily those of the Florida Department of Transportation This report, serving as the final report for Task Order #1, has been prepared in cooperation with the Florida Department of Transportation-Office of Toll Operations, in partial fulfillment of (Eleetronic Toll and Traffic Management Retainer Services) Contract No. B-A90j), State Job. Nos. 97000-1322 and 99900-5000, Journal Trans. 492()265500648900700, and CUTRAccount No. 21-17-248-L.O. Sujeeva (Anu) Weerasuriya, Graduate Research Assistant and Michael C. Pietrzyk, CUTR ITS Program Manager, Principal Investigator

PAGE 2

TABLE OF CONTENTS Page EXECUTIVE SUMMARY ......................................... .. .... ..... 1 1. INTRODUCTION .............. ... .... ........................... 2 2. BACKGROUND ............. ... ................ ....... .... ......... 3 2.1 Technical Papers ..................... ................. ...... ... 4 2.2 Principles of Technologies ...... ................................... 6 2 3 Review of Exis ting Florida Toll Rate Structure and Vehicle Classes ..... . 11 3. FIELD INVESTIGA TIONS .. ... .... .... ..................... ....... 13 3 .1 Current Classification Process . . . . . . .. . . . . . . . ... "13 3.2 Existing Problems at Toll Plazas ..... ................... ......... 15 3.3 Field Tests of AVCEquipment .................... . ..... ........ 18 4. COMPARATIVE EVALUATION OFAVCDEVICES ......... ............. 24 4.1 Sensing and Activating Module (SAM) . ......................... 24 4.2 Sonar XR-500 .............. .................. . ............... 26 4 3 Autosense II .. : ...... ................................ ..... . 27 4 4 M bil . ... . . 29 o tzer ............ . .... . . ................. .......... 4.5 Ground-Hog ... . ....... ........ : .. ... ...... .. ... .. : .. : : 29 4.6 Light Curtains ............. . ." . ............... : ...... ..... 30 4.7 IDRIS .. ... ... .. ........................... .... ......... 35 5. SUMMARY OF FINDINGS ................ ... ... ...................... 36 6. RECOMMENDATIONS ... .................. ... ..... ................ 38.

PAGE 3

BffiLIOGRAPHY .............................. ............. ......... 40 APPENDICES A. List of Contacts ......... ... : .... .......... ..... .......... ... 4 2 B. Manufacturer Specifications for Light Curtains .. .......... ............. 44 C. Man u facturer Specifications for IDRIS ............ ." .... ......... .. 45 LIST OF FIGURES 1. Typical Manual Lane and Automatic Lane Configuration in Florida Turnpike System .... 14 2. Schemati c Diagram of Major Parts in a Scanning Area [Source-STI Brochure ] ........ 31 3. Light Curtai n i n Operation (Light beams are shown for clarity. However, they are not visibl e in the field ) [SourceSTI Brochure] ...... : . ............. ........ ; ...... 32 LIST OF TABLES 1. Comparison of Detector Technologies ................ . . . . ............... 10 2. Vehicle Classification in Florida .......... .............. ... ...... ....... 12 3. Technica l Data of SAM .. .. ............................................... 25 4. Differences among Light Curtain Manufacturers ........................ : ........ 33 5. Light Curtain Clients (Toll Agencies) in the U.S.A. Contacted by CUTR .............. 34

PAGE 4

EXECUTIVE SUlviMARY CUTR' s objective in this task order was to evaluate the capabili ties of various automatic vehicle classification (A VC) systems, primarily those based on pulse laser, ultrasonic, and active infrared technologies. The basis of this evaluatio n was formed from available reports, brochures, and field tests prepared and conducted by other agencies. The existing overhead ultrasonic vehicle separation component of the A VC system has generally become a sufficient maintenance problem as to be "unacceptable" to the Florida Department of Transportation (FDOT) Office of Toll Operations (OTO). A side-mounted version of this device cou l d improve accuracy and ease of maintenance, but insufficient testing has been conducted to confirm this expectation. A new improved version of the overhead has recently been deployed, and a newly-developed algorithm for inductive loops (IDIUS) appears promising. Other vehicle separator technologies are available, but none seem capable of offering the reliability The fact that at least nine ETC agencies have chosen light curtains for vehicle separation begins to indicate the direction the industry is moving In combination with triggering loops and treadles, light curtains . appear to provide the most efftcient technique for automatic vehicle classification currently. However, IDIUS and the improved transducer for ultrasonic overhead should be equally considered at this point. It is recommended that field tests be conducted by the OTO to confum accuracies of the light curtain, IDIUS, and improved overhead ultrasonic vehicle separation alternatives under various traffic and environmental conditions and determine software integration requirements fo r reconciliation of conventional and ETC toll transactions. Therefore, this report is to be considered as a documentation of interim findings Eva l uation of Automatic Vehicle Classification Systems . 1

PAGE 5

I. ll-.'TRODUCTION A total of seven (7) automatic vehicle classification (AVC) technologies were evaluated under this . task including: vehicle magnetic imaging, radar, 'video tracking, video trip line, active infrared, pulse laser, and ultrasonic (piezo-electric is a used to transform electrical energy to mechanical energy in devices). Different technologies offer devices w itb. several capabilities. Evaluating tb.e advantages and disadvantages of different technologies and devices will help the Florida Department of Transportation (FOOT) to make a more informed decision on a suitable AVC device. The primary technology emphasis of this evaluation was on pulse laser, ultrasonic, and active infrared technologies However, toward the e .nd of the eval uation, FOOT re-emphasized tbe focus on light curtains. The task order scope d i d not all ow CUTR to conduct large scale independent field testing of new A VC devices, nor did it allow visits to any field testing of the new devices outside the State of Florida. However, CUTR was able to borrow an ultrasonic device (Sonar XR-500) from FOOT and conduct small scale field tests. Therefore, this report was compiled on information gathered from limited field tests (mostly conducted by other agencies), site visits within the state, and te lep hone conversations with manufactures and toll agencies. The existing A VC system includes loops, treadles, and an overhead u ltrasonic device from Cubic (currently known as TransCore). While the overhead mounted ultrasonic device separates vehicles, the treadle counts the number of axles for each vehicle. Often, the ultrasoni c device fails to accurately separate vehicles creating an audit problem to the State of Florida Office of Toll Operations (OTO). Modifying or changing the existing vehicle separators was necessary for toll audit purposes. Therefore, the main objective of this study is to evaluate the ability of A VC devices to accurately separate vehicles. Evaluation of Automatic Vehicle Classification Systems . 2

PAGE 6

2. BACKGROUND Automatic Vehicle Classification (AVC) systems apply advanced technologies to improve v.ehicle . classification for toll collection systems. Vehicles can lie classified by number of axles, dimensions (e.g. height, length, wheel-base, height over first axle), weight, number of occupa nts in vehicle, and purpose for which a vehicle is being used (e.g a special class assigned to taxi cabs). Toll agencies use these classifications to determ ine the correct toll for each vehicle. The OTO has an axle-based toll structure. The OTO's curren t AVC system (Cubic's overhead < ultrasonic device) frequently fails to separate vehicles due to malfunction of the device's acoustic membrane. As a result, the existing system has been determined to be unacceptable by OTO, particularly in consideration of greater vehicle processing speeds associated with SunPass implementation (the state s electronic toll collection program soon to be deployed). ClTfR's research investigating current AVC systems started with a thorough literature s urvey using several library databases and the Internet. Through various contacts, the CUTR p roject team was able to visit field testing of Autosense (I, II, and III) and SAM. Sonar XR-500, a side-mounted ultrasonic device from Ocean Motion, Inc., was tested at the University o f South Florida campus in Tampa. Only a few objective field tests of automatic vehicle classification devices exist in the U.S.A. It may be because that other components of automated traffic surveillance systems (e.g., automatic vehicle identification, automat i c toll collection, and automatic traffic data collection etc.) receive more attention than A VC systems, and, therefore, more resources are funneled in to those automation activities. Several technical pape r s have been published on new technologies appJ.ied to veh i cle classification and laboratory environment testing of these devices. However, sorne of these tests, conducted in laboratory environments, cannot truly represen t actual traffic and weather conditions. Evaluation of Automatic Vehicle Classification Systems 3

PAGE 7

Background research on existing AVC field tests and lite!'8lllre on A VC started with searches through Internet and libracy databases. As previously mentioned, the search was limited to AVC 0 tests performed in the U.S.A. A total of 6 database sources were used to search field test 0 0 reports/papers The fo llowi!Jg is a list of databases where the search was comple ted. A Yahoo, lnfos e e k, Lycos, M agellan, and ExiteWorl d Wi de W ebbased search methods B. Engineer in g Index ERIC G e n eral AcademicUnivers ity libracy dat a b ases C. Procite CUTR library d atabase D The Intelligent Trans p ortation Infrastructure Web site E. TRIS TRB-based database F Institute of Transportation Engineering Journal 2.1 Technical Papers Six technical papers directly related to AVC systems were found through the literature survey. The papers discussed treadle and loop to acoustic and neural networ k AVC systems. Most of the papers were based on ne wly-designed de v i ces t hat were not currently a vailable in t he market. Reel (1996) described the history of vehicle classification in Florida and its deve l opment to the present time period. His brief discussion was focused on vehicle classification data, gathered at the state's permanent traffic count stations that feeds into th. e FOOT's Roadway Characteristics Inventory (RCI) database. The FDOT reports these classification data in its Annual Vehicle Classification Report, axle correction factors, traffic flow map, and allocation of truck percentage to every segment of the state highway system Cosentino, et al. (199 6 ) discus s e d a development and field evalu at ion of a fib er-optic sensor-base d traffic classificati o n s y stem T h e fibe r-optic sensor d etects vehicle axles an d th e TrafiCOMP II Model241 (from P eek T raffic Inc ) interprets the signals t o classify v ehi c les and to estimate speeds and gaps, etc. Although field tests were conducted with the fiber optic sensor the results were not Evaluation of Automatic Vehicle Classification Systems 4

PAGE 8

compared with ground truth sources such as video taping the actual test and manually verifying the sensor results. This system is more suitable in conventional traffic count stations than in toll plazas. The properties of these fiber-optic sensors reduce the potential of lightning damage to the electronic AVCsystem. Gattis and Lee (1992) described a development and field testing of a photoelectric A VC system. The system employs an inductive loop detector and a pair of photoelectric sensors emitting modulated infrared-spectrum light beams, wh ich are reflected off of raised pavement markers The presence of a vehicle activates the loop detecto r prompting the system to classify the passing vehicle The pattern of the infrared light beam interruptions enables the system to infer the vehicle's speed, number of axles, and number of tires The number of axles and tires defmes the vehicle's class for assessing the proper toll S i gnals from the activation go through a special microprocessor into a personal computer for interpretation, display and recording. The system was tested initially at a high-speed (50 to 60 mph) at an Oklahoma Turnpike Authority site during 1990 and 1991 An accuracy of 95 percent was reported during the field tests. James and Sampan (1995) represented a method of vehicle classification using neural networks based upon acoustic signals. The acoustic signals, detected by tbe AT&T SmartSonic sensors on Highway Route 460 near the Virginia Tech Campus in southwes t Virginia, were used as inputs to the network. Vehicle classifications were performed in real time. The network's success rate in separating vehicles into cars and trucks was 96.5 percent and the success of fou(-c)ass classification was 95.5 percent To make more informed purChasing decisions, many agencies test vehicle detectors, though it can be an expensive and time-consuming task. The technologies and the changes in succeeding versions of each application place a significant burden on the ability of agencies to do their own testing in a cost effective manner. Further, lack of standard test protocols for vehicle detectors makes it difficult to avoid duplicate testing (Gillmann, 1996). Hamrick and Schoenmackers (1996) discussed Evaluation of Automatic Vehicle Classification Systems 5

PAGE 9

a plan for a National Vehicle Detector Test Center (NVDTC). The NVDTC is envisioned to conduct bench tests (to be completed in the laboratory shop), simulation tests (to be completed in the shop and laboratory), and field tests (to be completed at a field site where various types of sensors are Gillmann (1996) stated, NVDTC is far from being a reality jn the near future. 2.2 Principles of Techno logies Many transportation agencies have been using inductive loops and road tubes to determine the vehicle class. However, difficulties in obtaining accurate vehicle classification data on multilane highways, assessing the validity of the data, and maintaining inductive l oops have encouraged the transportation agencies to find new alternatives to inductive loops and roadtubes. Several new technologies have been developed for automating the traffic data collection. The basic principles of each technology are summarized as follows'. A. Infrared (Passive and Active) Infrared waves (heat waves) have wavelengths ranging from about lmm to 7 x 101 m. These waves, produced by hot objects, are readily absorbed by most materials and appear as heat. One advantage of infrared rays is tha t they can penetrate fog and haze better than visible l ight. Active infrared devices aim a lo w energy infrared beam at the target area and analyze the reflected infrared signal for the presence of a vehicle. Passive infrared devices detect vehicles by measuring the infrared radiation that naturally rad iates from the heat of the detection zone and the change in heat caused by the presence of a vehicle. 1 Field Test of Monitoring of Urban Vehicle Operations Using Non-Intrusive Technologies, Minnesota Department Of Transportat ion, Final Repo rt, 1997 Evaluation of Automatic Vehicle Classification Systems 6

PAGE 10

B. Magnetic (Passive and Active) Passive magnetic devices measure the change io the earth's magnetic flux created when a vehicle passes through a detection zone Aetive magnetic dey ices, such as inductive loops, apply a small electric current to a coil of wires ;w.d detect the change in inductance caused by the passage of a vehicle. C. Microwave (Doppler, Radar, and Passive Millimeter) Microwaves are short-wavelength (about lmm to 30cm) radio waves. Doppler microwave devices transmit lo w energy microwave r adiation at a target area on the pavement and then analyze the I signal reflected back to the detector. According to the Doppler principle, the motion of a vehicle in the detectio n zone causes a shift in the frequency of the reflected signal which can be used to
PAGE 11

' E. Ultrasonic (Pulse and Doppler) Ultrasoni c waves are sound waves whose frequencies are from 20kHz to I 00 kHz, which is beyond the audible range. Because of their high frequency and corresponding short wave lengths, ultrasonic . waves can be used to produce images of sma l fobjects. When an alternating voltage of very high frequency is applied to a crystal (such as quartz or strontium titanate), the crystal will vibrate at the same frequency as the applied voltage As the crystal vibrates, i t emit s a beam of ultrasonic waves Transforming electrical energy into mechanical energy in this manner is called "piezoelectric method." The energy transformation is reversible in that, if some external source causes the crystal to vibrate, an alternating voltage is produced across the crystal. Therefore, a single crystal can be I used to both transmit and receive ultrasonic waves. Pulse devices emit pulses of ultrasonic sound energy and measure the time for the s i gnal to return to the device. Doppler devices emit a continuous ultrasonic signa l and utilize the Doppler principle to measure the shift in the reflected signal. F. Vid e o Video devices use a microprocessor to ana l yze the video image input from a video camera Two basic approaches, tripline and tracking, are used to detect traffic. Tripline techniques monitor specific zones on the video image to de t ect the presence of a vehicle Video tracking techniques employ algorithms to identify and track vehicles as they pass through tbe field of view Some systems use a combination of th e two methods. G Laser Laser (light ampl ification by stimulated emission of radiation) is a light source which produces an int ense unidirectional beam of coherent light. Lasers are avail able that cover wavelengths in infrared, visible, and ultraviolet regions. Evaluation of Automatic Vehicle Classification Systems 8

PAGE 12

Scanning laser beams are used to detect the presence of vehicles. Vehicles operating in high speed, high volume conditions can be detected and classified. Output from tbe device is processed to produce 3-dimeosional images which are compared with stored templates of various vehicle profiles . to determine vehicle classification. Evaluation of Automatic Vehicle Classification Systems 9

PAGE 13

Table I. Compariso n o f Detector TechnologiesZ Technology Advantages Disadvantages Ultraso ni c compact size, ease of attachment t o existing structures performance may b e degraded by variations in temperature and a i r turbulence Passive infrare d gr e ater viewing distance in fog than with visible performance potentially degraded by heavy rain o r snow wavelength sensors Microwave Doppler good performance in incl e ment weather can not detect stopped or very slow-moving vehicles direct measurement of speed requires a narrow-beam antenna to confine footprint to single lane in forward-looking mode Active infrar e d greater viewing distance in fog than with visible performance potentially degraded by obscurants in the wavelength sensors atmosphere and weather direct measurement o f speed I Magnetometer can detElC( small vehicles including bicycles difficulty in discriminating longitudinal separation between useful where loops can not be ins tall e d closely spaced vehicles Acoust ic potential for identifying specific vehicle types by their signal proceasing of energy received by the array is required to acoustic signatu r e remove extraneous background sounds and to identify vehicles Video image roadside processing allows low data rate transmission different algorithms usually required for day and night use processor multiple lanes observed possible e rr ors in traffic data during trans i tion from day to no traffic interruption for installation and repair nigh t susceptible to atmospheric obscurants and adverse weather zsource-RTMS Performance and Qualification Test History, EIS Electronic Integrated Systems, Inc Brochure Evaluation of Aut omatic Vehicle Classification Systems 10

PAGE 14

2.3 Review of Existing Florida Toll Rate Structure and Vehicle Classes As a possible solution to the problem with individual vehicle separation, the toll rate structure was examined to determine if a simplification could be made to remedy the to ll audit problem The toll rate of the Florida Turnpike system is based on the number of axles. OTO' s current AVC system has an accura te count of the total number of axles, but does not have the same accuracy in separating vehicles. CUTR investigated the possibility of counting only the total number of axles for auditing collected tolls. Consideration was given to OTO's current toll structure, and suggesting any suitable modifications The State of Florida utilizes two different toll collection structures; ticket systems and coin systems. The following two formulae are being used to estimate the toll for differ ent vehicle types. In ticket system: In coin system: Toll = toll for a 2 -axle vehicle 2 (no. of axles) Toll = (no of axles 1) (toll for a 2-axle vehicle) The current vehicle classification used in Florida is shown in the following table. Evaluation of Automatic Vehicle Classification Systems . 11

PAGE 15

Table 2 V chicle Classific a tion in Florida Class Indicated Axles Vehicle Type 1 N/A unassigned 2 2 2-axle vehicle 3 3 3-axle vehicle 4 4 4-axle vehicle 5 5 5-axle vehicle 6 6 6 axle vehicle 7 7 7-axle vehicle 8 8 8-axle vehicl e 9 :.9 # of axl e s in vehicle 9 If the toll rate structure is such that every axle is charged at the same rate, the t o tal n umber of axles . can be used to audit the toll collecto r during a given shift. When the t otal number o f axles passed in a lane is multiplied by the per axl e rate, i t will give tb.e total transaction dollars for that particular shift. To examine the current toll structure, l et a 2-axle vehicle pays a toll of $x at a toll plaza. Consider a vehicle with "n" (>2) nu m ber of axles go through the same toll plaza. If the toll per axle (i .e., $x/2) is the same for every vehicle, an axle-based audit system may work. At a t i cket sys tembased toll plaza, toll per axle is, -n 2._ = .=. n 2 This means that the toll rate structure in this system has the same per axle rate for every vehicle. Therefore, an axle-based audit system may work at ticket system-based toll plazas Evaluatio n of Automatic Vehicle Classificati on Systems 12

PAGE 16

At a coin system based toll plaza, toll per axle is, (n-l)x ,. x n 2 The toll rate structure in this system charges a different per axle rate depending on the numbe r of axles in the vehicle. Therefore, an axlebased aud it system may not work at coin system based toLl plazas with the current toll rate structure. If the toll rate formula for ticket system is used at coin system based toll plazas, an axle-based audit system may be successfully imp l emented In other words, vehicle separating devices are not needed for ticket-based system, because an accurate axle 1 count device can be use d to audit tolls. However, with the deploy ment of electronic toll co llection systems, vehicle separation for individu al vehicle assessed tolls is mandatory Therefore the previous discussion on auditability by total axle counts becomes moot. Capabilities of the vehicle separation will dictate the extent (number) of separate vehicle classifications 3 FIELD INVESTIGATIONS 3 I Current Classification P rocess The existing A VC system installed in manua l toll l anes in the Florida turnpike system consists of an arming loop, treadle, and overhead AVC device. A schematic diagram of a typical manual toll l ane and ali adjacen t automatic lane is shown in Figilre I As a vehicle a pproa ches the toll collector s cabin, the arming loo p near the cabin detects presence of the vehicle. The toll collector detennines the number of axles in the vehicle a n d manually e nte rs it into the system. The system then calculates the toll to be collected After completing the transaction, the vehicle is class ified at the end of the toll lane. First, the overhead A VC device Evaluation of Automatic Vehicle Classifica tion Systems 13

PAGE 17

detects presence of the vehicle underneath it and sends a signal to the treadle doctor board through an RS-232 cable. As the vehicle pas ses over the treadle, it sendS signals to the tread l e doctor. The treadle doctor then records the number of axles. After the vehicle leaves-the treadle, the AVC device detects the vehicle's absence and sends a signal to the tread l e doctor These signa ls are fed into a lane control software which registers the number of axles of the vehicle and estimates the toll r equired. This toll is registered in a computer which provides a total s u m of transactions at the end of each toll collector's shift. At the end of each shift, the plaza supervisor will check the toll collector's revenue with the AVC s ystem's report for audit i ng purpos e s Ck'othed AVC Autom gat (TYP a 1ie .) r< Traffie Hgh t (TYP. ) b AnnU\gLoop Treadle 6' Loop< I = I I '\ Automatic coin machine (TYP. ) Manua l Lane Automatic Lane :> ... How F low Figur!' 1. Typical Manual Lane and Automatic Lane in Florida T u rnpike System Evalua t ion of Automatic Vehicle Classification Systems 14

PAGE 18

3.2 Existing Problems at Toll Plazas If the AVC device fails to separate individual vehicles or if the treadle fails to count axles, the lane control software will not classify vehicles. Whet! the AVC system will not classify a vehicle, the syst em will not record a transaction In other words, the sys.tem will not expect the collector to collect a toll during the transaction. Therefore, when the tolls collected by collectors are audited with the total recorded transactions from the AVC system, the AVC transaction will report fewer total dollars than collected For example audit differences from $2 to $200 we r e observed during a visit to the Anderson mainline toll plaza at Veteran's Expressway in Tampa. Such audit unbalances exist at many OTO plazas. Through field visits and phone conversations with several l Florida Turnpike system toll plaza managers and technicians, the CUTR project team gathered the current experiences with the existing AVC system. Their experience and comments are summarized as follows. 3.1.1 Veteran's Expressway The CUTR project tearn visited the Anderson mainline toll plaza at Veteran's Expressway on February 20, 1997. The purpose of the visit was to gather first hand information regarding the operational and maintenance problems associated with the Cubic A VC device. During the visit, the technicians showed how the AVC system works. The team members witnessed vehicles being classified by the toll collectors and the AVC system in real-time According to the plaza. technicians, transmitter cones of the Cubic device n .eeded to be replaced frequently (varied from 90 days to one year). The replacement frequency was based more on the available supply of new transmitter cones than anything else (i.e., replacement would be more frequent if adequate supply was available). The height (above the road surface) of the AVC device has been reduced from 18' to 15' to improve the accuracy. Performance of the AVC system has not been affected by fog or rain. However, the A VC device signal has been noted to cbaQge frequently when the temperature drops below 32"F. The technicians had experienced insect nesting problems blocking a membrane in the tranSmitter cone. Especially, when spiders make webs and lay eggs in the Evaluation of Automatic Vehicle Classification Systems IS

PAGE 19

transmitter cone, the AVC system does not function properly. The plaza supervisor showed an example of an audit difference of $200 at the end of a toll collector's shift. 3.1.2 Toll Maintenance Office, Orlando The regional maintenance supervisor and a technician of the OTO were contacted on March 12, 1997. Among the Florida Turnpike system toll plazas contacted bY. CUTR, the Orlando toll plazas have the least number of problems directly associated with the AVC system. They have been able to successfully identify the leading cause for many problems that l e d the AVC system to m alfunction and they have been able to take necessary steps to correct the problems. The existing AVC devices were also lowered from the plaza canopy height to 16.5' above the road surface. Toll plazas close to swamp areas, have experienced problems with spiders (such as spiders making webs and laying eggs) that led the AVC to malfunction. Some problems associated with A VC system in the Orlando area had not been d u e to the Cubic AVC device itself. They were attributed to human errors and treadle doctor errors For example at one plaza, a lane had a steep gradient and the AVC device mounted under the canopy was not level with the road surface Leveling the A VC device with the road surface corrected this problem. In other cases, toll collectors were found to classify vehicles before they entered the lane or after they left the lan e Some collectors were found not closing a transaction before a vehicle left the lane. These mishaps have caused the AVC system not to record vehicles and even to misclassify vehicles. Mter instructing the toll collectors, many human errors which had led the AVC system to malfunction have been reduced to a minimum Occasionally, they have found that some treadles were not properly fastened to the pavement. This led the A VC system t o misclassify vehicles. The variance between the dollar amount collected and the total transactions reported in the AVC sys tem per toll collector's shift has been between $0 .10 and $40.00. Evaluation of Automatic :Vehicle Classification Systems 16

PAGE 20

3.1.3 Okeechobee Toll Plaza The plaza technician was contacted on March 13, 1997 and gathered information on problems associated with the current AVC system. The te chnician mentioned that the transmitter cone of the . current AVC device needed frequent replacements (2-3 day cycles) Recently the Cubic AVC device was lowered to a level of 15' above road surface as well. The difference between the dollar amount collected and the amount expected by the A VC system of a toll collector's shift has risen to $1,000. The technician also men tioned that the AVC system fails to function properly during heavy traffic. 3.1.4 Cypress Creek Plaza The plaza manager was contacted on March 1 3, 1997 to gather information on problems associated with the current A VC system in Cypress Creek Plaza. Similar problems regarding frequent replacements (1-2 days) of transducer cones were received. The Cypress Creek Plaza has also been faced with the same toll audit problems. However, this plaza has quite an unusual audit problem with some shifts where the AVC system reported a negative total transaction Through field visits and telepho!!e conversations, the following observatio ns were made as common occurrences at OTO' s toll plazas: when the AVC device operates properly, the system works well when the AVC device does not operate properly, toll audit differences occur at various levels insects damage the transducer cone when l owered to a height 15' above the road surface the A VC device accuracy improved the transducer cone needs frequent (ranged from one day to one year) replace ments replacing the cones is difficult and sometimes takes several days to gather required tools (e.g., every toll plaza does not have a bucket truck) the transducer supplies are hard to get and are costly ($300 each) Evaluation of Automatic Vehicle Classification Systems 17

PAGE 21

CUTR investigated the principles behind the transducer cone and the insect's reaction to the ultrasonic waves. A physics professor pointed out that the nylon stocking material, used to cover the cone opening, may block much of the ultrasonic waves. His suggestion was to find another . alternative to the nylon stocking. cover for the transducer. An entomology pmfessor stated that an insect predator, which is found in Florida, sends some form of ultrasonic sound waves to caich its pray. T herefore, whenever insects h= an ultrasonic sound, they tend to move away to avoid being captured by the predator. With these explanations, it was difficult to find a reas o n for insects being attracted t o the tran sducer cone. N ew transducers ($18 each), manufactured for outdoor use (previous membranes were believed to be for indoor use) by Polamid, have been recently installed (second week in June) at tile Cypress Creek Plaza. CUTR contacted a technician and the plaza manager on July 16, 97 to gatller information about tllis new installation. The new cones, installed with a protective nylon stocking cover, are functioning successfully. According to the plaza manager, the re ported audit differences have been significantly reduced or eliminated since the new installation. The new cones have been in operation for about four weeks, and insect related problems have not been reported during this period Monitoring of this improvement should continue. It is also understood that Polaroid is recommending further enhancements to the transducers requiring re-engine ering 3.3 Field Tests of AVC Equipmen t To determine whether the new devices could perform the data collection functions under a variety o f difficult geometric, environmental, and traffic conditions, many agencies perform field tests. However, the current task order scope did not allow CUTR to perform rigoro us field tes ts with the new A VC devices. Therefore, CUTR depended on findings from field tests conduct e d by other agencies in preparing this report. CUTR project team cond ucted some very limited field tests with an AVC device (Sonar XR.-500) manufactured by Ocean Motions. The project team also made field visits to Schwartz Electro-Optics, Inc. in Orlando, FL and to toll plazas on the Venetian and Rickenbacker Causeways in Miami, FL where new A VC devices were being tested. Eval uation of Automatic Vehicle Classification Systems 18

PAGE 22

3 .2.1 Sonar XR.-500 The Sonar XR.-500 is an ultrasonic device that outputs one-dimensional swfacc profiles of vehicles in a traffic stream This device is a modified version of OTO's current Cubic AVC device. The CUTR project team conducted four on-line field tests with the Sonar XR.-500 during February 1997. These tests were conducted to evaluate the accuracy of the device as a vehicle separator CUI'R received clearance from the Division of Environmental Health and Safety and the University Police at the U niversity South Florida (USF) to conduct tests with the device at the Tampa campus. Equipment used in the fiel.d tests included: Sonar XR.-500 with a piezoelectric transducer, reflector I swface (cardboard fixed to a metal frame with a supporting arm), laptop comput er p o rtable power .generator extension cords (:Z-nos), video camera with a stand (tripod), 8 mm video cassettes. The test sites are located within the USF campu s in Tampa. The first tes t was conducted from 9:45 A.M. to 10:4 5 A.M at the westbound, right tum only lane at Laurel Drive & Alumni Drive, a 3 legged, unsignalized intersection. The transducer was attached to a timber pole which was stuck in the ground and was about 6" away from the curb edge of the right-turn only lane. A cardboard plate of 23" x 30" was used as the reflector surface. The reflec t or surface was placed on a raised island and was kept in place using several bricks. A video camera was use d to record the monitored traffic on the right -tu rn only lane This recording was later used as the ground truth data. The l aptop computer was used to record the profile reading in real-time The second and the third tests were conducted at the southbound right-turn only lane a t the same intersection from 12 P.M : to 12:30 P .M. and 4:30P.M. to 4:45P .M. respect ively. The equipment set up was s im ilar to the first test During these two tests, the sonar reading fluctuated between a maximum distance setting ( maximum distance where an object is expected) and a closer distance despite the presence or absence of a vehic le Therefore, the tests were terminated ear ly. Evaluation of Automatic Vehicle Classification Syste ms 19

PAGE 23

The fourth test site was at the USF Visitor Reception Center Information lanes which have a similar layout to a branch bank's drive-through teller lanes. The reception center has four lanes with each having access to an informat ion window. Vehicles approaching .the reception center stop at the informat ion windows and then leave the information center. Since the lane closest to the main information window attracted more traffic, the traffic on that lane was selected for the field test The equipment setup (except the reflector surface) was similar to that of the previous tests. A 2.5' wide brick column was used as the refle<.10r surface during this field test. The test was conducted between 9:15A.M. and 10:45 A.M. After about one hour into the test, a fluctuating sonar reading was observed. The device was then turned off and on again. This quick fix worked for a few (10-15) minutes, 'after while similar fluctuations were observed. Accuracy of the profile data was checked against the ground truth data fro m the video recordings. It was found that when the Sonar XR-500 was operating properly, the accuracy was 100 percent. However, the device failed 40 percent of the time during the field tests. The toll plazas that CUTR contacted had similar experiences with the existing Cubic A VC device. Ocean Motion had used the same device to conduct a field demonstration at the Dolphin Center Toll Plaza in Miami on November 14, 1996. A single vehicle was used during this demonstration. The tniiler hitches were simulated by a plastic soda bottle hooked to the trailer hitch connector. The transducer was mounted horizontally and perpendicular to the traffic flow. The results suggested the following: The Sonar XR.-500 detected fixed and moving targets across the full width of the test lane. Highly curved contours of the test car returned echoes from its sloped windows, roof edge, front and rear bumpers, and its streamlined driver-side mirror. The membrane style transducer was able to detect a simulated trailer bitch. Evaluation of Automatic Vehicle Classification Systems 20

PAGE 24

3 .2 2 Autosense II A member of the CUTR projec t team visited Schwartz Electro-Optics, Inc. (SEO) in Orlando FL on March 7, 1997. SEOwas conductingon linetestswith Autosense I, II, and IU on SR 441 dur ing . . the week of March 3, 1997. Autosense devices are mainly o v erhead-mounted pulse laser devices . The Autosense I can measure one-dimensional vehicle profiles. The Autosense II device is capable of measuring the height, width, length, and speed of vehicles. I t can classify up to I 0 vehicle classes. The latest version, Autosense Ill is thought to be capable of measuring vehicles across multip le lanes. Because vehic l e classification and traffic monitoring capabilities of Autosense II are well suited with the current needs of OTO, the main emphasis of the visit was on Au t osens e II . A rotating drum and a pair of mirrors are utilized in Autosensc II to emit p ulse laser beams at 45 kHz frequency. Along with the moving parts, the mirrors and circuits are containe d in an environmentally sealed unit. During the on lin e field tests, the Autosense devices were mounted on a mast arm over a southbound lane. The mast arm, a t a 25' height from the road surface, was cantilevered to a nearby light pole The CUTR project member observed the on-line tests with the Autosense devices wh ere outputs were p ro jected to a monitor. Autosense II captured all th e vehicles driving through the t est lane during the observation During the field visit, SEO design engine e r s mentioned that Autosense II has an accuracy o f 98% in vehicle c lassification. The ob serv e d tests were conducted during a steady speed, mixed traffic (including passenger cars, motorcycles, vans, and trucks) flow period. However, stop-and-go traffic conditions exist in toll plazas From the existing Autosense II design stand point, the device may run into difficulties i n moni toring vehicles traveling below 5 mph speed. SEO design engineers confirmed t hat the device can be modified for accur a te ve hicle classification during such conditions. Evaluation of Automa t i c Vehicle Classification Systems . 2 1

PAGE 25

3 .2.3 Sensing and Activating Module (SAM:) Lockheed Martin, Inc., is conducting initial field tests on an electronic toll collect ion (ETC) system being installed in toll plazas at Venetian and Rickenbacker Causeways, in Miami, FL. The system consists of light curtains, loop detectors, SAMs; Mark IV ETC antennas, treadles, and enforcement video cameras. The SAMs were installed to detect presence of vehicles to assis t Mark IV antennas that read in-vehicle transponders Two SAM units, monitoring traffic from the side, were installed at each lane just before the toll attendant's cabin. The lower SAM unit is about 1.5' above the island level and the other unit is about 1.5' directly above it. This is the frrst installation of SAM in the U.S., and the fJ.rSt side-mounted installation known to exist in the wodd, according to an MBB (manufacturer of SAM) technician on the site. Two members of the project team fust visited the toll plazas on April I 0 and I 1, 1 997. The tests conducted during these two days were focused on the ability of Mark IV antennas to read tags on windshields or dashboards and deduct the correct toll from the corresponding account. The CUTR team observed SAM units monitoring vehicles in a test lane. A m em ber of the CUTR project team later visited the Rickenbacker Causeway toll plaza on May I to observe field tests. A few days before the fie l d tests, CUTR was informed tha t the SAM units would also be tested on May I. Having arrived at the site, the membe r found that the testing was postponed without any prior notice to CUTR. The same field tests were re--scheduled twice on June 5-6 and on June 25-27. However, CUTR later learned that these tests were also postponed due to difficulties associated with system software. The latest test schedule for tests is July 16-18, 1997. 3.2.4 Georgia Field Tests' Several veh i c l e classifiers were tested to determine the accuracy in classifying vehicles, in measu ring vehicle length, and in measuring axle spacing. A total of 13 sensors and classifier configurations were included in this test conducted by the Georgi!l Tech R e search Institute. The tests 3 Accuracy of Traffic Monitoring Equipment, Georgia Tech Research June 1995 Evaluation of Automatic Vehicle Classification Systems 22

PAGE 26

were conducted during May 5-7, 1993 alid Septeo:iber 9-16 1993. All the classifiers tested used a combina tion of magnetic loop detectors and piezoel ectric axle sensors. A sidemounted video crunera provid .ed ground ttutb data for the The classifi cation accur acies . . from tbe tests ranged from 63.5% to 79: I%. A significant part of the low accuracies were attcibuted to the fact that the classifiers bad difficulty in distinguishing small pickup trucks (class 3} from large cars (class 2) based on length and axle spacing. When the classes 2 and 3 were combined, the accuracies were reported to range from 78.8% to 96.2%. 3.2.5 Minnesota Field Tests Field tests on traffic data collection devices were conducted during February 1996 to January 1997. A total of eight technologies (active infrared, passive infrared, passive magneti c, radar dopp ler microwave, pulse ultrasonic, passive acoustic, and video) were evaluated in these tests by Minnesota Guidestar. Traffic data such as traffic volume and speed were collected from 21 devices during these field tests. The field tests were conducted at an intersection and a freeway section in Minneapolis, MN. Some of these devices (e g., Autosense I, Mobilizer, and Video Trak 900) also provided vehicle classification data_ However, the level of vehicle classification varied from device to device. Therefore, the vehicle classification accuracies were neither reported nor analyzed in their fmal report. The following are the relevant results from tb e Mi. nnesota study': Active infrared technology exhibi ted good potential for ve hicle detection during freeway tests. Doppler microwave technologies exhibited good potential for de tecting moving vehicles during freeway tests. Data collection performance at the intersection site was poor 'Field Test of Monitoring of Urban Vehicle Operations Using Non-Int71lsive Technologies, Minn esota Department of Transportati on, Minnesota Guidestar, and SRF Consulting Group, Inc. Evaluat ion of Automatic Vehicle Classification Systems 23

PAGE 27

> Pulse ultrasonic technologies hlive moderate potential for detecting traffic at both intersection' and freeway applications. Video devices require extensive installation and setup time and are not as accurate as other . technologies. However, video devices have the advantage of offering a variety of traffic data and surveillance information. Video devices require extensive calibratioD, over serial communication lines. Lighting conditions were observed to affect some video devices, particularly during the transition from day to night. 4. COMPARATIVE EVALUATION OF AVC DEVICES During this evaluation, the primary emphasis as mentioned previously was on devices that use pulse laser, active infrared, and ultrasonic technologies. The comparative evaluatio n is based on accuracy, configuration, ease of installation, cost, durability and maintenance, proven capabilities, mean time between failures, life cycle of equipment, sensitivity to environmental conditions, and health safety criteria. However, information on all these factors was not available or could not be provided for many devices. 4 .I Sensing and Activating Module (SAM) J:\.1BB Business Development GmbH in Germany manufactures, and Spectra Systems, Inc. in Boca Raton, FL distributes, SAM units in the U.S.A. The SAM unit, which can be mounted overhead or sidefire (horizontally), can detect moving and stationary veh i cles; cyclists, and pedestrians. It operates as an active pulse laser device that often uses a reference background for better accuracy, because vehicles have a certain percentage (typically 20%) of surface that absorbs or reflects a laser beam away from the receiver. If a road surface is selected as the background (when the unit is overhead mounted), the condition of the background could change due to rain, snow, and dust etc. The SAM unit is capable of automatically calibrating background and re-calibrating background Evaluation of Automatic Vehicle Classification Systems 24

PAGE 28

(whe n weather conditions change ) and noise level. The unit has a special and unique analysis technology that enables SAM to offer 12 aetection areas {windows). These windows are adjustable and are individually by a built.in micropro=sor Table3 Technical Data of SAM Laser Pulsed laser, 850 nm (near infrared range), Class 1 Beam Shape SAM Narrow, focused beam SAM-E I widened beam that creates a continuous line Detection Range SAM 0-30 m (0-98') SAM-E 0-10 m (0-32? Power Supply 21-27 V DC Power ConSillllption <2.5W Dimensions 250x 130x 100mm(9 .8''x5.l''x3.9") Weight 1.8 Kg (41b) Environmenta l Protection lP 67 water proof Temperature Range -20 to 60 C (-4 to 158 F) Relative Humidity O%to 100% Operational Weather Conditions Operational under all weather conditions Interface Serial RS 485, half duplex, bus capability Parallel differenlial presence signal no vehicle =+3V vehicle present--3V differential pulse (count) signal Detection Time <20 msec Safety Eye safe (class I laser) Accuracy . Vehicle presence> 99.9% E valuation of Automatic Vehicle Classification Systems 25

PAGE 29

SAMs are used in European countries to gather traffic iofonnation, and the units have been mounted over the traffic lanes in evecy working installation. Th e biggest cballenge for SAM is to detect trailer hitches when tbe unit is mpunted on a sidefiCe. Small trailer hitches are typically ab0ut 1' to 1 .5' above the road swface and are 2 "-3" in diameter : For SAM units to detect these trailer hitches, the operating manual suggests that it should be installed at a height of 3 '-4' with a 5 angle facing d o wnward Being a Class I laser product, SAM is considered safe for eyes. Since the SAM units do not have any moving mechanical parts, they are lik ely to have fewer maintenance problems l The SAM units were introduced in the U.S.A. vecy recently At the time of this report preparation, there was no working SAM installation reported in this country. Even though Lockheed Martin MIS was testing SAM units in New Orleans, LA ; the CUTR project team was not a ble to observe proven capabilities of SAM units at the installati on in Miami, FL As p reviously stated, the tests that were scheduled on April !'0 have been delayed until July 16, i n dicating that Lockheed Martin MIS is nt lcnst 3 -month.s behind the test schedul e. They arc working to resolve the system software problems in the Miami installation and to restart the online field tests on July 16, 1997. 4.2 Sonar XR-500 The Sonar XR-500, manufactured by Ocean Motions, is an ultrasonic devi ce that can be mounted overhead or from the side to the traffic lane This device consists of a small microprocessor unit which is based on a channel ultrasonic transceiver and electrostatic/ pi ezoelectric transducer The microprocessor unit can yield output in both analog and digital formal A fixed reference background with no obstructions between the transducer and the surface is essential to improve the vehicle detection accuracy. The Sonar XR-500 measures distance to the closest surface using ultrasonic sound waves. It can measure profiles up to 90' away from the transducer. Evaluation of Automatic Vehicle Classification Systems 26

PAGE 30

The transducer emits sound waves through a conical fashion. As the angle of the cone decreases, the accuracy of the output increases. Narrow eonicai beams are best suited for trailer hitch de te ction Trailer hitchers are set at different heights in cars, pickup trucks, large trucks. Ther efore, when the transducer is horizontally mounted to detect vehicles, several transducers may be required to detect every trailer hitch. However, a horizontally mounte d cone is easy to replace in the event of equipment failure As experienced during the limited field tests at USF, the Sonar XR-500 would be easy to install However, despite the absence or presence of a vehicle, the output could fluctuate between an observed distance to the vehicle and a set defaul t distance that sends faulty vehicle separation signals to an AVC system. 4.3 Autosense II Autosense II, manufactured by SEO, is a pulse laser device It employs two scanning l aser rangefinde r beams, at a 10 angle, to measure a three-dimensional vehicle profile.used for vehicle classification. Eq uipped with a diode-laser range finder, Autosense II has a rotating polygon to scan across a 12' wide lane The polygon rotates continuously in one direction at a constant speed The narrowly spaced laser beams permit the detect ion of two-inch wide tow bars and closely spaced vehicles moving at high speed. A laser diode in the Autosense II unit produces a pealr current pulse. A trigger pulse fro m the scanner control circuit triggers the laser at the p r oper scan angles. When the laser beam m eets a vehicle/road surface, part of the beam gets reflected back to the scanning device The optical c ir cuitry in the unit converts this optical radiation reflected from the vehicle/road surface to fiiSt, an equi valent electrical analog of the input radiation and fmally to a logic-level signal. This logiclevel signal is processed within the range counter logic to y i eld analog range data, which is read by the microprocessor. The anal o g rang e data is sent through a range-measurement circuit, known as Evaluation of Automatic Vehicle Classification Systems 27

PAGE 31

> a time-to amplitude converter ('TAC). TAC ou t put is then converted to digital by a faSt !Obit analog/digital (AID) converter Pulsed tinie measurements are read and converted into distance measurements by a micropfocessor. When vehicles are not present, the range measurements are equal to the distance to the road. When a vehicle is present beneath the sensor the distance to the top surface of the vehicle is measured and provides a transverse height profile of the vehicle on each scan. The vehicle speed, estimated from the time interval between the interception of the two laser beams by the vehicle, is used to space the transverse profile by a straightfotward geome t ric transformation and to obtain the full three- dimensiofial vehicle profile. The Autosense II can detect the following 10 vehicle classes: 1. Motorcycle 2. Motorcycle with trailer 3. Passenger car 4. Passenger car with trailer 5. Pickup van, or sport utility 6. Pickup, van, or sport utility with trailer 7 Miscellaneous truck or bus 8. Miscellaneous truck or bus with trailer 9. Tractor with one trailer I 0. Tractor with two trai l ers The SEO claims 98.5% accuracy of the uriit in vehicle separation The Autosense II instrumentation is housedjn a n environm entally sealed container purged with nitrogen gas. The manufactures claim that the device needs very little maintenance (e.g keeping the glass window clean to avoid range errors). They also reported a 15,000-hour (625 days) mean time failure for the unit. Evaluation of Automatic Vehicle Classification Systems 28

PAGE 32

Texas A&M University has included the Autosense II in their research on trucks and weigh-in motion (report is under preparation). The tests were performed adjacen t to a bridge in Austin, TX. CUTR learned that the unit performed satisfactorily in separating vehicles but performed poorly in measuring truck speed Laser radiation of the unit belongs to Class I laser. The laser radiation of Class I is considered not hazardous and eye safe. The Autosense II installation is easy aod the cost of units is high (about $6 ,000 each in bulk quantities). Performance of the current unit under stop-and-go traffic conditions has not been tested so far. Therefore, the applicability of this u nit for manual toll lanes or mixed-use ETC lanes i s still unknown. 4.4 Mobilizer Mobilizer is a video tracking system used at intersections and freeway and priority corridors. The Mobilizer is manufactured by Condition Monitoring Systems of America, Inc in Long Beach, CA. It can measure traffic volume, speed, queue length, vehicle classification, turning movements, and can detect pedestrians and bicycles. However, the vehicle classification capability of the mobilizer is limited to two classes (1-trucks & buses and 2-cars). It contains tbree major components; Smart Sensor Interface (SSI) that extracts vehicle data from the video camera and transmits it at a low bandwidth to the Mobilizer Advanced Tracking System (MATS) that reduces SSI data into traffic flow information at the Traffic Management Center by tracking vehicles through an entire camera scene, and a Roadside Equipment Interface (REIF) that transforms SSI output into a format used by existing signal controllers for local signal control using a Mobilizer system Since the Mobilizer is a video-based system, i t does not create any safety concerns. 4.5 Ground-Hog The Ground-Hog Model G-2 is an in pavement, self contained (servi ceable or replaceab le) traffic monitor coupled to a high performance, 2 -wa y radio link for wireless transfer of collected traffic data. Traffic detection and measurement are accomplished by vehicle magnetic imaging technology that measures the magnetic influence of.a motor vehicles passing through the earth's magnetic field. Ev aluation of Automatic Vehicle Classification Systems 29

PAGE 33

This unit can measure traffic volume, speed, vehicle class, occupancy, swface and sub-surface temperature of the road. The number of identifiable vehicle classes i s limited to six. Ground-Hog is another product from Nu-Metric. However, the accuracy of this model in vehicle separation has not been documented or reported at the present time. Seminole and Sarasota Counties were planning to install Ground-Hog units at several intersections on a trai.l basis However, this installation has not been completed at the time this report was prepared. 4.6 Light Curtains Light curtain, also known as the vehicle scanner, is an active infrared device It consists of a transmitter a receiver, a controller, two tower enclosures, and interconnecting cables. The transmitter houses a linear array of LED' s (light emitting diodes) while the receiver houses a linear array of photo detectors. A transmitter and a receiver form a scanner. Th e controller operates the scanner and controls the input and outputs to the external sys t em Tower enc losures, also known as environmental enclosures, are used to protect the transmitter and receiver The towers are made of s tainless steel or aluminum. Each LED and its corresponding photo detector forms a scanni n g beam. Only on e LED-Detector pair is acti ve at any given time, to prevent problems caused by interaction between beams. In normal operation, the controller pulses each LED in sequence, and the corresponding photo detector current is measured. The detector's current depends on the amount of beam it r e ceives. The controller interprets this current to determine if an object is in the scanning area (Figure 2). Evaluation of Automatic Vehicle Classification Systems 30

PAGE 34

Transmitter Reoc:iver 0 0 0 0 LEOs De tectors 0 0 0 0 Range Connector Connector Figure 2. Schematic Diagram of Major Parts in a Scanning Area [SourceST!Broehure] When a vehicle passes between the transmitter and the receiver, some beams will be blocked Blocking of one or more beams is interpreted as a presence of a vehicle Figure 3 shows a typical light curtain in operation. The environmental en closures have heaters to prevent condensation and frost from building up on the glass, and a fan to keep the scanner from overheating in hot weather. These added features are included as accessories to the enclosu res. Evaluation of Automatic Vehicle Classification Systems 31

PAGE 35

Figure 3. Ligh t Curtain in Operation (Light beams are shown for clarity However, they are not visible in the field ) [Souxcc-STI Brochure] Scientific Technology Inc. (STI) in Fremont, CA; Banner Engineering Corp in Minneapolis, MN; and Sick Optic -Electronic, Inc. in Eden Prai rie, MN are the three major manu factures of light curtains in t he U S A Variations between each mode l range from LED spacing to the type o f accessories included. Some important variations are listed in Table 4 Evaluation of Automatic Vehicle Classification Systems 32

PAGE 36

Table 4 Differences among Light Curtain Manufacturers Description STI Sanner Engine ering Sick Optic-Electronic, Corp. Inc.+ Model# VSS604& series ME):\L60 16A and FGS 1200-111 MEEL6032A series Beam Spacing 0.5'', 0.75", or 1.5" 0.38 or 0 .75" 0.28" (7 mm) or Or combination of beam Same spacing per pair 0.79 (20 mm) spacing is possible Vehicle i when at least a beam is when 6" or more of scan when at least a beam Identifi cation broken length is is b ro k en Response Time 6.2ms 18 ms 15 IDS Power Requirement Interface Cost Note: ++ 115 V ACor24 VDC 105-125 V AC 24VDC RS -232 or RS-422 RS-232 or RS-485 RS-4&5 $3570 S3 302 $41&! The light curtains have been used in industrial applications only. They were not aware of th e light curtain applications under. toll plaza environment, and are investigatin g the possibility. The manufacturer requires t hat at l east a 6" of scan length be blocked to d etect vehicle separation_ TransCore contacts in Orlando, FL mentioned that they h ave redesigned the Banner light scanners for vehicle separation applications in the Orlando Orange County Expressway Authority (OOCEA) toll plazas. The cost includes for a complet e lig ht curtain pair of 4' high with aluminum ** enclosures and 0.5" LED beam spacing. The cost includes for a complete light curtain pair of 4 high with enclosures and 0.375" LED beam spacing. *** The cos t includes for a complete light curtain pair of 4 high with enclosures and 0.28" LED beam spacing. Evaluation of Automatic Vehicle Classification Systems 33

PAGE 37

CUIR contacted several toll agencies where light curtains are used for vehicle separation. A list of agencies, overall observed accuracies, and other important comments are summarized in Table 5. Every toll agency contacted that uses light curtains is pleased with the high accuracy and low maintenance required. Few agencies mentioned that some of their light curtains (at pre-classification positions) were run over by vehicles (especially when trucks changed lanes). Some agencies plan to protect light curtains by installing concrete or steel barrie .rs close to the light curtain enclosures. Table 5. Light Curtain Clients (Toll Agencies) io the U.S.A. Contacted by CUIR Agency Make Overall Accuracy Comments Kansas Turnpike Authority STI Very accurate* Needed a heater to overcome condensation problems Metro Transportation STI Very accurate* Needed a beater to overcome Authority, NY condensation problems California Transportation STI 99% Had no problems with fog or Corddor Agencies, Foothill condensation Corridor Dulles Greenway, VA Banner Very accurate* Snow piling caused problems OOCEA,FL Banner Very accurate Used Rain-X on glasses to minimize condensation Greater New Orleans Banner 98-99% Strongly encourage purchase Expressway Commission LA enclosures from the manufacture Note :*-Either initial tests were done and the reports are not available, or do not know of any tests and generally had no trouble with light curtains. + Kansas Turnpike Authority had tested both STI and Banner light curtains before making a decision on the installation. To function as a better AVC system, light curtains should be used with a device (e g., treadles) that could count axles. Light curtain and the axle counting device can co-exist on the same vertical plane. While the light curtain detects and separates each vehicle, the treadles for example, can count the Evaluation of Automatic Vehicle Classification Systems 34

PAGE 38

axles of each vehicle. However, a flaw e>Qsts in this configuration. That is, if a person walking across the lane blocks the light curtain beams and tread on the treadles that would stimulate the treadle to count the action as an axle, the AVC system may identify the person as a vehicle. This can . be remedied by having a loop adjacent to the treadle. Loops can diStinguish the presence of a vehicle . from a non-vehicular object (e.g., a person). Therefore, the signal from the loop can be used as a doubl e check before registering a blocked signal as a presence of a vehicle. Such a configured A VC system can be used to classify vehicles into any number of vehicle classes (where vehicle class is defined exclusively by the number of axles). 4.7 IDRIS IDRIS is a t echnique developed to enhance the output of a vehicle detector. The technique has been tested with a standard inductive loop layout and has bee n shown to identify individual vehicles in a traffic stream. The IDRIS system comprises of inductive loops and associated outstation equipment network manager(s), and instation(s). It includes six algorithms to improve input data quality and to collate and analyze the data. It can measure volume, speed, and vehicle length. During the initial tests conducted in U.K., the IDRIS has exhibited a vehicle count accuracy of 0.2% under congestion and of 0.01% under free flow traffic conditions. Initially designed for automatic incident detection in multilane open highways, Peek Traffic, Inc. feels the same technique could be used for vehicle separa tion under slow moving traffic conditions within a single lane at a toll plaza. CUTR met with Peek Traffic representatives on July 9, 97 to invest i gate this matter further During the meeting the expressed interest in conducting limited field tests of IDRIS atAnderson Mainline Plaza in Veteran's Expressway. Currently, CUTR. is coo rdinating this field test with O T O and Peek Traffic Evaluation of Automatic Vehicle Classification Systems 35

PAGE 39

5. SUlVIMARY OF FINDINGS Through literature surveys, vendor brochures, secondary field test reports, and conversations with professionals, the following important points associated with A VC devices were found: Active infrared technology exhibited good potential for vehicle detection during freeway tests. Doppler microwave technologies exhibited good potential for detecting moving vehicleS during freeway tests. Data collection performance at the intersection site was poor. Pulse ultrasonic technologies have moderate potential for detecting traffic at both intersection and freeway applications Video devices require extensive installatio n and setup time and are not as accurate as other technologies. However, video devices have the advantage of offering a variety of traffic data and surveillance information. Video devices require extensive calibration over serial communication lines Lighting conditions were observed to affect some video devices, particularly during the trl\Dsition from day to night. Full-scale field tests on SAM are currently not available in the U.S. Several agencies are still conducting tests on SAM's ability to separate/detect vehicles. Currently, SAM units cannot be repaired in the U.S. SAM units do not have any moving mechanical parts. Many toll agencies have a growing confidence in light curtains. Light curtains can detect trailer hitches effectively Light curtains have a very high accuracy rate (98%99%) in vehicle separation compared to other AVC devices. Light curtains have low maintenance requirements. Light curtains can be used with treadles and loops for better classificatio n accuracies. Light curtains do not require a reference background for vehicle separation. Evaluation of Automatic Vehicle Classification Systems 36

PAGE 40

SAM, Autosense II, and Sonar :Jrn.-500 units, if mounted horizontally,' require a reference ba7kground (e.g., wall, large column, et c.) for accurate vehicle separation. Light curtains, SAM, Sonar XR.-500, and Autosense.ll eye safe. Heavy snow, fog, and condensation affects the lig ht curtain's functionality. Evaluation of Automatic Vehicle Classification Systems 37

PAGE 41

6. RECOMMENDATIONS Based on the findings of this limited evaluation, CUTR recommends that the OTO focus its attention . . on light curtains apd IDRIS as future alternative means for vehicle separation (in continuing comparison to the improved overhead ultrasonic transducer). With the deployment of SunPass, accurate vehicle separation will obviously become even more critical. as other ETC agencies have found. Light curtains are becoming the technology of choice for reliable vehicle separation among other experienced ETC agencies. We feel this is a strong indicator of industry acceptance. The recent breakthrough ofJDRIS shows tremendous promise in improved vehicle separation capability, > as well as other applications related to ETC operations. However, the findings and general information contained in this evaluation are from secondary sources and we would advise OTO not to unconditionally invest in this system wide technology conversion without fLISt-hand evidence of accuracy and reliability under the current nine vehicle classification categories ( impact of fog and condensation are primary concerns), integration requirements, and maintenance requirements. A preferred configuration in conjunction with triggering loops and treadles must also be confirmed. At this point, due to Dade County's continuing software integration problems, consideration of the SAM technology for vehicle separation should be put on hold. No conclusive fmdings can be reached until testing is completed. (Since this is the first installation of SAM in the U.S.A., CUTR recommends not to disregard this technology totally) In combination with site visits to existing light curtain installations (identified in this report) it is also recommended that sufficient field tests be conducted under real-world traffic and environmental conditions to confirm capabilities of light curtain and IDRIS in comparison with the improved overhead ultrasonic transducer capabilities. These field tests should include at least two of the light curtain manufacturers identified in this report, IDRIS, and Polaroid transducer. CUTR believes as with the previous ETC testing, that these vendors would be willing to come to a predetermined test Evaluation of Automatic Vehicle Classification Systems 38

PAGE 42

site at their own expense. In these tests, it is recommended that in lieu of the deployment schedule for SunPass that the "winner" be declared the vendor of choice. t Evaluation of Automatic Vehi cle Systems 39

PAGE 43

BIBLIOGRAPHY 1. Cosentino, P. J., Grossman, B., Taylor, C Eckroth, W., Tongta, R., ll!ld Zhao, T., "Fiber Opt ic Traffic Sensor," National Traffic Data Ac q uis i tion Confer.ence, Albuquerque, NM, May 5 9, 1996 p .p. 629-643 2 Dodd, D "Video, Digital Imaging, and an Automatic Vehicle Classification System," ITE 1994 Compendium of T echnical Papers, p p 109-113. 3. Duckworth, G L., Frey, M. L., Remer, C. E., Ritter, S., Vidaver, G., "Comparative Study of No nIntrusive Traffic Monitoring Sensors," Proceedings of SPIEThe Internatio n al Society for Optical Engineering, Vol. 2344, Bellingham, WA, 1995, p.p. 16-29. 4 Elliott, C J Fisher, H Pepin, J., and Gillmann, R., "Beyond Scheme F," National Traffic Data Acquis i tion Conference, Albuquerque, NM, May 5-9, 1996, p p 17-31. 5. Gattis, J. L and Lee C E., "Trial Implementation of a Photoelectric Sensor System to Classify Vehicles," Transportation Research Record 1364, Washington, D.C., 199, p.p. 179-185 6 Gillman n R. "Vehicle Detector Test Center Status," National Traffic Data Acquisition Conference, Albuquerque, NM, May 5-9, 1996, p.p. 502-503. 7. Hamrick, J. and Schoenmacker, R., "Testing Manual for Testing Vehicle Detector; Traffic Surveillance, and Control Dev i ces National Traffic Data Acquisi tion Conference; Albuquerque, NM, May 5-9, 1996, p .p. 506 -510. 8. James, R. D and Sampan, S., "Vehicle Classification Using Ne ural Networks Ba-sed Upon Acoustic Signals," ITS Annual Meeting 1995, p.p 975-982. Evaluation of Automatic Vehicle Classification Systems ' 40

PAGE 44

9. Reel, R., "Vehicle Classification: Florida's Perspeetive," National Traffic Data Acquisition Conference, Albuquerque, NM, May 5-9 1996, p.p 89-97. I 0 Se.rway, R A. and Faug4n, J. S., "College Physics," 2nd Edition, Saunders College Publishing, New York, NY, 1989. 11. Tilley, D. and Thumm, W., "College Physics, "Cuinmings Publishing Company, Menlo Park CA, 1991. 12. Web Site, "Electronic Toll and Traffic Management," at http://www.ettm.com, March 4, 97. 13. Autosens e II Operator's Manual, Schwartz Electro-Optics, Inc Novembe r 1996. 14. Installation Instructions, Sensing and Activating Module, MBB-BD Sensorsysteme, Oc tober, 1996. 15. Traffic Observation Module, Laser Detection for Traffic Data Acquisition, MBB-BD Sensorsysteme, October, 1996 . 16. Piezo-Electric Automatic Vehicle Classification System, Final Report, Oregon Department of Transportation, Materials and Research Section, Report# SHRP-OR-RD-92-01, 1991. 17. Field Test of Monitoring of Urban Vehicle Operations Using Non-Intrusive Technologies ; Final Report, Minnesota Department of Transportation, Minnesota Guidest&r, and SRF Consulting Group, Inc., May 1997. 18 Accuracy of Traffic Monitoring Equipment, Georgia Tech Research Institute, Georgia Institute of Technology, June 1995. Evaluation of Automatic Vehicle Classification Systems 41

PAGE 45

APPENDIX A. List of Contacts Evaluation of Automatic Vehicle Classification Systems 42 .

PAGE 46

List of Contacts Syste:m Technician An.du:;oo. Malnline Toll Plua, Vetctlt'\ Elrpta$w.ay Ttl: 813-91$ Scralo Aguilar Applioation Engineer Sctentifio Tol: 800-221-1060, SI0-608-3400 (x-3lll) Pa.!Br.tten DulcoEog"61Z-S#-3 164 Made. Cantelli Manaaec. &sia.ecrin and lntea;ntion Ttansport3tioft Systems and Suvioo1 Lockheed Martin JMS Tel: 5044S47600 Doo.C..U on Toll System Project 1\.f.ana,gar Offic& of Toll Flori4a Department of'naruportatlon 9S4-S70-S39<> ICoy D. Dame< Oonllbm:U Eog;oeer Tel: 612,...$44-3164 1 D.. Haft Soulh Ftcrida Maial.tnanee Coo"""'*tor Otnoe ofToll Florida O.putmtnt of Ttaasporbtioa. T o l : 561-481 lk Dconis ICillio$'< ,..,..,.. Dcpotlxn
PAGE 47

APPENDIX B. ManUfacturer Specifications for Light Curtains Scientific Technologies, Inc. Banner Engineering Corporation Sick Optic-Electronic, Inc Evaluation of Automatic Vehicle Classification Systems 44

PAGE 48

VEHICLE SCANNER PROfiLING SCANNIR S Feat ures profiling scannm for scanning moving vehicles. Specifically designed for vehiclo detection and classification in auto mated toll systems. Three beam spacings available: 0 .5, 0.75, and !. 5 inches (12.7,19. 1 and 38.1 mm). Ma.,imum range of 80 feet (24.4 m), provides high excess ga i n when used for normal lane widths. High degree o f immunity to all kinds of ambient l i ghting. Four separate outputs: RS-232 or R S-422 serial data, parallel data, relay and optional analog out put Non-volatile RA.'I to store user's settings. Dt!ault mode suitable for many applications. Towers serve as heavy duty envi-ronmenta l e nclosures. -Rugged sta inless steel design or a luminum with corrosion resistant paint. Fan prevents overheating in sum m e r -Heater for tho optical window prevents condensation and frosl Special software ignores small objects such a s snownakes while still detecti n g narrow traile r hitches. @ CE 1!4.! SCIIN1U'IC 11CHNOLOGII S INC Description STI's VSS6000 Vehicle Scanner is a family of high speed profiling scan ners spuifically designed for vehicle detection and classification in auto mated toll syst e m s A complete Vehicle Scanner consists of a trans miller, a receiv e r a controller, two tower environmental enclosures, and interconnecting cablos. The transmitter houses a linear array of LED's (light emitting diodos) while the receiver has a lin ear array of photodetectors. Each LED and its corresponding photode tector form a beam. When a vehicle passes between t h e transmitter and receiver some of the beams will be blocked, and the vehicle is detected. 'fhe Vehic l e StaMer comes in scanning lengths of 2, 3, 4, 5 and 6 feet (610, 914, 1219, 1524 and 1829 mm) Each length is available with beam spacings of 0.5, 0 .75, and 1.5 inches (12. 7,19.1, and 38.1 millime ters) providing resolution of 0.75, 1.125, and 2.25 inches (19. 1 28.6, and 57.2 millimeters) respectively. C aJit 11800f221 USA and canada

PAGE 49

VSS6000 S1R1os PROFILING ScANNoRs -----------------------Vehicle Separation The VSS6000 provides a relay out put for simple on or off vehicle dete<:tion in separator applications. A special sortv.are function enables reliable vehicle detection using this simple control output. Some less sophisticated vehicle separators have difficulty distin guishing a tailgating vehicle and a trailer following a vehicle. When these systems are set to detect small objects, they distinguish between the tailgater and the trailer but small objects such as birds, insects, and even snowflakes can be erroneously detected as vehicles. When the system is set to detect only larger objects they fail t o distinguish the tailgater and the trailer. STI's Vehicle Scanner overcomes thi s difficult problem with a special software function that allows you to set separate tum -on and turn -off sizes for the outpul A high turn-on size prevents false detection of small obje<:ts as vehicles, while a low tum off size ensures the scanner detects only one vehicle if the vehicle's pro file changes to a narrow hitch and then to the trailer. C.tassiflc-ation The VSSGOOO is ideal for integra tion into classification systems. The scanner includes an 8 bit parallel output and an RS-232 or RS-422 serial data output that can be used for this purpose. An optional analog output is also available. For these .systems, the scanner provides a series of measurements as the vehi cle passes. The classification system records these measurements and interprets them to determine the class. Note that the scanner does not directly classify the vehicles, but pro vides measurements for an external system to perform this operation. The outputs lor classification and detection can be used simultaneously. Custom Scanners For applications requiring longer scan lengths or different beam spac ing including variable beam spacing, please contact STI. For appropriate quantities, STI will consider cU.stom hardware and software modifications to the Vehicle Scanner system. E.nvJronmentat Towers To protect the scanners from the elements and vandalism, the VSS6000 has rugged environmental enclosure towers. These are available in 316 stainless steel or aluminum with a special corrosion resistant polyurethane paint. The enclosures have a heater to prevent condensa lion and frost from building up on the windo1v, and a fan to keep the scanner from overheating in hot weather. The VSS6000 has a range of 80 feet, which provides an excess gain of 25 when used at the typical lane 1vidth of 16 feet. This gain minimizes the impact of dirt and grime. on the optical performance of the scanner. Setup The Vehicle Scanner uses a micro processor to control the system. Although the factory default settings work in many applications, some applications will require other set lings. Setup Is easy using a personal computer or an ASCII terminal. Alter changing the settings you instruct the controller to store them in non volatile RAt>!. A switch setting allows you to start the scanner using your custom settings or the factory defaull Setup commands are fully documented in the installation man ual available upon request. Spectal Functions The Vehicle Scanner incorporates many special functions. These are built into the operating software and can be activated through the setup commands. A brief synopsis is pre sented here, and further information is available in the installation manual or by calling STI. DoubleScan-improves resolution by a factor of 2 for some applica tions. Sgnchronous, hardu;are and soft ware triggers -allows you to update the scanner's outputs at regular time intervals or upon demand. Maskingallows you to manjpu late the reported status of each beam. Screenallows you to set a mini mum value to be reported. Quiet mode-inhibits reporting of repetitive zero values on the serial data output, minimizing the amoun t of data to be processed. Delta mode -reports only changes on the serial data outl)ut, minimizing the amount of data to be processed. Heartbeat-forces serial data out put reports at regular time intervals in spite of Quiet and Delta modes. Used to indicate the scanner is func tioning e1>en if no vehicles are detected 011tr a long time period.

PAGE 50

.y t .. Drawing ln. 0.0. (33. 3 mm) INDICATORS Transmitter only ...:::;_ _ TRAIISMITIER AOILt 0 .30 ln. DIA (7.6 mm) 0.8751n or1.131n. (22.2or28.7mm) KNOCKOUTS (4) BOTTOM & SIDES VEHICLE SCANNER V$56000 .r--SCANNING ZONE 1 .311n. (33.3 mm) RCEIVEII CABLE _:::::::,_ 0.30 ln. DIA (7. 6 m m ) CONTROllER/ ENCLOSURE

PAGE 51

. VS$6000 SRIES PROFIUNG SCANNERS Dimensional Drawings 755 = !l SG 0 Vebi cle Scanner fr z o 05 242.6 1 3 I AIK. 6.55 168.4 FROIIT YIEVI zu 7 SlOlS (4) us 41 C (MI G.) M 1 1.16 Z$3. 5 tlO. ) 52.1 A _J. 6 &.9 l FRONT VIEW VSS600o .. o .so SoJ'io-$ Dlme"sions ira/mm can Lenglh A 36 [ 914.4 3.!1/901.7 41 47 5/1 206 5 60/1524 7211828.8 71.511616. 1 1] 0 4.33 110.0 SIDE VIEW B 26.5/673.1 3 8 51977 9 50.5/1 2 82 7 62.5[1587.5 MTG HOLES (8) 7.6 274.3 8 c 3.375 85.7 V> :;; 5I SI'DE VIEW 28 .5/723.9 40 .5/1028. 7 52. 5/1333.5 .5/1 638.8 4.0 0 101.6 .. .. 2 .05 52. 1 BACK VIEW D D 29 .&a5u 41.6/1056.6 53.611361,1 65.6/1666 2 77.611971 o l.tll. 255.3 ..

PAGE 52

VEHICLE SCANNER V$56000 !.!!!, X Ul& D I A \ 3.. 20.5 1 80 IJdlG.) SlOTS (4) J..!--U5 'ui' ; ... 1 us == C (MIG. ) VSS60000.75 Scrios Dimen .1ions ln/m m ___r-o.r -A __l 1.80 45.7 Scan Lenglh A 8 c 3.375 lli... 119.1 l__J : :==::sl-o .?o 17.6 8 3 .(10 Mill 78.2 u c c ll
PAGE 53

V556000 SERIES PROfiLING ScANNERS ____________________ ___;_ Drawing VSS6000!NC..A\. Dimension ln/mm Scan Length A 8 241509.6 40.1nD19 31.Dn87 38!914A 52.1/1323 43.011092 48/1 2192 64.111628 55.011397 60/1524 76.f/1933 67.011702 --,5.57 144.0 132.1 1 .51 38.4-DOOR FASTENER_/ (3 PLCS) BASE PLATE SIDE VlEVI :..!_1= =_, 5/BHUHUT MAX AREA I : : r: t! : : 1.05 26.6 o: '// "-DOOR : V "-sWING AREA 1-t_J._. : _ ____ ...U.----l 226.1 TOPVIEVI .I-8 FRONT VIEW 4 .25 i65.1 107.9 SYM 19. 1 A 1D.4 264. 2

PAGE 54

Vtlllt\l Str.NMlR VSS6000 -a.n I 1<5.3 I ... LOUVERS DOOR LATCH (SPLCS) VS S 6 000 INC S.S Dimension i n /mm Sea B 141609.6 42.511080 31.51800 361914.4 54.511384 43 S/IIOO 4811219.2 66.511689 SS.S/1410 6011524 78.511994 67.511710 7211828.8 90 .8/2306.3 79.5/2019.3 OODR SWING AREA SIDE VIEW I o _ 8.1 I O.l!J 9.7 .lJ!. 31. 8 U&SYIOI 123.9 I I I I I FMl ;-1 +-----1 .!:!! m t 78. 7 I BASEPLATE I I CUTOUT I l _l_ ) ------I.1S SYI4 ---1 206.5 1 ---------25U TOP VIEW 13i. l -. . . . . . FRO I I! VIEVI OIA u .. ,.. ........ A 8

PAGE 55

V$56000 SIRlES PROFILING 5CANNIJIS ----------------------Specifications Beam Spacing 0.5, 0 .75, or 1.5 inches; 12.7, 19.1, or 38.1 mm. Other spacings available -consult factory. Scan Lengths 2 3, 4, 5 or 6 feet; 610, 914, 1219, 1524. or 1829 mm. Other lengths availableconsutt factory. Operating Range 16ft. (4.9 m) with excess gain of 25. 80ft. (24.4 m) with gain of 1 Minimum Deteclabte ObJect Beam Spacing Minimum Ootectab-le Object 0.5 in (12.7 mm) 0.75 in (19.1 mm) 0.75 i n (19.1 mm) 1.1251n (28.6mm) 1.51n (38.1 mm) 2.25 in (57.2 mm) Nominal, varies wilh application Scan Rate so x the number of beams set plus 200 ps. When the serial port Is used to com municate excessively long data, the communication time may exceed the scan time. Scan time will be delayed until communication Is complete. Serial Communications RSZ32 or RS 19 ,200 or 38,400 baud rate. Parallel Output 8 bits, 0 lo .s VDC transition Sources or sinks 16 mA. Optional connection kll pro vides terminal b lock w iring for lhe parallel oulpul (Model No. VSGSOOPCK). Separate DATA VALID blllndicates when o u tpulls in a valid state. Analog Outpul Model no. AIC (Opllonat) Selectable for -41o mA, to +10 VOC, or 0 to +10 VOC. Analog Relay Card Model No. ARC (Optional)' Provides 4 relays driven b y the analog output. Each an lndependenl P .Otantlomater to set the 1um-on point, an Indicator LEO, and an SPOT e lectromechanical relay. Parallel Relay Card Model No. PRCOZ (Optional)' Provides four relays driven by the parallel output. Each relay has SPOT conlacts (raled for 100mA at25 VOC maximum) and Is driven by one bil of the parallel Oulpul. Two cards are required to cover all eight bits of the parallel output Optoisoiator out puts are available in place of relays (Model No. PRCS). (Optoisolators are rated for 40 mw max. power, 75 mA max. current, 40 V max.)

PAGE 56

Par all el Connecllon Card Model No. P CC-02 (OpUonal)' This car d p rovides screw tenninal connections for the parallel out put Standar d connections from t h e controller board are through ribbOn connector pins. Thi s car d allows for easy w iring without using ribbon cable. O plion Card s Each row of the following cha rt sho\vs one o f t h e allowable c .omb inatioils of output card s t h a t.can b e Installed No olher combinations a,re possible. AIC-02 ARC-02 PRC(S) PCt:-02 1 1 1 1 1 1 1 1 2 Relay Out p u t SPOT ; 25 v o c maximu m voltage 100 mA maximum current. Do not use t h i s relay t o control AC voltage s Power Requirements -Scanner VAC, 230.VA C <10%, 50/60Hz o r -2 4 VOC 2 VOC, 100 mV ripple maximum, 30 VA powe r c onsu m p tion Environmental Towers -120 and 230 VAC versi ons: 240 VA m aximu m -24 VAC versi o n s : Scan Length Power Requirem e nts 2 4 1 n (610 mm) SOVA 36 in. (914 mm) 80 VA 48in .(121 9mm) 130VA 60 i n (1524 mm) 175 VA H ardware Trigger -Opti c a ll y i s olate d input. Signal tran s ition b etween 0 a n d .5 VDC at 1 0 to 20 mA requi r e d Positive signal is aetNa. Be sur e to limit current to 20 mA maximum. I nput circu it does not provide limiting. -If trigger persist s for l e s s than the scan t ime, then o n e updat e of the outputs w ill occur. -If trigger persi st s f o r long e r tha n o n e scan time, then the output s will update contin uously w h i l e trigger persists. -Minimu m trigge r time is 200 ps. Transmllter and Receiver Cables Maximum Length: 100ft (30m) Environmental raling S canne r: NEMA 4 12, tP 65. E nviron m entatTower. NEMA 1. ll!..rl.! S CIENTIFIC nCHNOLOGIU INC. V HICLE SCANNER V556000

PAGE 57

V556000 SIRlES PROFILING SCANNERs---------------------Temperature Rating Scanners a lone: 32 to 131(o to ss C). In environmental enclosures: -22"to 131"F (-so to SS"C) Consult factory If your application exceed these temperatures. Endurance:. -X, Y,ZAJ
PAGE 58

' . ----------------------VE HI C U SCANNER V 556000 Shi pping Wolg hls For complete systems includin g transmiUer, receiver, controller (with enclosure), cab\es, and env\ronmenta\ tower enclosures (approximate). EnctOsuce Matettat A luminum Stainless Steel Selection Charts Transmitters Scan Length 24 inches (SI D mm) 36 inches (914 mm) 48 inches (1219 mm) 60 inches (1524 mm) 72 inches (1829 mm) Retelvers Scan Length 241nches (610 mm) 36tnch" (914 mm) 48 inohes ( 1219 mm) 60 Inches (1524 mm) 12 lnclles (1829 mm) I 241n 36 in 61Dmm 9 1 4 mm 841b 1021b 38. 1 kg 46.3 kg 901 b i 1101 b 40.8kg 49.8kg 0 5 in<:h (12. 7 mm) VSS6G240.SOX VSS5036.60X vsssoo-o.sox VSS60600.60X VSS6072.SOX O.SI.1ch (12.7 mm) VSS6024.0.50R VSS60360.50R VSS6048 0 .50R VSS60EOO.SOR VSS5012.0.5DR Scan length 481n 16Gin 721n 1219mm 1524 mm 1829mm 120 l b l1381 b 11531b 54. 4 kg 61.2 kg 69A kg 1301 b 1150 lb 1701b 59. 0kg 68. 0 kg 77.1kg Beam Spaeirig 0 .751nch (19. 1 mm) 1.51noh (38.1 mm) VSS60240.75X VSSW24-1.50X VSS6036.0.7SX VSS6030-1.50X VSS6048.0.75X VSS6048-1.50X VSS6060.75X VSS6060.SOX VSS60720.75X VSS6072.5DX Beam Spaeing 0.75 incll ( 1 9 1 mm) 1.5 inch (38. 1 mm) VSSSOZ4-0.75R VSS60241.SOR VSS6036.75R VSS6030-1.50R VSS6048' 0.75R VSS504a.t .SOR VSS6060.0.75R VSS6060.50R VSS6072.0.75R VSS5072.60R Aluminum Envlronmental Tower E n t loS\Ires All en-closures include heater ventilation fan at1d yfC>t k lOt transmittus or Operating Voltage for defJO$leJ and fan Scan length 120 VAC 24 VAC 230 VAC 24 inches (610 mm) VSS602HIIC.AlAC120 VSS602HIIC.ALAC24 VSS602HIIC.AlAC230 36 inches (914 mm) VSS600HNCALAC120 VSSSOOHNc-AlAC24 VSS6006-ENCALAC230 48 Inches (1219 mm) VSS604HNCALACI2 0 VSS6048-ENC.ALAC24 VS$6048-ENC.ALAC230 60 Inches (1524 mm) VSS6DmNCALACI20 VSS6060-I!NC.AlAC24 VSS6061WJC.AlAC200 60 lnclteo (1524 mm) VSS6060ENC.AlAC120 VSS6060EIICALAC24 VSS0060EIIC.ALAC230 721nclles (1829 mm) VSS6072-Eilc-AlAC120 VSS6072ENc-AI.-AC24 VSS6072-Eilc-AL 'AC230 Stai n less S teel Envlronmontal Tower E nclosures All e nclosures include heater and ventilation fan and wor1< for transmitters o r receivers. OperaUng Voltage tor defroster and fan Sean length 120 VAC 24 VAC 230VAC 241nch,. (610 mm) VSS6024-ENCSSAC120 VSS6G24-ENC.SSAC24 VSS50e4EIICSSAC230 3Sincltes (914 mm) VSS6006-l:NC.SSAC120 VSS5036-ENCSSAC230 481ncltes (1219 mm) VSS604&-91C.SS.AC24 VSS6048-ENC-SS4\C230 60 inches (1524 mm) VSS6060EIICSS.AC120 VSS6060-ENC.SS.AC24 VSS6061HNC.SSAC230 72 Inches (1629 mm) VSS607HIICSS.AC120 VSS607HNCSSAC24 VSS6072EIIC.SS.AC230 Calh 118001221 .. 7060 USA and Canada ' ..

PAGE 59

VSS6000 SERIES PROFILING SCANNERS---------------------Conlrollor, Firmware, and Documentation VSS6000-cTLfl2 Controller with HEMA 4 enctosure, for AC 120 or 230 V VSS6000-CILR2PCB VSlHW1 VSS6000DOC2E Cables Coble length 25ft (7.62 m} 50ft (15.2 m} 75ft (22.9 m } 100ft(S0.5m} Options Model No. AI= ARC-02 PR= PRCS PCC-02 25194-DOtO 25194020 Orcloring C'hockli$1 Controller clrcu1t bovd assembly (no fOt VOC Standard fimwtare for operating con t roller Documentation kit, including installation manual and demonstrationfsttop software Transmitter Cable Receiver Cable OCX OAR-25 acx-5o OAR-50 OCX-75 OAR-75 OCX-100 OAR-100 Descrlplion AnalOg isolator Card: p rovides a nalog output orO l o +10VOC, -10 to -t10VDC, or current loop o t -4 to -20 rnA Analog Relay Card: prondes 4 SPOT relays d r l"'n by all3log output. Each has a potentiomete r t o set the poin t and an LED indicator Parallel Relay Cord: protides 4 SPOT relays drl"'n Ill' parallel ou(put. Each has an ktdk:ator LEO. P rovides 4 opto Isolator outpul$ d rlv$n by paralle l output. Each has an Indicator LED. Paraltal Connection catd; p rovides screw terminal <:onnettions f o r parallel output. R ubbBt mounting gasket for aluminum t owers.. RU:bb&r mounting ror stainle$$ steel towers. For one complete Vehicle Scanner system order the following items: One transmitter One receiver (must be the same beam spacing and scan length as the transmitter) \vo tower enclosures One controller-specify if operating voltage is other than 120 VAC. Firmware kit Cable: trans m i tter t o controller Cable: receiver to controller Documentation kit Any option cards. C.l'lll!.,..,.t:',,. .. e,.uu"''"'r.::a c ,.,,. Applications Ass1stcmce If you need assista nce in selecting the proper scanner or need to discuss possible custom configurations p lease call our applications engi neering department._ A CAunoN STI scanners are not designed for and sh.oul d n o t be used for personnel protection (safety) applications. STI has a wide range of safety products designed for use in these applications. Refer to the Safety Products section in this catalog or contact STI for information on these products. ..... . : ..

PAGE 61

. Light Screen System . : . . . . : . : . .... ' : 1 :., .. ,.,, . ... . .. . MIRIl\RRl\Y features -The Ban ner MINI-ARRAY i s a p r ogrammabJe measuring light screen ideally suited for inspection and profiling app l ic Mions. lt excels a t the fly product sizing And p rofoling, !dg e-gui d illg and center-guid i ng, loop r eosion control, hole detection pans :oun ting. die ejeet\on ver\facadon. ao.d )imila r uses. Each system consis ts of a :on t roller module an :m i ne r and receiver, and wo int erconnecting cab le s r he MINI-ARRAY's :ompac1 controlle r module ni\y be eit her factory Ol' Jse r programmed for any me or two of ten neasurement modes and my one of four scann ing no d es. Progmnuning is ccompHshed vin the on1roJicr"s built-in serial nterface using a PC ompat1ble computer unning Windows" 3. I 'v'indows' 95, or OS/2'. software' displays for analysis and senso r receivers are available in .. llrray l engths ranging from 6 inches 10 4 _'feet in 6 increments. plus 5 and 6 foot models : All are available with I ___ either .375inch (9,53 mm) or .750-in ch (19.0 5 mm)' beam spacing, which io e\the r 32 or 6 beams per foot, of IJ'IY length. Sensors with ::; 3/8: inch beam spacing h ave a sensing range of up to 20 feel (6.1 meters). Sensors with 3/4inch beam spacing have a sensing range of up to 55 feet (16.8 meters). System status. i ncluding sensor alignment, is disp l ayed via L E D i nd icator s on bc:uh. the Jntc-o\ler and th<:. sensors. The controller moduie offers tw o discrete oul p uts (o n e reed relay and one de relay) which can be separately assigned to any of the ten measurement modes. The MINIARRAYs programming versatility also a llows t he user to "blank" areas of t he array to ignore objects that must pass u ndetcc r ed \hrough \he-ligh\ U tiliz ing the MINIARRAY 's se rial interface, up to 15 MINI-ARRAY controllers may be assigned separa te IDs and either p l aced on an RS485 party line or controlled by a hos t Scannlng can either be set for continuous opera t ion o r triggered by a dev ice connected to the optically isolated "Gate input. In lieu of a presence sensing device "Hos t Mode may be used in which scanning is initia te d by either a host computer or a prog.ta1n mable logic controller Compact Emitters and Receivers A complete MIHI-AI\1\AY system includes the following components: I Model MAC-I controller module (includes programming softw are ) 1 c.a. Senso rs: emiuer and r eceiver of equa l length and bea m spacing (see model listing on page 6) 2 Sensor cables (see a vailable lengths on page 7)

PAGE 62

Meas u ring L i g h t Screen MINIARRAY System Measurement Mode s ihe MAC-I controlle r i$ a bl&Jlly v ers a ti l e mlcrocontroller-based module. Jt is configured usi ng the su pplied Banner softw a re vi a its bu ilt -In RS232 interfac e U $ing a PC-compatib lc. computer runn i ng \V'mdows 3.1 or 9 5 or OS/1.". ihe MINI-ARRAY Sy stem may be pr ogramm ed for any one o r two of the followin g Scan Analy s i s Modes: First Beam Blocked ( FBB ): The eootroUer identifies the location of t he first b locked beam. Bc&ns a r e numb ered, beg inning at the cabled end or the senso r M d continui n g in sequence to t he owosite end of t he sensor. Last B e a m B loeked (LB B ): The oontrollcr ide n tifies the lout i on o f the last blocke d beam Total B e am s Blo c ked (TBD): The controller totals t he number of bloelced beams. Contigu o u s Beams Blo c k e d (CDB): The controll e r totals the nurnbcr of contigu o u s beams blocked in eac h group of bloclced beams along the lec>g th of the sensor. First Beam Made (FllM) : The controlle r identifies t h e IO<:ation of t he first u nb l o cked beam . MINIARRAY Programming Features L3s t B e a m Made (LllM): The controller the location of the l ast u n blocked beam. Tota l Beams Mod e (TBM): The controller t ota l s t he number of unblocked beams. Co ntiguous Bea ms Ma d e (C B M): T h e controlle r t o t a l s the number o f contig uous unbloc k e d beams in eac h group or unblocked beams along the length of the sensor. All Da ta (A.L L): The con trolle r passes all beam con ditio n d a ti.' for every scan t o the serial int e rrace ror -analysis by a host comput et. or controll er. Vehicle Separation (VHS) : Controller output II L i s energized whe n ever six or m o r e inches of curtain length a re bloeked (i. e. con tiguous bems bloeked), and de-energizes when all beMnS become unbloc'id; unstable outptt c onditio n s "ch a tt er ing" of the output) .. scanning condition exactly n of the set points. Scan # is t he n umb er scans o f the array th at are before the a s sociated The contro ll e r may be ;r ... .. .... .. from 1 to 9 coosecutivc seaJ\'. : d ata must be identical for co n secutive scans in output s 10 be updatOfl, '; . . ;: . ... .. "(i:ll. ;V) : . "" .. .. .. Outp u t controller mOdule, cAn be prosrammed t o function as lhe alrm output for the module's.,-"1::.; self-diagnostic whene ver it not tO scnn an al ysis O m p ut #2 m{ be. .. serve lrig,ge;r the scan of )lnO
PAGE 63

. . .. : .. . . . . MINIARRAY Scanning .. __. ..... .. . . The C?ntrol modUle. may be configured .-'< Edge only t he beams for aa'fOrle' o( f our the top edge of an object is iil: the Hght screen. (NOTE: \ : Straight Scan is the default mOde.:AUJ..: .. utop edgfi . refers to the edge of the beams are. scanned:.irLsequence from object pasSjng nearest to t he top end of eila to. t\ic.j>pposjtcfeM 9fthe f!!" the i. e. the top of the light sensors. : : ._ s creen) Each scan begins six bams I ... _ _. ... -..J .. prior t o beam blocked during IntcrJaced :--c;Jh e The scan scan with a sran-ted:beam SCctwccn the las t iteXt to Scan mode limit s the scan analysis Ia" receiver. The cin)iiier is. then mode' selection to L llB (Last Beam reactivated t o establish a .t>ea m to .. .' Blocked). last Ieceivt r tO Skip Scan quickens sensing response str aight scaDs ; . cime at t he expense of decreased sensing resolution within one third resolut ion The Skip Scan mode allows 'of scanning from one to seven beams to be skipped figure, below. "' '1.\ ':;..._ . durin g each scan. For example, with : one beam skipped, only beams #I, 3. 5, "'': ./ 7, etc will be interro gated With two ResolutiOn i s inctaasad :..... beams skipped only # 1 4 7 JO etc. i n the middl e 113 of "\viU be interrog' ated, and range .. .. ..... .. ........... , > ;, c Controi$fscanning : . ""'' controller may be for either c ontinUous scanning: The GAlE "'-,\"P''''' energized by applic ation of JO ..9c!;' 'Gating is tYPically acc6rhpliShcd using a de presence _y sensing device. Alternatively, die MINI-ARRAY System may be ... p rogrammed for "Hos t Mode" ._. control, ,Yhich allows scanning to be initiated by a host computer o r a lo g i c contro!lci: (PLC) . -. r \':'<:;.: ., :.,; Host Control The MINI-ARRAY System can communica te with either a host computer or a controller via either RS232 o r RS-485 seri.al protocol. The host can direct the M IN1-ARRAY Sys t ein to scan on demand andfor receive the scan data directly from the MINI-ARRAY Sys tem in binary or ASCII form. Se1ectabJe communication baud rates are 9600, 19200 and 38400. Programming of t h e Controller Module Configuration of the MINI-ARRAY control modu l e is done using the Banner-supplied software and an ordinary PCcompi\tiblc computer Programmed system conti _gurations. called "Parameter Setup Files" (o r PSFs), can be sto r ed in t he controller modul e's non .. volatile memory. The supplied software can c r eate and store multiple PSFs in computer files for i nstan t call up of a particular confl_guratiOI\ whenever it is needed. The Banner software also two o t her useful features. An Alignment screen displays the individual status of each beam along t he en tire length of t he array. The same screen also summarizes the total beams blocked and unblocked and indicates the beam numbers of the first beam b l ocked the fust beam unblocked t he last beam blocked, and the last beam unblocked This screen is invaluable during setup for monitoring and analyzing exactly what is being seen by the light screen The second feat ure, a Diagnostics screen indicates prob lems with th e emiuer or 'he rec eiver.

PAGE 64

Measuring light Screen MINIARBAr System B MEL6 . A emitter 201 BMRL6 . A receiYer BMELI8 .. A emitter 505 BMRLIS .. A re c eiver BMEL30 . 1\. emiuer 810 BMRL30 .. A receiver BMEI.A2 .. 1\. emiuer 1115 BMRI.A2 .. A receiver ,. , BMEL60 .. 1\. emitter 1 572 BMRL60 . A receiver 7.9 1 43 1 9. 9 448 31.9 752 43.9 1057 61.9 1514 Sensors 5. 6 2 133 5 .25 17.62 4 3 8 17. 25 29. 6 2 7 4 3 29.25 41.62 1048 41.25 59.62 1 5 05 59.25 MAC-1 Controller Module 100.0 mm < (3 .94") 110.0 mm (4. 33") 3 8.1 mm squa r e ( 1 .50") Cab l e s are order e d 1 s eparat ely (see page 7) M odule may be mounted directly to a s urface usi ng supplied hardware or onto standard 35 mm DIN rail.

PAGE 65

. MINI ARRAY Measuring Light Screen System . ... 'rv .. ;!!: .. .... -.,.' _.,._.[:. BML3016A Emiuer 752 mm 4 0 BMR!-3016A Receiver (29.62 in) BMEL36i6A:.:.'miiter . c<"' BMEL4216A Emiuer 1057 mm 5 6 BMRlA216A Receiver (41.62 in) BMEL48 16A : BMRL4816A "-I' BMEL60!6A Emiuer 1514 mm BMRL6016A Receive r (59. 62 in) BMEL7216A .Emitter : \JJ!.N.':lJim" 96.BMRL7216A Rece i ver : . 318" Beam Spacing 0 6to 6.1 m (2to 2 0 f t ) for sensors< 4 feet 0 6to 4.6 m (2to 15f t ) for sensors> 4 feet 314" Beam 0.9 to 17 m (3 to 55 ft) for sensors < 4 feet 0 9 to 14 m (3 t o 45 ft)forsensors > 4 feet Nott: Moximum is spuifitd at the poitt! wllere 3x t:tcess gai11 rtmains 3/8" Beam Spacing 3/4'' Beam Spacing 19.1 m m (.75 in) 38. 1 mm (1.5 in) I n t erlaced Mode: 12.7 mm (.5 in) Interlaced Mode: 25.4 mm (1.0 io) *Note: Assumes uming i s in middle Olletllird of scanning range (ste page 4 ) 55 m icroseconds per beam, plu s 1 m iJiiseoond pro..--essing lime per scwt Emitter 0 .10 amps max.@ 12V de Nott: Maximum tum'llf i:s for tJ 6 /()0l sms or Receiver 318-inch beam spacing0.75 amps max. @ 12V de 3/4-inch beam spacing 0.50 amps max.@ 12V de Sensors CQil!'.ett toro-.ooller using tw05<000.Ctoc quick disconnect cables (ore foc emitter and one for recciver), ordered separntely. Usc only BIIMCr cables, whidl inco!porme a "twisted pair" for noise immunity on RS485 data communication lirles. Cables l11f2SUre .32-i>:h (8.1 mm) in and are shielded and PVCjacke>ed. Conductors rue 20-gauge. Eminerand receiercables may not exceed 250 feet, ..ct>. See bol10m of page 7. Emitler Red LED Jights for p roper oper.uion Si:<: see Figure, pag e 5 Receiver Oren = sensors aligned (> 3x excess gain) Yellow= m arginal alignment (lx
PAGE 66

MINI-ARRAY Measuring Light Screen System . MINI ARRAY Controller Specifications 16 to 30V de@ I .25 iunps max. (see current requirements for sensors); controller, alone, (without sensors con ne cted) requires 0 1 amp. MJ.Nl ARRAY sensor input (5 emitter and receiver wire in parallel to five tenninals GATE input is optically-isolated requires 10 to 30V de (7.5K input impedance} for gate signal. Outputs are inactive for S seconds afrer system power up. Maximum response time for the discrete ou tputS are two scan cycles. A scan cycle rnclude s a sensor scan p lus any serial data transmission. Serial transmission (jf activated) follows ovcry sensor scan. Output I (OUT I) Reed relay contact rated 1 25V ac/d c max., 10 VA max resistive load (non inductive). I Output 2 (ALARM) Open collector NPN transistor rated 30V de max., ISO mA max, shan circuit protected; may be configured as a second data analysis output. a system alarm output, o r a scan trigge r output for a secondary device (e.g. came ra MINI-ARRAY etc.}. OFF-STATE Leakage Current: ON-STATE Saturation Voltage: < lOpA@ 30V de IX), DIAG I (green) Indicates power is applied to the mod u le, DIAG 2 (red) Indicates receiver fai lu re, DIAG 3 (red) Indica tes emitter failure. Size: see Fig u re, page 5 Material: Polycarbonate Raring: NEMA 12 (IP 52) -20 to +70C ( 4 to J58F); 95% re l ative humidi t y (non-condensing). Cables (2 required per system) Model Length QDC-SlSC 4.6 m (15 ft) cable, straight QD connector QDC-525C 7.6 m (25 ft) c able, suaight QD connector QDC-SSOC 15.2 m (50ft) cable, straight QD connector

PAGE 67

;.,., ll"o fo11<1 :.... 0 ;.t r. .. When yoifllily your ... the confid.cnce;of-de!lling W)tll the . ;,:::!!....... __ :.: l'' ' nation'slargest; mos.r.knowledgeable and experienced : company. We have the broade s \ line of in sensor )nd\fstcy. !We can )landle any s(nall, utilizing.;. thc.mo'st"ailvanced'inamifacturmg _of : more than 1 just three dayswithin --............ _,,,_ ... ..... ... .. . .. _:. : :? { : : : Just a$.irnP9.,t;laJ\t;,>e have _the . largest photoeleeuic sales and .. a Jnost knowledgeable dtstributo 'i s and sales :-<.): ...... . .. !:>J :w.or\dWi,de. . ""'' ... !5y wlii'iever 'you're and w e're-ready to help you wit h your applications, give you excellent seJVice support. . When you add it a ll up, you'll find more value in Banner products. For more information or applieatic:ms assistance: ' . ... . Call 612 ..... ---E ' .... . .. !" r Banner Engineering Corporation . P.O. Box 9414 ' Minneapolis, MN 5544Q ', Phone: 612-544-3164 Fax: 612-544-3213 .............. .. WARNJNG! The!e photoelec-tric presence sensors do NOT include the self c h ecking redundant circuitry necessary 1 0 allow their u s e in perSOnneJ safety application.s. A sensor ot controller failure or malfunct ion can result in either an energized or a de-enccgi?:cd co ndition Ne\'cr use t hese p:oducts as sensing for perso;tnel proteciion. Their use as a safety device may c reate a n u nsafe condition which cou l d lead to serious injury or death Only MJNI.SCREJ;N", MULTI .SCREEN'", MACHINEGUARD. a nd PERIMrrfERGUARO Systems, and other systems so designated, are desi$ ned to meet OSHA a n d ANSI ma<::hine safety S-tandards for point-of-operation guerdmg de\'ices. No other Banner sensors or co-ntrol s are designed to meet these s1anda.rds, and they must not be used as sensing devi ces for personnel protection. 1508002 DOCUMENTED OUALITV .

PAGE 68

I t TECHNICAL DESCRIPTION FGS 100 co FGS 1800 -Safety Llaht Curtain (Grid Principle) t> HAZARDOUS AREA AND POINT -oF-OPERATION SAFEGUARDING t> SMALL HOUSING D IMENSIONS, MAINTENANCE-FREE . . . : . . . . . t> FULLY ELECTRONIC !iii:K optic .:: . : . ' . .; . . / . .-. ' . . . . . .:,:_. .. . . .. : . . . .. : ... . : : : .

PAGE 69

2 Contents Page I General 3 7 1.1 Features . . . . 3 2. A pplication Possiblll ties and Condition s 4 8 2.1 Areas of Application . 4 2.2 Applic.tion Conditioi\S 5 2.3 Comer MirTOrs . 6 !I 3 System Design 6 9 1 4 Theory of 9.2 Operation 0 7 4.1 Mode of 7 9.3 4.2 Me>nlng of LEOs . 7 4.3 Ho51/Gueu ea.ade (Daisy Chaining) . 8 4.4 Dia gnostia . . .... 8 s Mechanical and Electrical Installation ... 8 5.1 General ........ 8 5 .1.1 Safety Distance to Point of Operati on . 8 5.1.2 Distanc e t o Reflective S urfaces . . .... 9 5.1.3 Multiple Safety Systems per Application . . 9 5.2 Med\anical I nstalltion. Mounting Btackets . . 10 5.3 Electrical installatio n 11 5.3.1 Test Input (Sender) ... 11 5.3.2 Switching Outputs (Receiv er) . ... 1 2 5.3.3 RS-485 Data Interface Oiagnostia . . . 1 2 6 Specifications . .. 13 Page Agency Approvals Dimensional Drawings ... . 14 FGS . . . . 14 Connectio n Diagrams . . . 15 Selection of an Opt'l Electronic Safety System . . . 16 Selection Tabl e Sensors 16 Co mer M irror E U Europ e TOV Rheinland Am Grauen Stein 0 51105 J
PAGE 70

I General . FGS safety light curtains are active opto-e l ectronic protection devises (AOPD) w it h resolutions o f 14 mm or 30 mm. SICK FGS safety light cur1ains are self-checking ( AOPD) and belong to safety Category 4. They habe been designed for industria l applications and are characterized by: 1>High degree of reliabi lity ,... Maintenance .. fr ee components 1>-Practical size 1>Sturdy modul ar construction 1>Universal application possibiliti es 1>Simple i nstallation For practical application, the following specifications ar vali d : fig. 1. Threshold Values of an fGS Safety Ught Curtain ,... Maximum scanning range 1>Resolution (object sensitivity) 1>Maximum he i ght of p rotective field 1>Minimum height of protective field 6m/18 m Hmm /30mm 1800 mm 300m m The optical axis (center of exit wind ow) is marked on the housing. Features Special fea t ur es of the FGS safety .,. S maJI housing cross-section: light are : 52 mmxSS mm .,. Fast response time .,. Maintenance-free op e ration .,. Large excess gain .,. Co n trol reliable solid state .,. Possibility to chain numerous outputs sensors (to prevent p erso n .,. State of the art m i croprocessor from standing bet'IN'een s ensing technology with cus t om fields) designed !Cs ( ASICs) ... Construction in compl i ance ... Easy assembly a n d alignment with EN 50 1 00 Cat. 4 (self' .,. Internal te rminal strips or checking) quick-di sconnect ... Tested and approved for .,. Long range a n d high Europe, USA and japan resolution 1 General ' I l 3

PAGE 71

2 Application Possibil ities arid Conditions 2 Application Possibilities and Conditions fig. 2 Poi n tof.Operat\o.n s a .feguarding with an fG'S safc.<;y tight curtai n I -4 I f'rg. 3 Haurdous Area with a n FGS ia.fety tight curtain .. . .. :1.1 Areas of Application FGS safety light curta ins can be used f o r poi nt -o f-operation, (Fig. 2), are a (Fig. 3), per imeter. and entry/ ex i t (Fig. 4 ) safeguardin g a p plicat ions wit h short saf ety d ist ances. Typical a reas of applicati o n are : ..,. Presses in t h e me tal, p l astics, r u bber, leather and stone industries Punche s i n the le at her, text ile and plastics p r ocess ing industries F i lt e r p res ses .., Trimming presses and cutte r s Processing ma ch i n es and w e l ding presses Ass embly lines ).-Revolving t ra nsfer machines Automatic assembly e quipment 1> Paper cutt ing machines P. c

PAGE 72

Fig. 4. Entry/Exit SAfeguarding w ith an FGS taf c ty light curtain l .l. Application Conditions The machine o r system contro l must be able to be stopped e lectrically. T he FGS monitors the switching outputs f o r short circuits between them or against +V. The hazardous machine motion must be able to be stopped at all times. When Installing the FGS, keep in mind that reaching over. under or stepping behind the p r otective field to reach the machinge must be made impossibl e (Ag. 5). All pertinent regulations apply to installation and operation. Local authorities or the respective i ndustry associations are available for safety questions. Ag. S. The devices must be mounted $0 th.at the following arc not possib le.; reaching over, reachin,; u n del' and standi!'ll behind fields. -s

PAGE 73

3 System Design FGSS Fig. 6. HazardooJ S a re a safeguarding w ith th e FGS a n d a com e r m i rror. . Fig. 7. Th e FGS safety light curtaln COI'ISist'S of $et'lder r ece i ver. Between these is the 6 p rotective field. :1.1 Corner Mirrors The protective field be tween se nder and r eceiver can be guide d arou n d a corner using a comer mir ror (Fig. 6). This allows for safeguarding two s i d es of a hazardous area o r point o f operation. N ote that this reduces the scanning range 0.5 per m i rror. More th an t wo mir rors are not recommende d d ue to the alignment diffic ulties th 'IS may cause The safety light c urtain consists of a sende r an d a rece i ver (Fig. 7). Between these i s the p rotective field defined by the protected height and the s canning range. The length of each system Is dependent on the protected height requi re d The u p per and l ower edges o f the protective field are marked on the ho u s ings. Synchronization between t h e sende r and receiver takes place optically. i.e. electrical connect ion between sender and receiver is not necessary. The FGS is a modular construction. All optical and electrica l components are housed in one sturdy, yet slim profile

PAGE 74

4 Theory of Operation 4 Theory of . ' ... The (infrar ed ) l ight p u lses coded by the .sender are sent to the receiver, which evaluates them.lf an object. e.g. a finger or a hand(> H mm/ Sender > 30 mm). penetrates the protective field. bolh outputs or lhe FGS switch to stop the machine. As soon as the protective field Is cleare d, the outputvohag e automatically switches on again. 4.1 M o d e o f Operation The FGS operates in the guard only mode with external restart (automatic restart).lf other functions should be requ ined, e.g. PSDI mode (break mode), exiemal re lay monitoring, blanking, and others,lCU (Universal Safety Interface) devic .. are avai lable. With the addition ofthe LCU. p ractically all applicati ons are possible. 4.1 LED 8 shows th e function of the Indi cator lights on lhe FGS. Hos:t Sensor "I 4.3 Host/ Guest cascad e ( D aisy-Chaining ) In order to effectively prevent someone from standing behind the sensing fields, one or severa l FGS devices can b e connected in a daisy chai n configuration. External switc h ing is not necessary. Any desired combination is possible, ror exam .. pte, a 1200 rnm devke with 11 mm resolution for finger protection. with a 300 mm slave device. with 30 mm resolution to prevent someone from standing behind seming fields. When using the LCUPwith 'mixed" daisy-chain ing (e.g. host scn1or with ............... (Vet)
PAGE 75

5 Mechanical and Electrical installatio n j The guest sensor u nlik e th e m a i n sensor has no indicato r l i g hts It is I not des igned to be u sed as a s tand a l one device. T he H ost/G ues t j sensors with acce s sor i es are alread y : p re pa re d for daisy-chaining. I I 4.4 Diagnostics Th e FGS has a service i n d i cator that prpvid es quick diagnostics. Th i s gives inf orm ation r eg ard ing: ,.. s ignal l oss ..,. sender i nterfere nce ..,. receiver interference. Int e rferences can be locate d using the diffe rent blin k interval s. Detail e d d iagnost ies are available using a PC ( l aptop, e tc) em t h e data/ diagnostic i nterface. Software for th i s is a n option available from S ICK. I 5.: Mechanical and Eledricallnstallatlon . .. s.J General The FGS can be use d in '!'Y I environment (IP 65). R egu l at i on I safety distances as well as an a d equate protected height s hould be observ ed Special atten t i on shou l d b e given to t he danger o f reaching over. u n der, or standing behind sensing fie t d s to reach the poin t of operat i on. This must b e prevented mechanically or cover e d by ano th e r opt i c a l safety device (e.g. FGS da isy-chaini ng). 5.1.1 Safety Distance t o the Point of Operation The safety device requ ire s a certain m i n imum d istance b e tween t h e Resolution 14 mm 30 mm Apptoach spee d 2000 mml s (150 ms + 1 5 m s) 2000 mm/s (50 ms + 1S m s ) + 8 (14 11 mm) = +8 (3 0 -11 mm) 2000 m/s: 2000 mm/s S ms + 0 mm = 2000 mm/s 6 5 ms + 128 m m-= S< SOOmm 330 mm 258 mm App roa ch s p eed 1600 mmls (300 ms + 15 ms) 1600 mmls (0 150 s + 1 5 m s) + 8 (14 1< mm) = +8 ( 30 14 mm) = 1600 mls: 1600 mm/ s 31S ms + Omm'"" 1600m ml s 16S ms + 128 mm = S>5 0 0 m m 504 mm Tabl e t Example s of safety distance.s S. sensor and t h e point of oper ation T h i s e n sures that t h e poi nt of o perati o n ca n b e r eached only after the h azar d ous mot i on has b ee n st o pp e d (Fig. 10). T h e safety dis tan ce (acco rding to E N ns 999 a n d 294) is d e p en dent on t h e 551 mm ,.._ machine sto p t i m e I> r es po nse time o f t he safe ty device, ,.. resolution of the safety dev i ce, ,.. h a n d s pee d. Mi&d!e o f Depth o f Prote ctive f'".eld The machine stop time is a measurement o f t h e machine, hand speed i s 2m/sat up to 500 mm minimum safety distance; 1.6 mls if t h e minir:num safety d istance i s grea ter than 500 mm. mm Top of tool 'l'o M)(d tt>.e possibi lity o( SUnding bctwn the field t'Ft<: point of opetatfon this disunoo must be ma..inuinc.d a F i g 10. Cktcnnini n g machine S1opping time and s a f ety S afe t y d istance Sis calcula ted accord ing to the f ollowing formula: u p to S $. 500 mm : 5 =2000xT+ 8 (d -14) i f S>500mm S =1600xT+8(d 14 ) S = safety distance i n mm T = tot a l response time in rt:\S (m a ch i n e sto p time + FGS response ti me) d = reso l ution This r esults i n safety distances as shown i n Table 1.

PAGE 76

5.1.2 Distance to Reflective Surfaces 0 .Reflective surfaces that are present in the sensing fie l d of the sender and receiver can cause reflections that may result i n an o b ject not being detected (Fig. 11). T h e refore, a minimum distance (a) must be kept f r om the renect ive objects to the optical axi. s (straight-line connection FGSS/FGSE). The d i stance "a" is dependent on the distance between the send er and receiver unit, and on the a l ignment o f the d e v i ces (Fig. 12). 5.1.3 Multiple Safety Systems fig. 1 1 CorrectinstalbtiOt\,
PAGE 77

10 r----------------------------------------,' FGS with L-b'ackc.t 4' FCSwith djustabte brad:ct whh ll'lo
PAGE 78

S.3 Electrical Installation C onnecting the FGS tompon ';t:''t', ... 't ... using a 7-<0nduaor cable (fig. 15). The ''l.ll'! .;. .: ....... : ... h.'. ... ;. . -r: maximum wire cross seaion ts: t,,;..; :; ... ... ;, -''it.' tS mrn' w i t h out jocket; 1 min' wit h .. ., ... :: :--:. ...-.:,rt jacket. Both components have an internal termina l strip in the plugabl e s uppe r endcap. As an option, the end . cap can be equipped with a DIN . Cj\lkl<-d'JSCOnn ect receptade s E Both components a.-.; supplied wit h . . ,.-.... .. . .. .. 24 V DC ( 20 %). Power sup p lies ,7 . ) . ..... ' I ''i. ;. SO ms (fig. 16) 1 : .... 0< W l
PAGE 79

12 FGSE -----1 .. 3 I + 24 V I> i 2 ---'--t1Jo----C4'-fj...!1-, !max= O,S A 2 .. 1> 5! tRSas .. .. 6 I r (LCU/d1o,nostks) 2 -' ______ ..J Pf Fig. 17. Connectiol'! or FGSE Receiver. 5.3.1 Swlt
PAGE 80

6 FGS Safety Light Curtain Specifications 6 FilS Safety Light Curtain Specifications Protected He i ght 300 : . 1800 ,;,n, Scanning Range 0.3 ... 6 m I 0.3 .. 18 m Resol u ti o n 14mm/30mm Enclosure Rating I P 65 Supply Voliage Uv 2 4VDC% Max. 1\eslduall\ipple 5 v .. Power Consumption 0 9 ... t S A (noload) Synchronization opt ical, without separate synchronization channel Outputs 2 x PNP. 0.5 A, shon cirge U -2V v Response T ime :=:; 1 5 ms Connection Max. wire Cross-Section 1 mm2 with jad:Jr Humidity 15%-95% (not condens'1ng) Storage T emp eratu r e -25 ... +70 c Dimensions Length dependent u p on p rot ecte d height; Height see dimensional drawings. Housing Cross-Section 52mmx55mm . ll

PAGE 81

1 4 7 Dimensiona l Drawings 7 Dimensional Drawings PG Conn ector or ---.... Qukk Discon n ect (approx. 100 m m space teq uired for pluggi n g in) -M li)J < "' --l1S')L I 52 .. Op tiul Axis= Center of exi t window I -l21.$L u FGS 300 FGS 450 fG S 600 FGS 750 FGS 900 FGS 1050 FGS 1200 FGS 1350 FGS1500 FGS 1650 FGS 1800 Dimensions I n m m A 8 c D 300 384 224 80 450 534 374 80 600 684 80 750 835 675 80 900 985 82 5 80 1 0 50 1135 975 80 1200 1285 1125 80 1350 1435 1275 80 1500 1586 126 80 1650 1736 1576 -80 1800 1886 1 726 80

PAGE 82

' 8 Connection D iagrams 8.1 Conneetlon Diagram for Terminal Strip Fig. 8 show s the pin c o nnections for the FGS t e rminal strip: l eft, sender; r\g ht, rece iver 8.2 Connection of DIN Connector Rg. 19 shows the conneaioos for the 01 N connector. 8 C onnection Diagrams . .. . s E litttfitin liftiDiln 1234$47 1 2 3 4 s 6 7 o,_,_ & ...... ... :2:2 "" FG S FGS Fig. 18. Pin ---;... ) __ ) -__ l + Ag. 19. Pin
PAGE 83

9 Selection of an. Opto-Eiectroni<: Safety Device 9 Selettlon of :>n Opto-Eiettronlt S:>lety Device In the selection of opto-electronic safety dev i ce proceed as follows: Regulat.ions! 1 Look up the regul ations and standards applicabl e to that particu l a r application. I local authorities and profess ional ] organizations assist here 16 i Operating Range! j Determination or required operating range. The operating range corresponds w i th the width of t he area to be safeguarded. T he operating range s hou l d be chosen so that danger po ints accessible only through the protective field. Corner Mirror? Using a come r mirror offers t he possibility to safeguard two or t hree sides. Us ing corner mir rors the maximum scann ing range 0 5 m p e r mirror. Protected Height! Determination of required protected height The protected height must be chosen so that danger points are accessible o nly t h rough the prot ective field. ] 9.1 FGS Selection Table \ Modot Roso1"tlon Part No. I FCSs 300-11 11 1 012 ;oo -21 30 1 012 600 fGS E 300.111 14 101 2 SOl I -21 >() 1 012 601 FGSS <450-11 1'1 t 012 502 21 30 1 012602 ; Sl.2 Corner Mirror Selection Table Type of Mifrof" Range Protective field PNS 80 3m PNS 120 6m PNS 200 12m Vvhen ordering. p1tast indkat e protected height. _____ rFGs-s 14 1012 50,. 9.3 Accessory selection 30 1 : table FGS 600.11 1'1 1 012 SO:S i 30 1 O tl60S 1 fGSS 750t1 11 1 012 506 -21 30 1 012 606 FGS E 75011 1'1 1 012 507 30 1 012 607 f"GSS 90011 11 \ 012 50S t 1 30 1 0 12600 Part No. Bn<:ket,lixed ) 7 021 352 Mounting. Bracket. adjunllb1e 7 021 351 M o u n ting, 8 ra
PAGE 84

10 Weight of Systems 10 Wolght of Systems The values in the chart. below are 1 we i gh t s per system (Sender plus Receiver). Each component covel'S 1 approx. SO o/o of the total w e ight. The r e is no difference between 14 mm and 30 mm versions. I l ..... -.. ... 1gb Wt13M/syrtem I [mmJ I I 300 2,8 450 4,0 I 600 5,0 750 6,0 900 7,0 I 1050 8,0 J 1200 9,2 1350 10,0 1500 11,2 1800 13,4 \ I I I I I I I 1 7

PAGE 85

I I Instruction Label, Checks 18. II Instruction Label, Checks The Instruction Labe l (Fig. 20) must be placed a.'ter mounting and adjustment of the FGS (se l f adhes ive). It m u st be easily ac ce sSibl e a n d readable An ideal position i s the machine body Proceeding t h e cycli<: checks allows to d iscover manipulation and not permitted modifications. n ear t h e sender or receiver. FGS instruction plate and comeclrg liP '*"".l f'.l'f 11)C':IOIC:I s !OTI lOl'ltl't ttOt a.tt "ot t ) M l tot lhn CA'Oietftcn ol fla il cWoX t ttlwttn lr<)n9; .. !!er O"d lt(e-oh:llf l>.tl"'} 1M ()fO{fWt Cflt, ttd l o1p IJJ!V"Cll'c II O't'fot t:T ytkw $9\0'! let'(llq't$ \0 Cl tven. tnt It-En lt'4 nxl'lllt lo"';.CJ to: u sco l hrt o:ta tt poltetto e frlt SQJ t 6:1tor.c:t -..e;l bt 1'1 J.'le Oi'IC(S ()ft)'f()e(l OJ' lf'lt taQCIItlt FGS 14 mm resolution with LCU-P All enllcrl: If ...t..'l lt().(t(l t ( -;t'ul(l"( lilt 041 o""q; 01: n ectO"clo"
PAGE 86

1 1 Rating Label fore the e lectrical connection. I Be ch su eck the equation of the local pp l y vokage wit h the data on the Ra ti n g label (Fig. 21 ) -::: f """' .. ( ... .,.... .. ' I I l. ,, ....... ........ I w... ...... _. ..._,.._ UVi20'If>l ... .. 1 .....,, __ -:::1 I ....... _[ \.,,..., ... _.. w l ,....... .. :::.:..:: I """ ... ._ ......... I ::-=:::: ( o ... +Ssc I ..... ,._ .... """"'' I r. ....... "-lllo> l""' ( TYl'E 4 I '-""""'"''-[ """''"' IP$5 I tw .... ,,., 1 J .. [ I .. Fi& 21. Rating Lab
PAGE 87

APPENDIX C Manufacturer Specifications for IDRlS Evaluation of Automatic Vehicle Classification Systems 45

PAGE 88

0 0 Idris Count Accuracy Congestion free Flow No. Test Site Lanes No. of No. of Veh icles Error Vehldcs Error :.-t-.. ..... ; 2 N/A N/A >10,000 0 i North Wales t ;)fr",.tt:,fi { t 2 if;. .. ster ''>' . . ... , x ..... N/A WA >20, 000 0 . 25 f: 4 :WeSt L.OnetOO , 23,043 3 $1 591 2 Compari.$On or Actua l Accuracy Against Spceilication SpcclficaUon Actual :'" N iul< .;r: ";,.' !" 1 \; 1 i n 500 0 .2% 3 1 n23,000 ... ConQestloh .?_:. 0.013 .... :, ,';;;(;, 1 in 10,00 0 . :-( \\' .. ..... 0.01% 2 in 51, 500 0.004% Data Enhancement Techniques A breakth r ough which enhances vehicle detector output has been deve l oped by WS Atkins and Diamond CQnsulting Services The enhancement tech n iques are embodied In severa l algo ri thms des i gned lor use in roadside vehicle detection equipmen t such as outstations lor incident d eteclion syst ems, classifiers, census/survey apparatus and outstations for tolli ng. T h is t e chno logy has bean licensed to Peek T raffic, world leader in detection techno logy. s upplies a DBFO outstation which is capab le of meeting th e UK H i ghways Ag ency's fu ll specUicati on for shadow t olling equipment by comb i ning all the know n advantages of the ind uctive l oop detector with the enhancement t ech n i ques I ni ti ally tested with a standard induct ive loop lay out, two loops per l a n e the tech niq ues are ap p licable for any number of lanes. Th i s add iti on t o the ldr i s family enab les Indiv i dual identification of all vehicles, i rrespective of their position across the roadway o r rel ative to othe r vehicles Th is results I n the fo llowin g features. The vehic l e count acc u racy has only 1 in 10,000 (0.01%) error i n f r ee-flowi n g traff i c a n d 1 I n 500 (.2%) i n measured over any time period Vehicle speed is measu r ed accurately and consis t ently l or a ll speeds, regardless of t h e vehicle's lateral position I n rela t ion t o the se n sor site. Th e accuracy is reduced when only one loop o f the pair is operational. Vehic le length is measured accurately and consistently at all sp e eds even In congestion and regardl ess of the vehic l e's lateral posit i on in relation to the sensor site The accuracy of vehicle speed and leng t h measureme nt s Is not res t ricted by previous limitations imposed by t h e scan ra te of the detector T a ilgatin g vehicles can be Ide ntif ied a n d d iscr iminated from towing vehicles i n a// conditions, Including. congestion. Two similar vehicles in adja cent l anes are differentiated from one vehicle straddling two lanes The cou n t accuracy Is Improved by disc r imination between signal overs plll into th e adjacen t lane and a real vehicle alongside.

PAGE 89

Data Enhancement Techniques These techniques repr esen t a leap lorward i n dete<:tion accuracy with minima l change requir ed to existing tried and tested equipment and infrastructure thus capitalizing on all the benefi t s of known cost-effective detection sensors and I nfrastructur e. Applications I nclude: Inci den t Detection by improv ing the quality of input data to such systems a t the ou t station le ve l, the benefit Is saen In the reduction ot false a larms at the instation. As ve h ic l e data collec t i o n is I ndependent of ve h i cle speed and position, i n cident detection in congestion is easier to I denti fy. Tr affic Management the renge of data available has bee n extende d so managers have mo r e accurate Informatio n on whi c h t o base thei r planned maintenance ac tivi t i es, incident and emergency responses, a n d day to day activities. C ensus and Survey the reliability and ra ng e o f traffic statistics is incre as e d, enabl ing more complex combina t ions of data u sin g ti m e of day, vehicle c l assificat ion speed and lane usage to provide the complete picture of r oad usage for monitoring ano:! p l a n ning purposes. Vehicle Classification the abi li ty to measure veh i cle length I ndependently o f vehicle position and s peed, combined WTth the vast i m p r ovement In oonslstency and accuracy, enables use ol narr owe r b ands to give more detaile d categorization for tolling and census information. Tollingthe n ew U n iled Kingdom DBFO (Design, .Build, F i nance a n d Operate) r oad scheme s require solutions to some dlfficult p roblems in terms of performance. acct(racy a nd re lia bility o l shadow tolli n g equipment. Each d e tected vehicle cont r ibutes revenue according t o its class, and so vehicle data c o llection is subject to a very demanding specification which can b e m e t now. The extremely high accuracy allows improved auditing o f toll road usage enha n c i ng r even ues. A patent for these de tector enhancement t echniques Is pend i ng ldris i s l h e registered trade mar k o f WS Atkins Consultants Umited Fo r more information or to d i scuss your application require ments, contact Jan Cardozo or Peter Keen at Peek Traffic Sarasota. M25 Junction 15 to 16 Site Ref 4985A" Stop-Start Traffic !Lane 4 Lane 2 lane! '4870. f . 4875-. 4aari : ; . .. in' \ I I 4890 . 4 89 5 ,. . 4900 .. ... ,.. :. .:;;,' . ,: : PEEK. Peek Traffic, Inc. 1 500 N Washington Blvd. Sarasota, FL "34236

PAGE 90

: General Information Sheet: IDlUS is a: roadside automatic incidettt detection system. iJ Automatic incident detection by the very nature of its is an invaluable aid to improved road safery. -; IDRIS has been designed to be an integral part of an Advanced Transport Telematics system. It was originally designed, and is ideal, for restricted road conditions where it is important to detect hazardous situations promptly so that remedial action may be taken. The hazards that can be detected are: single stopped vehicle anywhere in the carriageway queuing traffic slow moving traf!ic vehicles travelling in wrong direction slow moving vehicle in fast traffi c The IDRIS system comprises inductive loops and usociated outstation equipment. network manager(s) and instation(s). It includes a combination of algorithms: one being located at the outstations improving the input data quality, and five at the instation collating and analysing all the data to give a complete picture of the traffic situation. Of the five instation algorithms, four run c<;>ntinuously while the tif\h, for the single stopped vehicle, is dependent on traffic flow rate. The system is based o n two loop s p er lane each 2 metres square and 2 met res apart at approximateiy I 00 metre spacing. T hese measur ements are nominal and a sit e spacing of 200m is achievable. It is hoped to increase this to 250m. The system is designed to be used with any suitable detector (cUJTent systems use de tect ors ) An installed system requires sufficient outstations at the chosen spacing on one carriageway, connected t o a network manager. A network manager is required for each carriageway ;md is capable of handling in exoess of20 outstations (more than 2 km of road). &ch outstation consists of sufficient loop detectors, a f our channel detector, an outstation processor and a communications interiilce in order to provide raw traffic data to the instation via the network manager. The outstations are synchronised to I ms :1: I ms. A detailed algorithm processes \he output of a standard loop detector to give a count accuracy of0.05% (i .e an error of l in 2000). 'll1e instation comprises communications interfaces together with a standar d processor containing sot'tware for the a l gorithm s and associated traffic statistics The IDRIS algorithms use a data packet for each vehic l e containing information d escribing the vehicl e s location at the time of acquisition (i.e which lane at which outstation), its direction, speed, length, clearance from the preceding vehicle, duration over the outstation site and any relevant time out information. From this information, IDRIS detects the vari ous hazardous conditions detailed above IDRIS is modular in sttuaure enab ling easy use of several different algorithms (e.g McMasters, HIOCC, Algorithm 8, as well as any or all of the six IDRIS algorithms) the specific site and performance demands ... ......,,f" lo + ' I o o o ,..,,..., ..o -. '"

PAGE 91

Techni cal Inform ation Sheet I IDRIS is a based road traffic incident dettion system, designed to det ect a number of anomalous traffi c conditions on a two-lan e dual caniageway. IDRIS is designed to detect the following conditions : a si n g le stopped vehicle anywhere in the caniagewa y at low flow rates queuing traffic general l y slow-moving traffic a slow-movi ng vehicle in faster traffic vehicles t ravelling in the wrong direction ' IDRIS consists of outstation sensors. outstations. a communications network and an inst:ation in the configuration shown in the system schematic below Outstatioo. 1 lORIS Ptoe=t Local Terminal IDIUS syJtem schematic System specification Maximum number of outstations per network manager Communications cable Maximum number of cab le pairs Maximum cab le distance between adjacent outstations Adjacent outstation time synchronisation Ou!Statio n count accuracy Outstation sensors Loop dimensions Distance between loop pair in direction of travel Distance between loop sites Loop installation standard Unks to other systems 20 Standard twisted pair. 0.9 !M' 3 1000 m lmslms than 0.05% Inductive road loops Approximately 2 m square 2 to2.5 m 80 to 200m UK Department of Transpon's Specification for Highway Works Subset ofccm X3. 28link to supervisory system Unk to CCIV system (optional)

PAGE 92

. I drYs ' Sheet 2 > : I .. ;,. i.$, f.;. ._, . . The IDRIS lnsta t ion }'?} . The instation COnSists of two parts, the network manager and the IDRIS processor. Nonnally these two units are co located in a service building local to ihe monitored site. However, the net\i>'ork manager is designed for a roadside envirorunent and can be located a considerable distance from the IDRIS processor . The networl< manager has been designed to meet the UK Depattment of Transport envirorunei)tal and EMC specificalions for roadside motorway communications equipment r .--.-----"':'-------------.--. -I I l I I I I I I I 1 To Supervisory Systea>. llS485 to nr.t outstatioo 1 I Network llS232 Manager Comms Algorithm X3.28 1 1 1 I 1 I 1 1 1 1-----i-> i..o<;al 1 Tennioal: 1 To Remot e Terminal (If required) 1 JDRJS pro
PAGE 93

r-- Id @) l ]e,c;linical lnforiruition Sheet 3 : t ( The IDRIS '. '' . An outstation consists ofia loop detector an-outstatio n communieations board, a display board and a mains power supply unit, 'all in a standard metal C:ase. T h e ou t station is designed to be installed in a roadside cabinet adja cent t o the loop site, although it can be installed up to 200 metres' cable length away, proViding Suitable loop feeder cable is used The ou t sta t ion has been designed to meet the UK Department of Transport environmental and EMC specificati ons for ro a dside motorway communicat ions equipment. -----.---------. -:--' I RS4SSfrom previous outswion I network manager I I I I I PSU I I :-1 I I H I I . Comms I Detector I Board Processor I I I ==l I Board ; l I -' Display board I I I Loops i RS4SS_ to next outstation I j --. ----------.-. ":"'------Outstation schematic Outstation sp e cification Dimensions S upply voltage Maximum supply current Supply frequency Operating temperature range Communications interf ace Lightnin g and transien t protection for t he l oo p s Lightning and transient pro t ection for the RS485 Loop inductance range (inc feeder cables) Loop osciUator range Maximum number ofloop s Maximum dis tance from l oops to outstation Maximum distance between outstations . 180 mm (h), 240 nun (w) 110 mm (d) 190 to 250 V ac (110 V a c available) 200mA 45 to63 Hz -1sc to +1oc RS485 2 wire active Z ener diode and flash over protection Transzorb 3 7 to 1 200 1M automaticall y tuned 65 to 87 kHz selectable in one of four ranges 4 300 m cable length (preferabl y within 200 m) 1000 m cable length

PAGE 94

. / Technical Information Sheet 4 An IDRIS Application "'= ... ...... . ..... Typically IDRlS is installed as part of lll integiated traffic management and operations scheme. 1in addition to IDRlS such a scheme may consist of variable message signing. lane eonuol signals, traffic swveillance using CCTV, envirorunental monitoring. emergency telephones and a supervisory system t o monitor and control all the sub-systems. Within the tra.flic management scheme IDRlS interfaces directly to the supervisory system which, in many cases, is based on a Supervisory, Control and Data Aquisition (SCAD A) package. Upon detecting an incident IDRIS notifies th e SCADA system, which in tum: I aleru the operator commands the CCTV system to switch the relevant cam
PAGE 95

. ' : Technical Sheet 5 .... ''i..: IDRI S ap)?lied to a 900 metre root) tunnel The example below shows how IDRIS could to a real situation. !Acation -a 900 metre two-lane singlt>-carriageway tunnel with no hard shoulder. Design constraints sensible trade-off between inter-site distance versus location ofinci\)en i loeal service building (SB) standard twisted pairs available along the carriageway pan, lilt and zoom ccrv cameras The tunnel schematic shov.n a ddresses the issues described above and allows t he portal areas t o be monitored as well. 2 3 s ' I 9 10 II 11 IJ -010 DIO 00 00 DO DO 00 DO DO 00 00 DO 0 t..m. D'O o;o 00 00 00 OD DO DO OD 00 DO 00 00 1-< lo t ,J. I lOOm : 900
PAGE 96

\ Information Sheet 6 .. IDRIS applied toia 1200 metre (4000 foot) confi ned road' . The example below sho.ws how IDRIS could be applied to a real situation. Locationa 1200 metre two-lane carriagewa y with both lanes running in one direction Design constraints sensible trade-off between inter-site distance v ersus location of incident local service building (SB) stan dard twisted pairs" available along the carriageway pan, tilt and zoom CCTV The carriageway schematic shown addresses the issues described above. I l ' s 7 I 10 -+-010 010 00 00 00 00 00 00 0 0 00 OiO OjO 00 00 00 00 00 o o 00 00 ..... lOCo ; ""'' Carriageway schematic II oo 0 0 u 00 00 I) 00 00 $1! The system schematic for the above carriagew&y. including remote access, is shown below. IDRIS _, RS2321mk . System schematic ..

PAGE 97

r TRAffiC nocm< funckd b) lOllil'll is eotirdy
PAGE 98

TRAffiC TECHN . : .,. .... .... ; ...... : .. .. ._.:_:_ .. .:.:.1:_.:..: .. ...... .: . 'Case . 8eha\'IOUr . i Re:;ott .:. i . : ,' .. : _ , ...... ...................................... .. t.:--:::..-:--.:-:;----.,..--.... ::.. .... ... 1 A "'hicle $tl'8ddles lhe lane 1 The vehicle Is not det.eete ................. r ........................... ........... ............ -......... r ............................ ................ -....... ...... ... .......... .. ...... ... r .. ... ............. ................... .. --.-----.. 5 i Two W:h'cSes i n a4jaecn t J ane$ 1 All fc>ur loops activate s.imu\taneously' ; Normal t.rafne panems. i nde pe ndent etOS$ their sites tog e ther 1 of site so m e vchide.s. who C3 n prove that chis is not inc\itabte should COOt3Ct t he authors whasc intcre1c is gu:u3ntee'-'0 may be 2 ph:ancom ehide, the o u m., c iol'l <::2n supply high qua l ity ro :tny system which r-equire$ it. ThiJ could be a hish-levd dac; collcion system for shadow tolling, or an i ncident detection Sf$ttm $Uch :lS lORIS or or-as of a c:.ompcehcnSive lt3ff.c management $Cherrn:. 'With rhe lORIS incident dettwe tion sysccm. this labdling is used by the main algorithms at the insu.tion as part of the identification of :1 single stopped vehicle. [n considering sevcnl consecutive and ttaddng pactieular vehicles '7he lORIS outstation algorithm contains ab.out half a dozen tests which consider the detector activation output, with each test addressing a different criterium, for example comparisons of detected vehicle speed and length" both answers being weighted. The wcishcins$ for a cw may vsry on the: o rder of testing o r if t he input conditiofts have p3ccieular ch2r3ctetis tics and nor aU testS ate used i n 3l1,drcum down che road, lORlS can soon con6rro o r coccea whether use four or case five occurred. chucby co n tributing co the system's low false alarm of one or '"""' .......

PAGE 99

Peek Sarasota MTS38Z 4 Channel Detectors Treadle IIP AVI Other Am tech I/Ps LANE COMPUTER Seria l Driver QNX4 ldris Detection Algorithm Amtech's Lane Processing Algorithm


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader ntm 22 Ka 4500
controlfield tag 008 s flunnn| ||||ineng
datafield ind1 8 ind2 024
subfield code a C01-00191
040
FHM
049
FHmm
0 245
Interim evaluation report : task order number 1 : Automatic Vehicle Classification (AVC) systems
260
Tampa, Fla
b Center for Urban Transportation Research (CUTR)
c 1997 July
650
Traffic surveys--Equipment and supplies
Intelligent Vehicle Highway Systems--Florida
7
Automatic Vehicle Classification (AVC) (TRIS)
2 TRT
710
University of South Florida. Center for Urban Transportation Research
720
Pietrzyk, Michael C.
1 773
t Center for Urban Transportation Research Publications [USF].
4 856
u http://digital.lib.usf.edu/?c1.191