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 nam 2200457Ia 4500
controlfield tag 001 028743217
006 m d
007 cr cn|||||||||
008 110816s2002 flu obt 000 0 eng d
datafield ind1 8 ind2 024
subfield code a C01-00219
Effectiveness of bus signal priority
h [electronic resource] /
Shireen Chada, Robert Newland.
Effectiveness of BSP
Tampa, Fla. :
b National Center for Transit Research, Center for Urban Transportation Research, University of South Florida,
1 online resource (v, 33 leaves).
Title from e-book t.p. (viewed Aug. 16, 2011).
Performed for the U.S. Dept. of Transportation Research and Special Programs Administration and Florida Dept. of Transportation under contract no.
Includes bibliographical references (leaves 31-33).
Traffic signs and signals
x Technological innovations
Traffic signal preemption.
Dept. of Transportation.
Research and Special Programs Administration.
Dept. of Transportation.
National Center for Transit Research (U.S.)
University of South Florida.
Center for Urban Transportation Research.
i Print version:
t Effectiveness of bus signal priority.
d Tampa, Fla. : National Center for Transit Research, Center for Urban Transportation Research, University of South Florida ; [Springfield, VA : Available through the National Technical Information Service, 2002]
Center for Urban Transportation Research Publications [USF].
y USF ONLINE ACCESS
USF ONLINE ACCESS
TECHN ICAl R EPORT STANDARD TITlE PAGE 1. Report No. 2 Govem men t Accesslo n No. 3 Reci pienrs Cat a log N o NCTR-416-04 4. T itl a and Subtitl& 5. Report Date Effectiveness of Bus S i gnal Prio ray January 2002 F i nal Report 6. Performing Organization Code 7 Aut h ot(s) 8 P erfomti n g O rgan izatio n Report No. Chada, Shireen and Newland, Robert 9. Perfo
Effectiveness of Bus Signal Priority Prepared by: Shireen Chada Research Associate Robert Newland Student Assistant National Center for Transit Research Center for Urban T ransportation Research College of Engineeri ng University of South F lorida 4202 E. Fowler Avenue CUT 100 Tampa, Flor ida 33620-3120 813.974.3120 h t!fl://www. nctr. u sf.edu/ Project Manager Gene Glotzbach, F DOT Contract Numbe r BC 1371 6 Florida Department of Tra nsportation 605 Suwannee Street Ta ll ahassee, Florida 32399-0450 850.414.4500 Researc h and Special Program Administ rati on U.S. Depa rtment of Transportation RSPNDIR-1, Room 8417 400 7' Street, SW Washington DC 20590 http://www.rspa.do tgov/ The opinions, findings, and recommendations expressed i n this p ubl icat ion are those of th e authors and not necessarily those of the Flor ida Department of Transportation or the U.S. Department of Transportat ion. The document was prepared in cooperation with the State of F lorida Department of Transportation and the U.S. Department of T ransportation. m
Table of Contents EXECUTI\IE SUMJvlARY ........ . ............... ....... ......................... ....................................... 1 I NTRODUCT ION ...... . ........ ... .................... .. ............ . ...... ........ ............. .. .................... 2 TERMINOLOGY ........................ ..................... ........ ............ ............ .. .......... . ...... ........ 2 EARLY \1\fORK ...... ... . ... ................ .. ........................................ .................................. 3 BSP STRATE GIES ........ .. .. . .. .............. ................... ........... ............ ................................. 4 BSP SYSTEMS ..... ....................................................... ............................ .... . .. ........... 6 PROS/CONS OF PRIORITY M E T HODS ..... ............ ..... ... .......... .......... . ............... 7 PROJECTS IN NORTH AMERICA ........ .............. ............................................................. 1 0 WHE N I S NON-TRANSIT TRAFFIC lEAST AFFECTED? .................................................. .. . 16 CHARACTERISTICS OF AN IDEAL BSP SYSTEM .... ......... ........................................... .... .. .. 17 PROFESSIONAL OPINIONS ON BSP ... .... ............. ...................... ........ .......... . ... ........... 18 SIGNIFICANT ISSUES OF BSP ......... ......... .......... ... ................................................ ... .... 22 IM PLEMENT I N G BSP .. ................ .. .................. .................... ... .................................... 24 CONCLUSI O N S .... .. . . . .. ... .... .. .... ... ... ............ ... ........ ........... . .... .................. ............ 29 REFERENCES ...... ....... . .. .................... .......................... . .......... ............. ........ ........... 31 APPENDIX A (BSP SPREADSHEET-SEE ATTACHED DISK) .... .. ........... .... ................ ....... A-1
TABlE 1 TABl E 2 TAB l E 3 TA8 L E 4 TABlE 5 TABl E 6 List of Tables SUMMARY OF BSP .... ........... . .............. ............................... 6 Pros and Cons ..... ......................... ... ...... .................. ... . ....... ... ........ . 9 PROFESSIONAl OPIN I O N S ..... ....... . .... . .... ..... ......... ........ . . . ........ 18 PREIMP L E MENTATION CHECKLIST P O INT SYSTEM .. .. ................ ... .. ...... 27 RECOiv1MEN DAl'IONS BASED O N SATURATION LEVEL . ........... ... ........... ... 27 OP ERATIO N A l AND DESIGN GUIDELINES ........ . .. ...... ............ ............... 26 v
EXECUTIVE SUMMARY E f fect iveness of Bus S i gnal Prio ri ty (BSP) s t udy was funded by the Nati ona l Center for Transit R e s earch (NCTR) at the Center for U rban T r anspo r tation Research (CUTR). This study eval uates BSP's impac t on traffic operations. The goal was to examine how d ifferent situations, such as the l eve l of conges ti on, p l ace ment o f bus s t ops p resen ce of e xpress bu s serv ice, and n umber o f transi t veh i c les on r oute r equ i r e differ en t t ech n iques o f BSP such as realtime or fixed-time based control. T hose t e chn iques a ls o utilize a var i et y of d i fferent cont rol s tr a t egies such as p ha s e suppression, s ync h r onization, compensat i on, an d green r ecall. In o rder to guide tr ansportat ion agencies in this d ecis ion-making p r o cess, a s p readsheet wa s deve l oped to h elp a n ag ency dete r m i ne i f s i g n a l p rior i t y i s viab le in a specif ic co r ridor A set o f guidel i ne s was then developed to d e t e r m i n e t he mo s t b en e f i cia l met hod of p r iority to pursue To establis h how BSP can b e most ef f ect i v e t en t r ans i t professionals (p lan ners and e n g i n eers) invol v ed in install i ng B S P p r ojects we r e intervi e wed. T r affic e ng i neer s i nterv iew ed stated that agencies must develop control str ateg ies that are app ropri a t e for th e g i ven i n t ersections i n their a rea. T ransi t p l anners i n terv i ewed for t h i s st udy exp lained t h at BSP d i d not c reate a n y s ig n i fi cant delays to non -trans i t vehicles A lit e r a tur e sear ch also u ncove red stat ements on t h e sig nifican t issues of BSP, s uc h as co n sidering the vo l u me of bus passengers, degree of existi n g congest i on l eve l and del ay t o non -trans i t vehicles. The l i tera ture s ea r ch reveale d the benefi t s of direct priority s tr ategy in a reas w i t h low congest ion a n d an i ndirect p ri o r ity str ategy for areas with a h igh d egre e of congestion. Agencies util iz i n g BSP around the count r y we r e stud i ed to determ i ne w h ich t y pes of p rior i ty a r e used i n ce rt ain con d iti ons and what technol ogy i s be i n g i mplemen t ed. It was de term ined that an e f fective bus p r i o r ity syst em m us t ensure transit vehicles i n all d i rections can be assisted w i t hout excess d elay t o non-transi t v ehicles BSP offers fewer benef its i n areas with extremely high bus volumes or very light t r af fic Th e most approp ri ate p r i ority method m ay be one that co mbi nes the var iou s e l e m e nts of existing pri ori t y t echn iques. Realtime con t rol strategy offe r s the greatest bene fi ts, a l thoug h any system must be de signe d w i t h the pa rticula r area needs in mind. A s a result o f the i nformat ion gathered, a fra mewor k for an idea l bus priority system was developed Based upo n t hat inform ation t h e pre-Imp l ementation Ch ecklist was designed. The checkl ist focuses on the mos t cr i tica l f actors in BSP and reco m mends pursui ng BSP if a n area has e n oug h characterist ics i n p l ace t o m a k e BSP e f fective The operational and Desi gn Guidelines for BSP assist an agency in choos i n g t h e mos t a p p rop ri ate BSP m e thod t hat complements the area characteristics. I
I NTRO D U CTI O N BSP i s be i ng more widel y deployed i n N orth Ameri ca to address traff i c congest i on, caused by t raffic s i gnals1 ior on street transit serv i ce Sig n a l prio r ity for t r a n sit veh i cles is expected to improve t r ans i t operat i ons and service qualit y. I n theory an impr ovement i n t r ansit performance p rovides add itiona l i ncentive fo r peop l e to swi t ch modes and reduce t r affic congest i on. This has been an issue of de b a t e, both among t r affic e n g i neers and transit planners. The traffic engineers r e s istance t o impl ementation has ofte n been based on a conce r n that ove r all traffic per for m ance may be undul y co m p rom ised when s i gnal timin g i ntended to opti m ize t r a f fic flow i s overridden t o provide a t r avel advantage to transit vehicles For purposes of t hi s report transi t signa l p rior i ty (TSP) will be r eferred to as bus signal pri ority (BSP) This researc h project evaluates how traffic operations are a f fected by bus s i gnal priori ty. The effectiveness of bus signa l pri o r ity o n transi t veh i cles a n d the i mp act on t raffic ope r ations a r e expl o r ed Fundin g for this st udy was provid ed b y the Nationa l Center for T r a n sit Rese a r ch (NCTR) at Cente r for Urban Tran sportat i o n Research ( CUTR) This study also addressed all the i m plicatio n s of BSP t hus g i ving a more objectiv e analysi s of the advantages and d isadvan tag es. I t attemp t ed to prov i de mi ssing information to trans i t planners and t raf fic engineers across t h e nation as t hey b e come more i nvolved in the dep l oyment of BSP. T ERMINOLOG Y T e rminol ogy i n th i s r eport i s based in part on the Sunkari et a l ( 22 ) stud y on models to evalua t e bus priority. )> Unconditi onal priority: P ri o r i ty is given when ever a bus detector requests it from s i gna l s )> Conditi ona l priority: I ncludes variables t h at may l i m i t p riority given, such as b u s occu p ancy queue l ength, and time since l ast priori ty was granted. )> Phase: Part o f the t raffic signa l time cycle allocated to any comb inat ion of traff i c movements r eceiving r ight of-way s i mu l taneously during o n e or more intervals. > Green extens i o n : Green phase i s e x tended w h en bus is n earby to allow i t to go t h rough. > Special phase: A spec i a l g r een phase is in j ected into the normal phase sequence whi le all othe r phases a r e stopped. 2
)> t runcation: I f the bu s arr ives at an intersection durin g th e beginning or m iddl e of a r e d p hase, t h e red phase i s t r uncated and green phase is in j ected to allow the bus t o go through )> Compensation: G reen tim e i s alloca t ed to a non-prio r ity p ha s e that was t r uncat e d to make up for l ost time. )> E ar ly start: vVhen bus arriv e s at i nte rsection during a r ed g reen follows quicker than usual. )> Gree n r ecall: A green phase is disp l ayed each cycle whether demand exis t s o r n o t. l> Synchronization: Ti min g groups of traf fic signals along a n arteria l t o provide fo r th e smo o th movem ent of traff i c with minimal stops. )> Phase s up p ression: One or mor e non -pri ority phases with low d emand m ay be omi t ted fro m the n o r mal phase s equ e nce. EARLY WORK The deve l opment of Intelligen t Transp ortation System (ITS) technology o pe ned up new possi b ilities for bus sig nal p r ior ity i nvolving real-ti me t rack ing Th e theory of bus p rior ity was estab l ished with experime n t s t hat took place as early as 1 962 Many studies and th eor ies were establis he d b ef ore th e 1 990' s t ha t l a id t he groundwork f o r the bus p r i o ri ty of t oda y. Some of the earl y work is outl ined h e re t o give an idea o n how BSP evo l ved over t h e l ast few decades In 19i8 Vincen t et a l used a Sus P riority Assessment Simulation (BUSPAS) t o test preemption con trol strategies at an i ntersect i on. They found that g r een extens i on on ly' was o f littl e ben efi t to buses and caused very m i nor delays to other traffic Cont r ast i ng l y, a green ext e nsion, red trunc.ation no compensation control method wa s of great benefit to buses, yet caused l arge d e lay s t o other tra ffic. A green extension, red t r uncation, com pensation control method gave buses som e be n efi t without the l ong delay s for other traffic. In 1981 j acobson and Sheffo (12) sho we d how bus p rior i ty i s made more effect ive i f one ca n c hange t he under l ying signal setti ng. For exam pl e, t he be n efits o f bus priority ane s mall when traf fic in a preem p tion direction i s much higher tha n c ross traf fi c flow A system compl e x enough to handle such issues is on l y now possib l e with th e em ergence of ITS technology An early bus priority plan was th e NW 7'' Ave nue Bus P riority System i n Miami, Flo rida (16). The testing r an f rom August 1 974 to March 19 76 Thi rty-seven traffic sig nals were equ i pped with optical s igna l p reempt i on equipment. The syst em used an i n tense stroboscopic I igh t t h a t could be 3
picked up by t h e rece i ver a t the i ntersection f rom approximat ely 1800 f ee t. Once a b u s was detected the signal was either ex tend ed until t h e bus came through (wi th obvi ous adj ustmen t s made i f th e bus was del ayed) or t he curren t s ignal phase was adjusted to i nclude a green phase for t h e bus to go t hrough T he r e was no stat ist ically s ig n i ficant in c rease in bus acc i dents and t otal accident rates of all veh i cles on N W 7 Ave nue ac t ually d ecreased N one of the pote nti a l safety concerns (two buses p reempt i ng t he same signa l from different directions, bus operators expecting a gua r anteed green, auto and pede str ia n confusion, automobile s clus t e ring a r ound the bus to get a green signa l ) m ate rial i zed du r ing t he t esting phase ('16). BSP STRATEGIES Certai n signa l systems and programs t end to fo r m t he bac k bone of m any BSP systems (fo r example many are based on t h e BUS-TRANSYT opti m izin g program ) whi l e many a reas choose to put t ogether an o r ig i nal BSP syst em with compl ete l y diff erent hardware, software and purposes I n som e cases an agency will c h oose a vendor t o i n stall an ent ir e system; sometimes they will purchase hardwar e fro m differen t supp l iers a n d merely use a vendor's software system; and sometimes an a rea custom i z e s t he c ontrol t hemselves using a variety of different h a r dware sources. Consequent l y t h ere a r e a l arge nu m ber of different system s i n existe nce that are enti rely unique to that a rea. Only some of the more pop ular system ven dors are o utli ned in t h is sect i o n To understand the var i ous systems mentioned in th i s secti on, t he basi c p riority concepts a n d t h e strategies that defi ne and diff erent i ate these co ncepts are outlined be low. Conrrol Suategies T he fou r kinds of cont r o l str a t eg ies for awarding BSP are: )> Realti me: Rel y o n cons t antly updated info r mation to m ake decis i ons regardi ng p r iori ty. A real-ti m e signa l cont rol mod el is more fle xible to changing conditions, hence is generally more eff ect i ve l> Fixedt i me: Fixed-t i me control applies a s i gna l contro l plan based on t he average cond itions of an a rea. Fixed -ti me control does not receive constant l y u pda t ed i n form a ti on; the best co n tro l scheme i s applied to the a r ea r egardless of actual cond i tions. 4
)> Schedule-based: Prior ity is awa rded based on t h e bu s schedu l e. I f t h e bus i s runn ing lat e then i t rece ives p riori t y th rough intersection s Schedule-based co ntro l is mor e e ffe ct i ve at reduc ing b us trave l times. Since schedule-based control does not need info r mat ion o f bus l ocat i ons i t r equires les s communication equipme n t. w h ich m ake s i t more c ost e ffectiv e. > Headway-based: Priority is awarded based o n t h e h eadway betwee n buses Buses avoid bunch ing up with oth e r buses i n t his co n tro l str a t egy. He adway-based control is more eff ect ive at red uci n g wa i t times On occas ion, t he term s dire ct or indir ect priority are used. D i rect pri ority p rovides solu t i ons fo r a partic u l a r bus when i t has reached o r is very c lose to an i ntersecti on. I nd i rect p riority l ooks to clear up congested inte rsec tions ahead of time so b uses ca n event ual l y t r avel thr ough w i t h little o r no congestion. Priority Co ncepts The two types o f basic bus p ri o rity concepts are: )> Acti ve prior ity : Each bus is detected o n approach to an inte r sect ion and th e s i gnals are then changed. Active systems can be a comb i n ation of r eal o r fixed-time con t r o l s t rategies, and schedule or headway-based contro l strat egies. Active concepts are more e ffective and w id ely used. > Passive priority: T raffic con trol dev ices are adjust e d to suit the bus schedu l e along the route i n genera l using a c ombination of fixed time and s c hedu le-based cont ro l strategies The literat u re search has indica t ed t hat passive prior i t y systems and f i xed -time systems are rare l y used. I n some applicat i ons, passive prio r i ty i s impl emented on l y at certain i nte rsections p r imarily i t favors roads w i t h s i gni fi cant transit usage, often close to t he buses origin point where schedules are most lik e l y to be adhered to -whi l e the entire co rridor has an active p riority system. Passive priority does have t he benefit of being lower i n cost, however i t has limited potent i a l to i mprove bus ope r ations. s
Ill/$ SICI>Ut I'KIOI/ITY Table 1 Summary of BSP Strategies Control Strategies Real t ime Priority changes based on constantly updated i n form at i on Fixed-time Applies a f i xed plan to make decisions regarding p r iority Schedule -based Priority awarded based o n the bus schedule Headway-based Priority awarded based on the headwaybetween buses Priority Concepts Active prior i ty Signal i s adjus t ed for each bus on detection at in t ersection app roac h Passive priority Signals are adjusted to suit the bus schedule along the route BSP SYSTEMS Bus Info r mation and Priority System (BIPS) offers bus priority withou t the large i n crease i n de lays to other traffic (5). S IPS i s an active real-time system that can utilize different me!hods of detection. It cons iders the e n !ire bus priority system as a network that co n tains i nteracting buses and intersec!ions. BIPS utilizes five modules to exchange infor mation and determine what kind of priority should be given to each bus The modules used are: l> Projection on Route Module receives bus-position i ng messages (wave based, dead reckoni ng and beacon systems may be utilized) and then returns i nformatio n. l>Traffic Flow Estimator Module receives traffic volume information and then returns congestion information and estimates the d e gree of congestion and t ravel time l> Travel T ime P rediction Module receives information about th e status of signals later on in the route of the bus and determines an esti m ated time for !Tavel t h rough the route. l>-Priority Processing Module commun icates with the travel time prediction modules to get informat ion on various data. l>Monitor and Control Module starts the other modules and communicates between them Microprocessor Optimized Vehicle Actuation (MOVA) is a bus priority system that utilizes a signal control system to analyze lane data and control signa l timing (6 ). MOVA is an active rea ltime system. The system has been tested on isolated junctions in England using tra nsponde r s and l oops in the road, although other me!hods of detection can be used with the sys tem. MOVA employs a variety of bus priority techniques depending on the s ituation of each priority request. If the !Tans i t vehicle is arriving when the light is a lready green, the green signal can be extended to allow enough time for !he transit vehicle to travel thnough the intersection under normal conditions. If the transit vehicle is a rri ving when its signal is in a red phase, the other phases can be skipped or truncated, depending on the situation. 6
E!1EC71f!ENEJ;f OF 8l/J' SJC/Ut I'KIOKITY Sp l i t Cycle O ffse t Opti mization Techn i que (SCOO T ) an active u r ban traffic contro l system, prov ide s real t i me opti mizatio n (4) Bus p r i o rity i s g i v en v i a g reen extension, green rec all, and r e synchron ization. Var i ous detect i on methods can be used w i t h the system such as transponders and AVL. The SCOO T software has t he benef i t of p r eventi ng early term i nat i on of bus p rior ity phases, extend phases, or recall th e p ha s e I n order t o perform such fu n ctions, SCOOT app lies d i fferent weights to each de cisi on. The dec i s ion i s t hen made to advance (bring the phase forward fou r seconds), stay (not change), or 'retard" (delay th e phase f o r four seconds) S e l e ctive Prior ity Ne twork Techn i que (SPRI N T ) i s an active f i xed-t i me system based on many pri ority p r i ncip les f rom a system called "PROMP T (PRi o rity and InfOr Mat ics i n Pub l i c Transport) (10). The SPRINT system de t ects buses with loops in the r oad surface and t r ansponders on the buses. The system decides new s i gna l timi ngs th at allow a bus to travel t h rou gh the nex t intersection usi ng green extension and r eca ll. Constra i nts are bui l t i n to r ed u ce de l ay to non transi t t r a f f ic. Max i mum ex t ens i on l imits are i n p lace. A f ter allowing one reca l l, t he system cannot pe r form another recall until Onf: c ycle has passed w ith no reca l l. Compensation i s also prov id ed. Signa l Progression Optimizat i on Techno l ogy !SPOT) uses a decentralized app r oach. SPOT, tested by Fox et a l (7,8), is the intelli gent s ignal co n tro l proces sor of Urba n Traffic Optimi zat i o n by Integrated Automation (U T OPIA) Th i s is an active real t ime BSP system UTOPIA con siders non trans i t veh i cles, along w i th transi t vehicles, w ithi n a hierarc h i cal decentrali z ed traffic adaptive contro l syst em The system opt i mizes the total t ravel time with constra i nts of average speed and saturatio n fl ows Prob lems are classi f i ed in t o a l ower l eve l (at the i n tersection) or a decision l eve l (over the entire a rea) I t uses t h e Texas I n s t r uments Radi o Ide n tificatio n System ( TIRIS) tags for au t omatic detect ion. These tags pass by l oops in the road and give the location of t he bus t o the compu t er. I t m i nimizes cost functions over a rolling hor i zon of two m i nutes and co-operates with the nearby i ntersections by exchang ing traff i c infor matio n Phase cha nge t i mes a r e limi ted by phase order and duration I i m its. The fi eld trial of the enhanced SPOT system performed i n L eeds, Eng l and r educed trave l time by approximate l y 10 per cen t. Non-transit vehicle trave l times were unchanged The "Opticom system by 3M uses i nfrared co mm un i cations betw ee n buse s and intersect i ons w i th emitters on t h e bus and de tect ors a t the i n tersection. I n a stand a l one system the driver activates t he em i tte r A lin k can be estab l is h ed to sw i tch off t he emi tte r s when t he doors are open so tha t p r iori t y i s not requested while t h e bus i s stopped Opti com bus prio r ity contro l systems have been impl emented at over 40,000 i n tersect i ons worl dwide and have been used t o give priority t o both emergency and transit veh i cles. T h i s i s an act ive real -t i me system. PROS/CON S O F PRIORITY METHO D S An AI -Sahili and Taylo r study from 1 996 (2) concluded th a t the max i mum benef i t a bus can gain under the tested con d itions was an ave rage of 75 seco n ds out of 22 minutes 41 seco n ds o r 6 percent. T h i s serves as ev i dence of the var y ing degrees of effectiveness of BSP. The t est conditions 7
i n that study were 14 intersec tions in Ann Arbor, Michigan. The test used only green ex t ension, red tr un c ation with and without co mp ensation, and skip phase with and w i thout compensation T hey co ncl uded that t h e overall bene fi t s gained by buses were not suffic i ent to counter the delay t o other vehicles. They a lso dete r m i n ed the best bus priority i s one whi ch comb i nes e l ements of various systems for each intersection. Other stud ies have produced greater be ne fit s with BSP, illustrating the impo rtance of c h oosing the correct system for a particu lar co r r i dor. A Weesner and Meyerkord study co ncu rred (26), suggesting any ef fective bus priori t y must be ta i l ored to sui t each area and n o one p lan can b e i mp lemen t ed unifor mly an d b e effect i ve. A study presented at the Internationa l Conference on Traffic and Transportation Studies i n 1998 (18) tested the effectiveness o f bus priority w i th a simple twoi ntersection a rter ial. Each link was 200 me t ers apart and the signa l s ran with a 60-second cycle length. The extensio n a nd tru ncation per i od was five secon ds for e i ther priority method. T he st udy per formed one test without a n y priority, t he n one with priority. The average reduct i on i n journey time w ith p r i ority was 1 0.3 percent, with max i mum r educt i ons r a n gi n g as high as 37 p e r cent. The study did ma k e note of the fact t h at in some ins tances a bus may act u all y take l o nger to comp l e t e a journey with priori t y. This is due to congest i o n th at m ay arise fro m prev i ous p r iority and t he bus missing a suitable time f o r priority to be gran t ed. H oweve r overall, priority was proven t o r educe b u s jou rn ey times. Non-transit veh i cle d e l ay was not exa m i n e d i n this study. J acobson and Sheffi d eveloped delay models for testing traffic impac t w i th signal bus preempt i o n in 1980. The results of thei r tests r e v ealed t hat the greatest benefits to all traff i c with bus sig nal priority could be obta i ned by adjusting signal cycle and phase duration By r eviewing e xi sti ng lit erature, t he au th ors have gleaned tha t each system exa mined i n this report has both p ros and cons, some of which are outlined i n the fo l lowing tab le 8
lii1'ECm'ENEs.F OF BU> .fJCNAt I'XIOKITY Table 2 Pros and Cons BIPS P r o Con Cons id e r s buses i n both direct i o n s )> The use o f an i n d i r ect p ri o r ity strateg y in heavi l y co ngest ed traffic that just tries t o clea r t h e way fo r oncoming p rio r ity veh icles may be less e f fective t ha n ot he r systems i n areas with h eavy conges ti o n. MOVA-P r os Con )> P roduces s i g nificant r educti o n i n d ela y )> S till i ncon clusive evidence of how it for buses a f fects some non-transit traff i c )> Ability to awar d d i f fe re n t types of pri o r ity )> Limit on the amount o f pri ority that can b e given to vehicl e s per hou r o )Ptlcom : Pro s Con )> Saved transit buses an average of up t o )> Optic o m does resul t i n some del ay to 15 seconds per i n te r sect i o n when no n transit t r aff i c W i th t h e OP TICOM impl e me nted in Phoen i x Ari zona syst e m tests i n Phoenix A r i z ona, the r e )> Buses ran cl oser to schedu l e and was a small i ncrease in dela y o f 1.4 riders h ip showed an increase i n percent t o ot he r traffic. C harlo tte, North Carolina )> Average bus run time was red uced b y 23.8 percen t and average trave l speed i ncrea s ed b y over 30 p ercent SCOOT P r o Con )> SCOOT ca n work on a short -t erm traff i c )> Bus ti mes are not sig nificantly changed i ssue, whil e a lso notici ng t r e nd s ove r for b uses goi ng agai nst the peak flow of time t raffic dur ing t h e PM hours i n re l atively un
SPRINT Pro Con );. The SPRINT system in london p r oduced )> The be low average test results trom an average of 2.0 seconds r eduction i n London ind i cate that the system works delay per j unctio n fo r buses on the mai n best in less congeste d areas. roa d and 6.4 seconds reduct ion on s id e r oads. SPOT P ro Con )> T he fie l d t ria l of t he en hanc e d SPOT )> Delay at certain intersect ions to traffic system reduced travel time by d r i ving i n opposite direction to p ri o r i ty approximately 10 percent. traffic. PROJECTS IN NORTH AMERICA Across North Amer ica, trans i t and highway agencies are work i ng closer toget her than ever before t o address transportat i on prob lems facing their communities. T his increas ingl y co operati ve envi ronment i s leadi n g these agencies t o embrace operationa l strategies that can increase the amount of peop l e transported, not j ust v e h icles. Alth oug h these strategies a r e b e in g em bra ced, skept icism still remai ns regardi ng t he i r effectiveness. Some o f th e case study projects across N ort h America a r e featured in thi s report to g i ve an overv iew of the type benefits and costs associated with BSP projects Los Angeles A Los Ange les regional tra n sit a u thority survey showed that bus passengers cite d the l ong d uration of journeys as the b iggest prob l em w i th the bus syst em (22). Bus speed had decl i ned i n t he area by 17 perce n t i n j ust a few years I n order to solve th i s p roblem the tran s i t aut h o rity i ni tiate d a pro j ec t that adopted many bus prior ity strateg ies from a program in Curitiba Brazi l (20). LADOT imple me nte d a $10 million signa l prior ity project on tw o demonstration corridors (Vent u ra Bou l evard and the Sant a Monica Beverly H i lls-Montebello rou te) in the City o f Los Ange les. The system was i nsta lled at 210 intersections an d used 331 loop detectors with more than 150 transponders. In i tial results from LADOT wer e 22 -27 percent re ductions i n bus trave l ti me. LADOT is i n the process of evaluating the TSP project a nd have proposed expanding the TSP program to other m ajo r transit corridors if the initial r esul t s hold up i n their final analysis. lO
liFFECTJnM1f.f oF Bl/S SJCMU. momr T he syst em gra nts ear l y g r een e xt ens i on, free hol d ( hold a s ign a l green until the bus passes t h r ough t h e inte r sect i on} a nd phase call (brings u p the selected phase that i s not n ormally act i vated, t ypically fo r queue j umper or l eft t urn } priori ty modes. B uses a r e d i spatched every 3-10 minutes and can be i nst ructe d to s low down o r speed up (with i n th e speed l i m it} to avoid bunching up with other buses. In orde r to reduce delay to non transit t raffic, softwa r e was c reated that p l aced l i m its on any g re en ex t ension to t en seconds. Even if the i ntersection i s saturated w i t h buses r equesting p ri o r i t y, t he syst em will grant onl y te n seco nds. Addi tionall y, a t high l y congested i ntersect ion s g reen extens i on is o n l y implemented i n every o ther cy cl e T h e Ventura Bou l eva rd r oute decreased bus trave l time by 25 perc en t a n d the Santa Moni ca Beverly H i lls Montebello r oute a l so e xper ien ced a 2 5 pe r cent r educt i on. I t shou l d be noted t hat i n a d d ition to p ri o r ity t echni ques, t h e bus routes had their n u mbe r of stops reduced from a two-tenth mil e m inim um be t ween s to ps t o an e igh t -t enth -m i l e m i nimum A d ditionally, low floor buses were used tha t are easier and quicke r to board. Furt he r more i n o r de r to e n courage passeng ers t o exi t from the rear of t he bus, large sign s were posted and the drivers verball y direct passenger s i n t h at direction. Despite these add i tiona l impr ovements, pro ject officials c redi t (subjective) 30 to 40 pe r ce n t of th e 25 pe rcent r ed u c tion to bus p r iori t y techno l ogy. The L A system uses a Model 2070 traffic s ig nal co n troller This is a first generation Advanced Transportation Co rltroller (A TC) des i g n ed b y the LADOT and Californi a Departm e n t o f T r ans p ortat ion. L A DOT staf f deve l o p e d all o f t h e software. Eac h i ntersect ion i s eq uipped wit h sensor s a t the con t r oller cabin e t and induct ive loops for each direction Buses equipped w i th transpo nders are de t ec t ed w h e n t raveling over loop detectors. The s e nsor u n i t in th e cont r olle r transmits the bus ID n u mber to the c omp uter. No priority is g r a nted i f the bus is ear l y o r run ning on schedule However, if the bus is l ate then th e cont r olle r will prov ide priority in order fo r the bus to catch up w ith the schedu le LADOT is currently testing a prototype syst em with AVL that would comm u n ic ate the arrival t ime informatio n of the next bus from the ATSAC (Automated T raf f i c S urveillance and Control Center! to cha ngeable message signs at the bus stops Portland Port l and has conducte d tests of bus p r iority w i th the 1993 Powell Boulevard t est a nd the 1994 Multnomah Bou lev ard test (11, 15) These test s r evealed t hat systems t h at requ ir e bus operators to manually request priority t o t h e t r aff i c signal w hil e ope rati n g the bus a r e not efficient methods. The bus driver often failed to call for priority A test o f Opti com (g reen exte nsion/r ed trunca tio n control) 11
was quite effective, al t hough th e soft wa r e r e qu i r e d e n hancements. The Powell Bou l eva r d test did revea l a be n efi t t o b us pr i or i ty but the two mile stret c h of r oad used f o r the test was too s m all t o reveal any s t atis t ically accu rat e figure s f or a c ity\Yid e i mplemen tati o n Howe v er, these stud ies did lead the way f o r a $5 million p roject tha t i n c l ud ed t he installa tion of opt ical em itter s on t h e e ntire stand a r d bus t leet (775 buses), and a r evisio n o f t h e AVL (A utomat i c Vehicle L ocat ion) syst em t o act ivate the emitter. I n addi tion a "2070" generat ion controlle r a nd optical detectors w e r e installed at intersections Th i s projec t p rov ides co mpl ete s i gnal p r i ority f or five major bus route s i n Portla n d. Tr i Met, Portl an d's public tra n sportat ion pro v ider wanted t o i mpr ove transi t perfo r mance and r educe dela y to buses i n order t o prov ide a m ore v iable alternative to d r i v in g a car Add i tio n ally P ort l a n d c r eated a B S P system with t he obj e ct i v e o f reducin g ope r a ti ng expense s to allow e i t her e n h anced h eadways or additi ona l services e l sewhere. T hey a lso wan ted to imple m e n t BSP w i thout increasin g d e la y to non-tra n sit vehic les. Furt hermore i n choosing their system, t hey impl e m e n ted Opt icom from 3M due t o the abili ty to a lso use the syst em f or eme rg e ncy vehicles. I t was important for Tri-Met t o i mp ro v e transit and als o impr ove emergency signal prior i ty and not i n terrupt any preemp ti on f or emergen cy ve h i cles. T he t ec hnol ogy behind t h e Portl and syste m is the Bus D i s patc h System (Tri -Met's AVL syst em ). An on-boa r d GP S (Gl o b a l Posi tion i ng System ) satellite r ece i ver de t e rm ines t he l oc a tion of t11 e b u s I f t h e bus i s r u n n i n g l ate (classi fied as ove r 7 m i n utes) t he n the Opticom e mitter is activated to in i ti ate p ri o r i ty The bus t h en rec e i ves p r ior i ty t hrou g h i n terse ctions. However all emergenc y ve h icle s have a high pr i or ity setting that overrides the trans i t' s Nlow" pri o r ity setting. To date this project has b een i mp l emented at six t y i n te r sections. T h e cu r rent sys tem o n ly initiates p r i o rity whe n a bu s is r u nning beh ind sch edu le. To use p r iori t y at other times 10 s p eed u p tra n sit would have r eq uired chan gi ng t he i r schedu les to a ccou n t for buses r un n i ng qui cke r The deci s i on was made t o wait u ntil new sched ul e s a r e c r eated to make c hanges. T ri -Met i s i n the m i ddle o f eva lu ati n g t h e system and have ye t t o release any doc ume nted i n for m atio n abou t the ben ef its of the system. Seattle The K i ng County Depar t m e n t of Transportation in Seattle i mple me n ted signa l priority i n the 2.1 mile Raini e r Avenue Syst e m (activat ed in the spr ing of 2000 } whic h i nclu des n i n e inters ecti ons five with prio r ity. A report from King Cou n ty (3}, r eported t h at they h a d many obj ect ives with t he i r BSP s ystem. They wanted to i mp rov e t he i r ab ilit y to transport p e o ple and i mprove t ransi t pe rforma nce. I n theory, an improvement i n tran si t p erformance provides addit i onal ince n tive for peo ple to swi t c h m od es. A d d i t ionally t he system w a s designed w i t h non transi t veh i cles in m ind, t o avoid any del ay t o those veh icles. Anothe r prima ry objectiv e was to a void implementing a 1 2
I imited contro l system that disrupts the f low of t raffi c. Subsequen tly, a system w ith a large amount of funct i o na l prior i ty control strategies was dee med necessaoy. Kin g County is the mos t recent major agency to i n itiat e a major transi t prior i ty project and produce a final report on t he i r initia l exper ie nces. Th e refore, d ata fro m the i r study is incl ude d i n more d etai l than most othe r case study areas. T he follow ing i n f orma tion i s s ummar i zed from the Rainier Avenue System Project Fina l R eport (27). The hardware for the system induded t he A mtech RF (radio f r eq uency) tags on t h e buses, Amtech lo g per i odi c antennas on the road, and the Transi t P r ior i ty Request Gene r ator. To r un the s ystem, The C i ty of Seattl e used a co mbi nation of con t rolle rs. The syst em i s made up of Eag le, PEEK LMD 9600 and Econolite controllers. The combination provided t he depth of function requi r ed to avoid any pro ble m s t hey experienced with a p r eempt only" type system. The y work e d w ith the controll ers to change timing s and deve l oped th e followi ng firmwar e impl ementat i on re quirements: )> Traffic signal s shall extend the i r green interval fo r approaching priority vehicle s ? Traffic signals shall shorten re d displays for approach i ng priority vehicle s ? Traffic s ig nals shal l not shorten any min i m u m or clearance intervals ? Traffic signal s shall not skip any phases ? T raffi c s ignals shall not break co ordina tion if a BSP serv i ce call i s requ ested. The A.M.-peak p e rio d along Rainier Avenue e xperi enced a 2.3 secon d/veh icle (13 percent ) reduction i n average intersec tion de la y. The i r s tudy also considered d ela y to t he m i nor street movements. M i nor move ments include all side street movements and the protected main s tree t le ft turns The A.M.-peak pe r i od featur ed a reduction i n minor movem e n t delay of 0.9 s eco n ds/ve hicl e (3 per ce nt). M i dday -peak period delay i ncreased slightly although t he le vel of serv ic e rema i ned the same Before and after studie s conducted by King County Metro show that TSP reduces the average intersection bus dela y by approximately 5 seconds per TSP-equi pped i ntersection a r ed u ction of 24 percent to 34 per cent for TSP.eligible buses. In addition intersection bus delay is r educ ed by an average of 34 percent when a bus i s eligible for priority treatment One cri tical result of the study is ther e were no side street cycle failures. This means the reduct i on in green time for sid e streets caused by TSP did not cause any side street veh icles t o wait through more than one g reen signa l cycle. King County concluded that transit priority p ro duces minima l side st r eet de lay that i s like l y un no ticeab l e to the d riv er. The exper i ences gai n ed from the fi eld study brought King County t o recommend "tailoring the speci fic contro l and transit operating stra t egies to the time periods where TSP i s being deployed 13
seems t o hav e t he gre atest poten tial for b enefit in terms of re ductio n in person delay." They also state "TSP implem ented i n a m oderate way (i.e., no phase s k ip p ing, minimal i mpact on coordinat i on, etc .) is virtually u nnot ice abl e by the motorist. The f i rst of 26 i ntersectio ns of the Regio nal Tran s i t Sign a l P riority System has bee n successfully deployed at the i ntersection of G enesee Str eet and R a i n ie r Ave n ue. The initial results show: }> A 5 0 percent reduct ion of s ignal re l ated stops b y prioriti zed buses }> A 57 pe rc ent re duc tion i n average delay f o r the priorit ize d buses }> A 13.5 perce nt dec rease i n i ntersecti on average pe rson d elay -;Avera ge intersectio n vehi cle de lay d id no t change )> A 3 5 pe r cent r educt ion i n bus trave l t i m e var i abili t y f o r the p rior i tized buses l> H f e cts to sid e street delay were i n s i gni fi cant and there was no significant change to non p r i o rity approac h queue l e ngths K i ng County Depar tment o f T ransportation r e p o r t s t hat the r eduction in delay at s i gnalized i nt ersections resulted i n trans i t patrons expe ri enc i ng a smoother and more comfo r ta bl e r i d e Add i tionally the r e duction i n stops reduc ed the transit operat or 's wor klo ad. Orlando The Cent r a l Flo rida R.egional Transportation Autho r i ty commonly known as Lynx, s t arted p r ovidin g service o n an improv e d downtown circu l ator l ymmo i n August 1 997. Even tho ug h thi s is a bus r api d transi t (BRT ) project, i t is f eatured here because two o f the i ntersect i ons have special phases for L y mmo buses. The service offers exclus i ve lanes f or t h e enti r e 2.3 -m ile r out e i n addition to signal p r e em ptio n AVL, and free fare, so th e r e i s no f are collect ion delay. The t otal cap ital cost for L ymmo was $21 m illion. T here i s no docu m en tation of costs for implem entati on and operat i o n of t h e BSP system for the two intersections. Orl ando did not use an outside r supplie r for hardwa re; they designed the system themse lves. B ecause Lymrno opera te s i n contra flow lanes all bus mov ement s a t intersections a r e control l ed b y specia l bus s i gna ls. To p reve n t co n fusi o n these signa l head s use 'white" l ines instead of the standard red, yellow, a n d g r een I ights. When a bus ap proaches an i n te rsect i on, a loop detecto r i n the bus l ane trig gers the i ntersection to allow t he bus t o p roceed eithe r i n i t s own signal p hase (e. g when making turns no t o th e rwise permitted) o r at the same t i me as other traffic is re leased when no conflicting traff i c moveme n t s a r e pe r mitted. One of th e dire c t b e nefits of hav i n g BSP a t two i nters ections is Lymmo is ab l e to kee p on schedu le. Wai ting time i s lower for lymmo compared to its predecessor freebee, i f dispatcher s wer e to us e 14
t he AVL system t o i nstr u ct drivers to ho ld ear l y buses to ad j ust t o t h e schedu l e i n o r der to ac h i eve a more even d i st ribution of headways Miami The South Dade Busway is a two-lane b us-o n l y roadway const r ucted i n a form er rai l right-of way adjacent t o US 1, a major art e r ia l. Th e Miami-Dade Transi t Author i t y (MDT A ) opened this bu sway in Febr uary 1 997. BSP for busw ay buses is available at six s i gnal i zed i ntersecti ons a l o n g the who l e cor r idor. The system uses typ e 170 cont roll ers by Safetran signa l cont r ol. The SSP timings and software wer e des i gned by the MDTA and i n ternal agenc ies rather than usi ng an o u t s ide vendor. I n the northern port i o n of the b usway, t he re i s very l i tt l e separation be t ween US 1 a n d the bu sway At these locat ions, the traf fic s i g n a l ization was desi gne d so t hat the busway inter s ect ion and U S 1 intersec tion would function as one signal syst em. The signa l facing buswa y t raffic i s red unless a loop d ete ctor senses an app roachi ng bus If a bu s is detec t ed and i f US 1 has a green s igna l the busw ay also gets a green. If US 41 has a r ed s ig n a l t he n t he busway and US I get a green s i g n a l and the lef t and righ t turns get a r ed s i gnal. I n the i n i ti al mon th s of operatio n there w e re 14 collisions between buse s and s i de s tr eet traff ic. I n every case, t he private vehicl e d river had not no tice d or had i gnored th e red sign a l a t t he busway. The signalizat i on was changed in August 1997 so t hat the busway signalizat ion was synchronized with US 1 from August 1997 to July 1998 there w ere five co l lisions, a m uch lower rate than i n the firs t mo n t hs of operation. R ight turns f rom side streets t o US 1 were facil i tat e d by i ns tall ing a r ight turn lane and right t urn s ig nals. Also, the drivers of t he tr ansit vehicles have been instr ucted to b e more a l ert and s low down at these i ntersections. Right t urns (south to w est) fro m US 1 t o s id e streets a r e permitted onl y on green arrow, to prevent righ t-turn i ng vehicles fro m colliding w i t h busway traffic. The buswa y cost $60 million: $ 1 7 million for r i ght-of-way acquisition and $43 m illion fo r construction (includ i ng ei ght miles of twol an e ro ad w a y and 1 5 stat ions). T here i s no documen t ation of costs fo r implemen t ation and operat i ons o f the six BSP in terse ction s The S outh Dade bus way seems t o have provided little (less than 10 percent) or no timesavings fo r tra nsit vehicles However, r idersh i p in the corridor i ncreased 49 pe rce n t on weekdays, 69 pe r cent on Sundays, and 1 30 per cent on Saturdays as of May 1998. This i ncrease in ri de rship i n d i rectly can be attributed to BSP. However, incre a s e d ridersh i p has resulted i n longer and more f r equent bus s topping. 15
WHEN IS NON-TRANSIT TRAFFIC LEAST AFFECTED? T he effe ct of p riority on non-t r a n s i t vehicles t ends t o vary dependi ng o n t he le vel o f c ong estion. I n gene r al, c ompensat ion to non -transi t traff i c is easie r in a reas w i th low congest ion. Docu men t e d tes t results of m ost BSP s y s t ems h a ve shown th at som e small delay t o nontransit tr a f f i c i s i nevita bl e. The f i eld t ri a l of the enhanced SPOT system pe r f orm ed i n Leeds r ed u ced travel tim e by approx i m at e l y 1 0 p erce nt. Car t ravel t imes were u n cha nged. However, most systems are tested o r implem e nted on jus t two or three i n tersection corrido rs, m e a n ing t hat test results m ay be m i slead i ng for p r act i cal applicat ions No one system has t h e p erfec t priori ty strategy for non-tra ns i t vehicle s S ome e l ements o f a p riority system t o e nsur e t r a f fic ope r a tions a r e l e a s t affected a r e discussed be low. T he box provides a summa r y o f this sec tion with a more d e tailed d iscu ssion be low. 'l> Real-t i m e control strategy is pre f e r able )l> D e la y can be m i n i m i ze d w ith con s t ra int s on max i m um and minimum g reens > Softwa r e with a weight i ng system has potential t o reduce delay to non transit vehicles )l> I n i ti a t e priority on l y when bus is r unni n g l a t e )> I nd i r ect pri o ri t y strategy Real -time c o n t rol strategy is preferabl e and has far more potential t o red uce dela ys to non tran sit t raf f i c. I f the syste m uses a p rog r a m s u ch as TRANSYT t o calcu late sig nal timing p lans and t h en selec t s whi c h p l a n best suits t he curr ent traff i c con d i tions, ope r a tors ca n a djust t he signa l tim ings i n the event o f conge stio n, ac cid en ts, or other u nforeseen events. A l though re a l-ti m e cont rol i s prefe r ab l e i t i s not wjthout flaws, such as: )> Some systems can not c onsider two or more tra n sit v eh i cles app roach i n g an inter section fro m different app roaches. Real-time signal systems r equ ire h igh-speed compute r s and communi cat ion syst ems t hat r aise the cos t of i mplementin g th e system. In addi ti on, computation times for each control decis i on can o ft en take too long However with p ro gressing techno l ogy, both these issues a r e becom ing less and less of a problem. J;> T h e sign a l timing p r ed i cti o n m odel s some systems use may not be re l ia b l e i n pract ic al applications because of variations i n t ra ffic volume, unforeseen events and the i m pact o f p rio rity itself. At t h e i nte rsection lev el, d e la y to non -trans i t vehicles can b e min i mized wit h con stra ints on maximum and minimum greens. I n o r de r to r educe delay t o n o n-tr a nsi t traffic, softw a re should be created to place l imits on any green extension to a max i mum amount ( e.g., 10 seconds) Ev en i f the in tersection is saturate d w i t h buses r eques ting pri o r i ty, the s ystem shou l d only g rant th i s 16
maxim u m amount of tim e. Add i ti o n ally, a t h ig h l y congest e d i ntersections green extens ion s shou l d be l i m ited to e very other cycle. Compensat i on s hould also be prov ided. S of twa r e w ith a we i ght i ng system at inte rsect i ons may be used to reduce delay t o non -tran sit veh icles. D i ffe r ent we i ghts can be created wit h the h i ghest wei ght mean i ng t he system g r a nts p r i o ri ty i mme diately, while the l owest weight indi cates n o p r iority s h ou l d be awarded The weig hts in between would g i ve green ex t ensi on o n l y a nd g ive g reen e xt ension/ recall w i t h a l eve l o f cons traint determined by the congestion o f non-transi t tr af fi c The r at i o bet ween bus headway and expec t ed headway const itutes the degree to wh i ch the bus i s said to be ru n n i ng be h i n d schedu l e. The we i ght would depend on how lat e the bus is r u nn i ng. Ano t her s i mp l e so l ution to reduce conges tion to non-transit t r affic is to in i tiate priority o n l y when t h e bus i s r unnin g beh ind sch,;,dul e I n heavy con gest ion, an i nd irect p r i ority s trategy m igh t wor k th i s t ries to clear traffic ahead of a bus S i nce bu s arrival time is ofte n difficuh to accuratel y de t erm i ne w i thout a f ully ope r ational AVL system, t h is t ype o f system may just attempt to clear as much congest ions as possib l e a l ong t he bus rout e. CHARACTERISTICS OF AN IDEAL BSP SYSTEM Systems operate with a variety of dif f erent tech n i ques that a i m to reduce delay to trans i t ve h i cles. The methods of prio r ity us,;,d are all relatively effec t i ve, w i th each system usi ng a mixtu r e of d i ffere n t p r ior ity techniques. Som e o f tbe ideal e l ements of a priori ty system a r e as follows: : Ability t o track bus moveme n t s accu r ately. Most systems dev el oped now a pply so m e form of automa t ed v eh i cle lo cation A system that constant l y tracks the location o f a bus i s idea l. Rather t han r elying on t he bu s to cross a sensor in the road a more advanced system would know if a bus has been delayed before reaching the next intersection Ar.y system that operates as a neNvork of i n t ersections is less effect i ve without constar.t monitoring of bus l ocation T h i s also co uld be i n teg r ated into real -time passenger in f orm ation serv ices to p r ovide estima t ed arrival t imes. : Abilit y to measure a nd record s tat istic s on t h e bus r o utes to form transit plans based on statistical a nalysis A l so consid e r t r a ffi c vol ume, passenger occupancy and other related f i gu r es. )'Ability to offer a wide variety of priority t echniques for differ ent situat ions. l> Ability to mini m ize delay t o n on -tr ansit tra ffic a n d offer compensation. 17
)> Ability to est i ma te cos t to both passenger and t r a n sit agency (based on ave r age d e lay) associated wit h e n act ing any g i ve n pr i ority method. PROFESSIONAL O PIN I ONS O N BSP Five transit p l anners and five traffic engineers were interv i ewed for this r eport. All profes siona l s interviewed wor k ed for tran sit agencies i n th e United States that have impl emented BSP systems. They agreed t o th e interv i ews on the basi s of r e m aining ano n ymo u s w i th i n the r eport. The t ab l e p r ovides a summal)l of their com ments with more det ai l ed comments be low. T able 3Prof essio n a l Opini ons Issue Transi t Pla n ners T raffic Eng i neers Effectiveness of BSP Reduce s de la y t o buses. Promotes transit as alternative mode. De lay to Non-Transit Traffic No no tic eab l e delay to side De lay cons i dered in design. t r affic. Where to P u t BSP P l aced on corridors with Picked r outes w i th most highe s t ride rshi p d e l ay. How to I mplement BSP Rel ocated bus s t ops Revised controlle r firmware and timinp; BSP T e<:hnology Success of v e ndor s sy s tems No BSP vendors ar e tru l y is dependen t on proven. i mplementation Public Perception of BSP Little p ublic reaction No feedback received. Effectiveness of BSP All trans i t planners invol ved in Bus Signal Prio rity (BSP) think BSP improves a transit agenc ies ability to move customers: )> One t ransit p l anner who has implem e nted BSP says that, i t clearly reduces delay to buses. All o f the traffic engineers concurred that i t does improve the ability to move people The t ransit p l anners all believe that BSP meets n e eds that are not being met by current congestion mitigation methods. One transi t planner stated t hat BSP provides a faster, more competitive tri p b y bus. This allows buses to provide a more viable alternat i ve for customers than merely driving: 18
)> "Whe n installe d on a cor ridor b a sis T ransit Signal Priority could r educe operat i ng expenses for the cu r ren t frequenc y allowing either enha n ced headways on that l i ne, or addit i o n a l or new service e lsewhere. Anot her transit p l anner conc u rred, stati ng that BSP can inc rease the numbe r of peop le who use buses i f the moveme n t and seat capacity of buses is a l so i mproved. A tra f fic engineer said tha t BSP i s just one way to promote transi t as an alternat i ve mode Another engineer pointed out: )> I t will take a number of years to install BSP at all inte rsec t i ons. There are a lso a number of i ntersections wher e t here are congestion problems that do not have transi t routes. It meets some very specific location p robl ems, but not all.' Delay to Non-Transit Traffic Although some lacked sufficient data t o substantiat e f i ndings no agenc y expe rienced any significant delay: l> Transi t planners inte rviewe d for th i s study expla i ned that SSP did not create any significant delays to non transi t ve h icles. One transi t planner d i d report a measured increas e in delay t o side street traffic o f 4 s econ ds pe r vehicle. However $UCh a delay was not notic ea ble to those non transit veh icles and the person de l ay rate actually decreased because of BSP. Ano t her tra nsit p l anner said that t hey conducted a detailed study to quan tify a n y traffi c operation concerns such as delay to nont ran sit vehicles. They looked at int ersection delay, s i de street del ay, cycle fai lures, and comp l aints They kept that in fo rm ati on in mind when des i gning the system. Regarding delay, a trans i t planner said that i n model i ng, no net loss was clear: l> A l though a few l ocations did expe r ience small delays, a few expe r ienced small improvement$. Furthermore, i n preliminary testing, no significant impacts or delay s wer e perceived. T hat agency is treating non -transit delay as a signi fi cant issue. I n 200 1, as transit signal p r iority is impl emented on their entire corridor, they w ill test the i mpact. In addition, a different transit planner stated that, "care is be i ng taken at e very intersection that adjustments w i ll not drastically impact o th e r traffic movements. N /9
A traffi c e ng i neer sai d that a l th ough they hav e no data for delay t o nontransit veh i cles because their syst em is not ope r at i n g yet, they do have political s u pport from the c i ty. Ano t her t r a ffic e ng inee r exp l a i n ed that t h ey d i d cons id er d 1 e potent i al d e lay w hen designing the syst em. Subsequently thei r s y s t em just gives ext en d e d ti me i n phases, because t hey were not willing t o s ki p phases en ti re l y. Where to Implem ent B S P In all cases, a hig h degre I n o r de r to det e rmin e s u i table cove r a g e for BSP, one t r a n sit p l anner d evelo p ed an Excel s preads hee t to estimate expected bus perfo r mance impr ovemen ts. That tool esti m ates prod uct i vity b y ana l yzi n g t he number of buses and passengers, bus arrival patterns, s i gnal phas i ng, a nd rel ative green /re d split t ime. Many transit p l an ners in stalled BSP on t h e corr i dors wit h the h i ghest rid e r s h i p and fr e q uency. A transit pl an ner e xplain ed: )- Thi s mean s t ha t the t ech nol o g y be nefits the l a rgest n umbe r of passenger s and provides th e greatest c h ance for en h anci ng t he eff i c i e nc y o f resources used : In order to ensure the system wou l d be effective "e xt e nsiv e test i ng occur r ed on th e te chno l ogy i tsel f. In a d d i tion, short test cor r i dors were i mp lemented and t ime savings measured : T hat transit planner a lso said an ana l ysi s o f t he b efor e a nd after stati s tics wou l d b e co ndu c ted to dete r m i ne the effec tiveness o f BSP. )A traff i c en g i nee r exp l a ined t hat th e y c hose those rou tes w ith s ubstantia l del ay f o r their BSP s y s tem s Another en g i n ee r s t a t e d t hat t hey used NETSIM to eva l u a t e th e s i m ulated e ffects b efore ope n i ng the syste m. How to Implement BSP Transi t planners d id make some changes to operat ing characteristics in order t o i mplemen t BSP. One planner exp l ained that t hey worked w ith t r a nsit sched u l ing t o tak e some time out of t he schedu l e Other steps taken w ere: )> Rel ocat i ng bu s stop s to p ro vide far-side s t ops where possible and retiming and opti mizing the signals. .20
The traff i c eng ineer s of t hat ci t y work etl with t raffi c contro lle r vendo r s to revis e con t ro lle r firmware t o p rovide enhanced signal logic. T he tr affic eng ineers a lso had to deve lop BSP t i m i ng s tr ateg ies t hat we r e ap p rop r i ate f o r t he i r corr id o rs. On the issue of t h e f a r s i de stops, a n other t ransit p l anner agreed, noting that most of t heir bus stop s had p r eviously been near-sid e o f intersections. He commente d: J> since BSP opera tes m ore effectively on fars id e stops part of the process has been to try to m ove stops to the Ia r s id e of s ignalized intersect ions: One traffic engineer did state they made no changes to exis t ing operating measures in order to i mpl e ment BSP. Anot her eng i neer stated t h at s om e c hang es we r e m ade to fine tun e tra f fic control. BSP Technology Mos t of the transit professional s be l ieve the fie l d of BSP vendors is unp r ove n. One trans i t p l anner used a det ect ion system based on RF techno l ogy t hat was ch osen and design ed based on a regio nal ove r s ight committee of loca l traffic engineers That transit p l anne r s tated that t he detection system i n of itse l f does not p rovide p rior i ty since that i s the traffic con t r oller's function: He went o n to state: l> As a n example I would not consider 3M Opticom (or othe r detect i o n system vendors) a BSP vendor since th e success of an Opticom system is completely dependent on how the p riority con t rol strategies a r e implemen ted.' An e n g ineer agreed, pointing out that no BSP vendo r s a r e t r u l y proven. It was agreed by many e n g ineers that: J> It is crit ica l for the tr-lffic e n g i neers to deve lop the priority contro l strateg ies. Nevertheless, one tra n s i t p l an ner d i d c hoose the Opticom system large l y because : l> "Th e benefits (of Opticom) can be double d by using the system for emerge n cy v eh i cles a t a high e r priority as well. Th e planners did experi e nce some difficu lties when they tested the loopComm em bedd ed loop system ( transducer techn o l ogy used to tag vehicles for d etection ) a nd TOTE radio systems ( t he rad i o f req uency part of t he d e tect i o n tag). 21
Public Perc epti on of BSP T he re was n o t an overw helming public re sp onse to BSP, although s ome impact was f e lt: )> One tra nsi t p l anner d id report s ome i n terest fro m the public abou t B SP. H e stated the r e have been several press r e leases. O n e other p lann e r commented that "not eno ugh o f the system i s o n -line for any one t o notic e y et." Howe v e r i t was noted that : )> 'In publ i c meeti ngs and mailings, we have d escr ib ed t he system, focusing o n t h e b e n e fits of r educ i ng transi t trave l ti me i ncre asing the number of intersections w ith emergency sign a l priority and ma k in g i t ve ry cle a r t hat bus p r i o r ity will not i n terru p t emergency vehicl e preemp ti on. p An engineer s t ated that although th e public l i ke l y notices the diffe r ent s ignals they have for buses, no f eedbac k has bee n rece ived. SIGNIFICANT ISSUES OF BSP The following i nformation i s summar i zed f r om an extens ive lite ratur e search. This section a ddresses t he more prom i ne n t issues r e g a r d in g the impl e ment a tion and i mpact of BSP. Impact of BSP Strategies )> U n cond i tiona l signal priority should be reserved for exp ress bus service d uring of f-peak hours ( 9). l> Unconditional signal priority shou l d be reg ul ated by placing l imits on g r een extension and r e d t runcat ion l engt hs, especiall y at i nterse<..tions w i t h busy c ross s t reets (9 ) )> Active sig nal priority may cause d isruptions a l o ng h igh l y saturat ed c ross street s that do not dissipate b ef ore the n ext p r ior ity c all (9) )> A real-tim e signa l contro l model is gene rally considered more f le xible in a cco mmoda t i ng bus operations t ha n a fixed time cont rol model. The perfor ma nce measures fo r a bus actuated system a re be t ter than for a fixed -t i me s y s t em when a lso cons id ering th e side street traff i c (17). However, d ue to the difficultie s o f processing on-l ine d ata c oncur re n t ly, both c on t rol mode l s fail to treat o n -lin e transit oper ations effectively ( 17). 22
);> Thou g h almost a l l systems p rovide r easonab l e contro l f eatures, some still fail to treat two or more transit v eh i cles co ming c o nc u r r ent l y from d i ffe ren t approaches (17). Network-Wide Impacts of BSP Ji> I t was not e d tha t when p ree mp tion too k p lace a t a high l y congested i n tersection (at saturat ion) preemp t ion effect s conti nued for several cycles (2). li> T he presence o f a s i ng l e heavy -vol ume in t ersection in the n e twor k can skew the ove rall networ k s tatistics. The we i gh t s of t h e s e interse ct i o n MOEs (Measu res O f Effe ct i venes s ) wer e very s igni fica n t i n t he ove r all ne twor k MOEs (2). l> I f s i g n i ficant changes are i ntrod u c e d to tra vel t i mes for genera l tra f fi c route c h o i ce effect s may result i n inc reases i n vehicu lar trav e l for v e h i cles u s in g t h e ma i n arterial, o r for those e xper ienc ing increased de l ays at m i nor approaches to i ntersec t ions where buses a r e g i ven pri o r ity (13). )> Temp o r a l e ffect s m ay acco mpany s u c h s p atial impacts, suc h as changed peak pe r iod tr i p starting t imes, to avoid i ncrease d congestion (13 ). )> The a dvantage of con sider i ng bus p rior ity i n a network i s to achieve a stable and opt i ma l co n tro l over t h e r elevan t i ntersections p r ovided tha t the p r edicted b us arr ival t i mes a t these i n te r sections are r easonab l y accurat e (5). Impact of Congestion > The di r ect pri o rity s tra t e gy i s s u i tab l e for traff i c s i tuat i ons w i t h a low deg ree o f conges tion (5). > The indi r ect p r i o r i t y str ategy is sui t ab l e f o r traffic situat i ons w ith a hig h degree o f conges tion (5). Bus T rav e l Times l> Th e hold i ng control parameter (deter m ining how l ong traf fic i s held) i s the most criti ca l d ecis ionma ki ng var i able i n bu s contro ls at bus stops Holdin g cont r ol can s i gnifica n tly impr ove t he regulari ty of b u s movement. With holding cont rol, the average wait time of passengers decreases. However, t h e average in-ve h i cle time of passenger s and the ave rage b u s t ravel tim e i n c reases. Unde r schedu l ebased co n t rol s t r ateg ies, ea rl y buses s hou ld b e 2 3
held un ti l t he pre-p l anne d s chedu le. Unde r headway -based con tro l strategies, early buse s should n o t be he ld un t i I the pre-pl anned headway ( 1 7). > Schedule-based co n t r ols are easy t o i mpl e m ent because t hey do not need i nformation of bu s lo cat ions. In ad d i tion, sche d u le-based co ntro l s i mprove t h e o n time perf orm a nc e of bus serv ice. This i s very i mportant for lon g headw ay s i tuat i ons (17) P lacement of Bus Stop s ? Changes in th e positioning of bus-st o p s could allev i a t e some p ro blem s associated with BSP (6) > If the bus-s t o p i s pl aced afte r the j unction, t he bus det e ctor cou l d be posi ti oned further upstream o n the approach g i v ing m ore adv a nced noti ce of t h e app roach of buses. I f a bus lane c ou l d also be added, buses wou ld be less i mpe ded fro m reach in g t h e juncti on (obvi ous l y ) and t he b en ef its would be e v e n great e r (6). > Fars id e bus s t ops sho u l d be used w ith ac ti ve signa l p r i o r i ty to ensure t hat sig n a l p r i o r ity calls a re not wasted a s t ransit veh i cles dwell a t bus stops (9) IMPLEMENTING BSP In orde r to assist a transporta ti on agenc y t o de term in e i f BSP is app ropri ate and p r a c tica l for th ei r area, a series of chec klists and spreadsheet s wer e developed I n addition to determ inin g appropr iateness of BSP, these tool s assist a n a gency i n determ i n i n g what changes may be need ed for p riori t y to fun ction ; and w hich method s of p rior i t y are best s u i t ed for t h eir area. Th e pre -imple m en t atio n c h ecklist was developed to determ i n e i f BSP is viab l e i n a gi ven area. T h e guide l ines wer e crea ted to determ i ne any necessary c h a ng e s for p r iority a nd wh ich met h ods of priority a r e advisab le. The pre-imp l ementat i on checklist is a tool t o c o nduct a ben efits assessment of BSP and id ent i f y which priori t ies to i mplement. Th e BSP assessment tool i s extracted f r om the SCRITS (SCReening f o r ITS), a spreadsheet analysis tool developed by t he F ederal H ighway Admin is tr a t i on (28) (http://www .fl1wa.do t.gov/steam/scrits htm) for estimating the user b enefi t s of Intell ig ent T rans portat ion Systems (ITS). T he too l allows p ract i t ioners to obtain an i n i tial i ndication of th e possib l e benefits of var i ous I T S applicat i o n s The revised SCRITS spreadsheet will be posted on the FHWA webs ite. .24
T he BSP o p e r atio na l a n d d e s i g n g u i del ines are f o r use by p r ac t iti o n e r s once t h e d ecision is made t o i nvest i n BSP These gui deli nes h e l p t o identif y what type of p r iority s hou ld be i mpl emen t e d based o n certai n prefere nces. Bus Signal Priority Analysis (BSP) B e n efi t Assessment Tool A p pendix A conta ins t he A n a l y s i s of Bus S ign a l Pri o r ity s p re a dsh ee t whi ch i s one o f 1 6 d i fferen t a pp l icat i o n s i n c l uded i n the SCRIT S sp r eadsheet. T he spreads h eet conta ins secti ons o n pre impl e m en t at i o n bus and traffic operations, costs and benefit s a s w e l l as an i n t e rsection specific guideline Ove r a l l t h e BSP s p readsheet allows the user to ca l cu l at e a benefit/cost rat i o t o d et ermi n e the effec t iveness o f BSP. Much o f the analysis requ ires infor m at i o n o n e may not have access to un l ess BSP had been impl emen t ed o r seriously evaluated. In addition, th e origi nal spreadsheet calcu l a t e d b e nef i t/c os t r at ios based on BSP per f o r manc e in oth e r areas. Thus, the r esulting r at i o m ay not b e accu r a t e i f some var i ables are i n p la ce that would adversel y affect BSP p erf orman ce. Th e cha nges made t o t he s p re a dsh ee t makes i t pos s i b l e for t h e u ser to detenmine if proceed i ng w i t h BSP and d eterm i n ing t h e v a r i ab les for t he b enefit/cos t ra t i o i s a p r actica l opti on for a par ticu l a r app l i c a tion a r ea Pre-Implementation Checklist The p r e -imple m en t at i on c hec kl is t cont a ins a series of questi ons t o assist agenci e s with dete r m i n i n g t he s u i tabil ity o f BSP. I n mos t cases changes can be made to i ncrease the v i abi lity of prio r i ty. The check list i s essential i n d e tenmi n i ng cr i t ica l f ac t ors that m a y i m p e d e BSP's be n efits The firs t two q u e s ti ons add ress e xpress bus serv i ce issues w i t h BSP. For i nstance, offpeak express serv i ce allows for uncondi ti o n al p r iority -saving t i me f o r transi t but m ay ca use delays i n o ther serv i ce a reas. Fars i de bus stops also ha v e an effect on BSP. For in stance, t h e e xist ence of far side s t ops (or at least a willin gness to re l ocat e s tops to far -side) increases the potential ben e fits of BSP. Alth oug h most BSP systems attempt to mini mize delays to nontra nsi t traff i c, i n cases with h i g h l y saturated cross s treets (1. 0 v/c r atio), priori ty can caus e compl i cat i ons. The v/c r atio measu res th e flow of traffic i n re l a ti ons hip to a t heor et i ca l d ete r m i nat i o n of t he capac ity. Subseq u ent l y a n agency w ith a l a rge n umbe r of h i gh l y satu r ated cross-st r eets may want to re-co n s id e r pri o r i ty, at least at t hose intersections. A l thoug h the exis tence of satu r ated c r o s s streets does not preclude BSP, some concessions m ust b e m ad e H e n ce t h e add i tiona l recomme n dation s a t the end of t h e c h ecklis t sugges t that the active priori ty m e t hod has limi ted potential with t h e h i gh l y saturated cros s streets 2:5
Relat e d to the saturated cross-streets a r e h e a v y v o l ume i n t ersect i ons i n the n etwo r k A n e twork t hat conta ins sever a l high l y conges t e d i ntersections may lim i t t he pote n t i a l of BSP. Although improvements m a y be noted, th ere may not b e enough benefits to jus ti(y implementatio n o f BSP. I f t he other var iables are in pla c e t o m ake priority a viable opti o n, the 'intersecti o n s pecific gu id elines section ( e xp la in e d in fu rther detail i n subsequen t section) can be used t o gaug e the i mpact of co ngest ion Many BSP systems i mp le m e nt t echniques to addre s s two transit veh icles simultaneously requesti ng prior i t y at the same intersection Howeve r i f this scenari o is like l y to occur f r eq uen t l y BSP has lim ite d poten tial. Many agenc ies award p riority to the rou te wit h t he most passengers, an effective t echnique i f simu ltan eous req u ests o cc u r r e l a tively i n frequently. H owever, if t his scenario is l ikely to occu r on a r eg u lar ba s is, a negative i mpact w ill arise on t h e d i re ct ion not r eceiving p rio ri t y. T h ere f o re thi s is a s i gn i fic ant var i a bl e t o cons id er. Finally, i f AV L techno l ogy i s a lread y i nstalled, BSP beco mes much easi e r to implem e n t and m or e beneficial. Based on th e users i npu t s to the series o f que stions, the spr eadsheet reco mmends t he pote n t i al implementat ion of BSP: J.> I f th ere a r e onl y one or two answers t hat make p r iori ty beneficial, the n t he o u t put of "changes needed for pri ority" appears. J.> I f the re ar e th ree benefic i a l answ ers t hen priority i s somewhat recomme n d e d. > Four a d vantageous answers out put a r eco m men da t ion to pursue p r i o r i t y. > Five or s ix pos i tive answers r e sults in pri ority be i ng s trongly reco mm ended. No r ecom menda tion is made t o not pursue priority as, by analyzi n g the answers, a few changes (such as bus s t o ps ) may turn p riority i n t o a ben e ficial p ro pos i tion 26
Table 4Pre-Implementation Checklist Point System Pre-Implementation Checklist If Answer is Yes If Answer is No Express bus service? P l us One Point No Point Express bus service during off peakl Plus One Point No Point Farside bus stopsl P lus One Point N o Point Highly saturated cross streets over 1.0 No Point P lus One Point vis rat iol Heavy volume intersections in the No Point Plus One Point network? Many instances of two transit vehicles N o Point P lus One Point a p proaching one intersedion? Do you have AVL technology installed? Plus One Po i nt N o Point Intersection Specific Guidelines This section rec ommends a priority strategy based upon the saturation leve l for a specific i ntersection (14). For example, if the saturation leve l is at 1 0 (v/c ratio) then priority may not be very effective. If the saturation leve l is un d er 1.0 and above 0.8, then 1 ().second priority is the bes t option. 10-second prio r ity works best i n this situation because any longer priority would result in too much delay. With a satu r ation leve l ove r 0.25 and less than 0.8, priority with some limits can be inst i tuted. Due to the l essened saturation, more priority can be i mplemented without causing any delay Un l imited priority can be considered with a saturation level under 0.25 because there is less possibility of delay. Table 5 Recommendations Based on Saturation Level Saturation Level .stral .. Under0.25 Un l imited Priority Over 0.25/Under 0.8 Priority With Limits Above0.8 1 CHecond Priority 1.0 Priority May Not Be Effective Operational and Design Guidelines If an agency decides to investigate or implement BSP, the operational and design guidelines provide another tool to help identify the type of priority that should be Implemented based on 27
certain preferences. Al t hough on e priority system may far outperform others i n some areas, i f the variables t hat m ake t hat system e f fec t ive a r e not in place, th e system may un der pe rfo r m. I n a d d i t i on, these guidelines consider var ious priority techn iques and methods, and o ffer suggestion s r egarding tec hnology or s i tu ation need s nec e ssary for that t ec h n iq ue to b e effect i v e and how it impacts the othe r technologies or metho ds ne ede d. Table f>Operat i onal and Design Guide lines 1. Are you willing to change operating characteristics for BSP? I f YES t hen implem e nting an active, re a l If NO then a schedu le based, fix e dtime time system is adv i sable. s trategy is best. Although t his does red uce t he benefi t s o f BSP, it r eq uires l ess adap t at ion of t he curren t s y s t em. 2. Are you willing to re l ocate stops to the far-side if not currently there? I f YES then the active priority system will If NO the n that l i m its the benefits possible have the most benefi t s with a n activ e o r io ritv s v s tem. 3. Do you want an active prioritv svstem1 I f YES then a realti me system with AV L i s I f NO then a passi ve pri o rity syste m is optimal. Sho u ld be noted that act i v e p r iori t y n eede d Note that t his limit s potenti a l has i n some c a ses bee n less effective i n benefits. areas with h ighly sat u rate d cross s treets. 4 Do you want unconditional priority? I f YES then n o te th a t t his is b est sui t ed to area s with low d e gree of cross street congestion. Uncond itiona l p rior ity is best sui t ed t o specifi c situations such as express bus service d uri ng offpeak Uncond i tional I p rior ity with no I i m its can cause de lays. I f NO th en you need condi ti o nal p rio r i ty It i s suggested t o use a wei ght i ng syst e m 5. Do you want to provide compensation? Selecting YES means you ha v e a concern Selecting NO means that non-transit de lay i s that BSP could de l a y non transit. not as sig n i fican t an issue, perhaps due to low cong e s tion l eve ls 6. Do you want to provide 11reen recall? Selecti ng YES is advisable i f y ou are o pe rati ng a passive pri o r ity system. Select ing NO is advisable i f you are ope r a t ing an active real time svstem. 28
EI'FECI1nilVESS OF Bw-S/GNALI'KFOIUTY 7. Do yo u want to provide synchron ization! Sel ec ti ng YES suggests you can p ro gra m s i gna l t i mings a l ong an arte r i a l t o prov i de m i nima l stops. Sel ecting NO means tha t a l ong sect i o n of synchro n ization i s not pract i cal along these arteria l s 8. Do you want to offer p hase s u ppression! Select i ng YES mean s cross-street saturation i s Sel ect i ng NO means crossst r eet saturat i o n is at low leve ls. too high. 9. Do you want to offer real t ime priority! I f YES t h en AVL and active pri o r ity is I f NO then a fixed-time or sch edu le-based advisable. system is recommended. 10. Do you want t o use a headway-based system? Sel ect i ng YES i s advisab l e to avo id buses Selecting NO is advisable if r idershi p and bunchin g up. f r eque ncy of buses is h igh enough to j us tify o n e bus on t he sam e r oute followi ng rel at ively close l y to another 11. Do yo u wish to apply different "weights" t o certain situa t ions? Selecting YES suggests you are using a real -Sel ect ing NO means you are using a fixedtime, active p ri o r ity system and have h i g h time schedu l e and do not have realt i me le v els of conges t ion orior i tv. 12. If using real-time do you want a l oop detector/tra n spond e r system or an AVL based system? Sel ec ti ng a n AV L s y s t em makes o n lin e i nf orm a ti on to transi t users and i mple m e nting var iables to when pri ority is offe r ed far more practical. 13. Do you want to offer priority only when running beh i nd schedule? Sel ect i ng YES l i m its benef its of r educi n g S e lect i ng NO means that schedu les do not travel time. 1\eed to b e update d and changed based on improv e d t rave l times. CONCLUSIONS to an bus priority system is to e nsure transit vehicles i n all directions can be assistw without non-transit to an y in the process. The degree to whi c h priority can be afforde d is limited by the poten tial d e lay t o o ther transit ve hi cles and non transit 29
vehicles that a r e delayed a s a r esul t of a ny gran ted p rior i t y. Most systems ar e able to p rov i de ampl e p r i ority to one direction. There is a lack of documented test results regarding the affect of BSP on non transit vehicles, especially for projects that are in operation in North America. The in t erv i ews wit h p l anners and eng i neers i nvol ved in installing BSP projects revealed that BSP, although i t has existed as a conc ept for abou t thr ee decad es, is still in re l ativ e i n f ancy i n terms of widespread i mplementation and study i n the United States. Performance monitoring and benefits assessment studies for BSP are scarce. The perception of traffic engineers is that, provided the timing strategy if tailored for the area, non-transit vehicle delay is not significant. BSP offers fewer benefits in areas with extremely high bus volumes or very light traffic. Some t est r e s ults conclud e d t hat th e overall b e nefits gai n ed by buses were not sufficient to counter the dela y t o othe r ve h icles. T he most appropr i ate pri o r ity m et hod may be one t hat combines t h e variou s elem ents of existing p r iori t y t e chnique s Overall, suff i ci e n t measure s m u s t b e tak e n t o pre v ent traffic opera t i o n s b eing negative l y affected b y bus priori t y Re al-time control strategy is the best solution a lth ough any system mus t be desi g ned w i t h t he part icu l a r area needs i n mind. A s the traffic engineers i nterviewe d agreed, t he eng i neers must deve l o p cont r o l strateg ies t hat a r e appropriate for the g i ven i n terse ctions in their a r e a. T he l i terature search d i d re veal the importance of impl
Jii1EC11PENE$ OF Ba; SICMil I'RIOKI1Y REFERENCES 1. Abde l fattah, A.M. s us De lays i n Mi. xed T raff i c: A Challenge for I nte l ligent T ranspor t ation Systems: 6 9 "' Annua l Meet i ng ofThe Institute ofTrans porta ti on Engineers UTE). August 1 999. 2 AlSahili, K Tayl or, W C "Eval uat i o n of B u s P r io rity S i g n a l (BPS) Strat egies in Ann Arb or, M i ch igan: Transp o rtat ion Research Record 155 4 TranspOrtation Research Board, 7511' Annual Meet i ng, )an 7 11, 1996 Wash i ngto n D.C. p p. 7 4-80. 3. Atherl ey, Ronal d. Rai nier Avenue S TSP F i e l d Eval uat i on. K i ng Coun t y Departme n t of Transportatio n 2000 4. Bowen, G.T sus Prior i t y i n SCOOT.' TRL Report 255, T r a n sport Resear c h. l abo r atory, UK 1 997 5 C h en, W., )argees, G., Drane C -R. New Approach for B u s P r iority at Signal ized I ntersections: T r anspo rt Proceed ingsConfere nce of the Aust ra l ian Road Research Board 1998. ARRB Transport Res ea r ch l t d Vermont, Australia. p p 39199 6. Crabtree, M.R. and Vincen t R .A. "Bus Priority i n MOVA.' lEE Conference Publication 4 54. 9 ' International Con f e r ence on Road Transport I n for mation and Contro l 21" 2 3 ' Apri l 1 998. L o n don. pp 85 89. 7. Fox, K., Ha ibo C., Montgom ery, F., Smit h M., ) ones S. "Sel ected Veh i cle Pri ority in the U TMC Environ m en t. I nst i tute for T ran sport Stu dies, Univer s ity of Leeds. 1998. 8 Fox, K., Montgom ery, F.O., Shepherd, S.P., Smi th, C., J o nes, S. and B i o ra, F oRtVE II proj ect P R IMAVE R A 2: B u s p riority i n SCOO T a nd SPOT us ing TIRIS.' T ra ffi c Engineer ing Co n trol 36. j une 1995. 9. Garrow M. and Mach e meh l R. Deve lopm ent and Eval ua tion o f T r an si t P ri o r i ty Strateg ies. Public journa l ofTr ans p o rt ation Vol. 2, No.2. 10 Hou nsell, N., Mcl eo d F. Aut omat ic Veh i c l e Locat i on and Bus Prior ity: The London System.' World Trans p ort Res earch: Proceed ings of 8" World Confere nce on T ransport R esearch Meersma n H Voo r de, E.V.d.V. Winkelmans, W Amsterdam : Per gamo n 1999. pp.279-92 11 Hunter-Zawors ki, K.M., K l oos, W.C. a n d Danahe r A.R. su s Priority at Traff i c S i gnals i n Port l and: The Powell Bo u leva r d P i lo t P roj ect." T ransporta tion R esearch Record 1503. 1995. pp. 29 38 Jl
EmCT1T/ENEf.f OF BUS SiCNAI. I'K!OJUTY 12. jacobso n,). and Sheffi, Y. NAn alytical Model of T raffic Delays under B u s Sign a l P r eempt i o n: Theory and Applica ti on T rans p ortation Research B, Vol. 15B, 1981. 13. J epso n D., Fer re ira, L. "Assessing Travel Time Impacts of Measures t o Enhance Bus Ope r ations." Road and Transport Research, Vol. 8, No. 4. 1999. pp. 41-54. 1 4 journa l of Publ i c Transportation Vol. 2, No.2 99. 15 K loos, W.C., Tumer, K.R. N l mplementing Traf fic Signal Priority for Buses i n Portland." Transportat ion Frontiers f or the N ex t Millennium: 69" Annua l Me eti ng o f the Instit ute o f Transporta tio n Engineers. las Vegas, Nevada Aug 1 4, 1999. pp. 5. 16. larson, R. M. Veh icle D e t ect i on Phase Il l: Passive Bus Detector/Inte rsection P r i ority System. Washi ngton D .C.: Federal High way Administration. C lassifier Handbook, Oct 1977. 17 Lin, C.$. Liang, P., Scho n feld P larso n, R. Adapt ive Contro l of Transi t Operat i ons." Novem b e r 15 1995. 18. liu, R., C l ark, S., Montgomery, F., Watling, D. NTraf fi c Management Measu res f o r Gu i ded Bus Operat i on." Procee dings of the Conference on Traf fic and T ransportat i on Studies N orthern j ia otong University, Beijing, Ch i na. Jul y 27-29, 1998. pp. 18 19 1 19. Miller, C., Deuser, R W att lewort h, ). Wallace, C. Safety Evalu a tio n of Priority Techn iques for High O c cupancy Vehicles Washi ngton, D .C.: Feder a l Highway Adm i n i stration. Final Report, Oct 1976 Dec 1978. 20 O'Mea r a M. NHow Mid-Sized Cities Can Avoi d Strangula tion: World Watch Magaz i ne. Sept/Oct 1998, Vol. 1 1, No. 5. Washington D.C.: Worldwatch I nst i tute. WWW: http://usin f o.state.gov / p roductslpubs/ l ivcomm/two htm. 21. Rothenberg, M. ). Priority Treatm e nt for H i gh Occupancy Vehicles: Project Status Report Washington, D.C.: Federal Highway Admin i s tration Interim Report May 1976-Ma r ch 1977. 22. Sunkari, S.R, Beasley, P S U r banik II, T ., F ambro, D.B. 'Model to Evaluate the Impacts of B us Priority on Signalized I ntersections." Transportation R esearch Record 1494 1995 pp. 1 17-23 Taub, A NRollin g On: System lets Traffic lights Wave Buses Through." NY Times. Sept 9, 2000. J:Z
24 Thompson, A a n d N ich o lls, J. S ignal PreE m p tion Us i ng Tr af fic Control Networ k s." T r affic Tec hnol ogy Internatio n a l. 1997. 25. Urba n T r a ffi c Cont r o l a n d Bus Prio rit y System O p e r ations a n d Mai n t enance -Final Report Washin gton, D.C. : Fed e r a l H i ghway Ad mi n i strat i on. F inal Report June 1976. 26. Wees ner,). F., Meye r kord, K.M. P i l o t Bus P r i o r i t y Syst em f o r Urba n Arter i a l s." Transportation Frontiers fo r t he N ext M i llenn i u m : 69" A nnu a l Mee ting of the Institute of T r a n spo rtation Engi neers Las Vegas Nevada. Aug 1 4 1 999. p p 1 2 27. K i ng County Metro F inal Rep ort : Rain i er Ave n ue South Trans i t Signa l Pr iority F i eld Eval uat ion. ) a n 2001. 28. User s Manu a l fo r SCRITS. Science Appl i ca ti o n s In t ernat i o nal Corp ora tion. http://pl an2op. f hwa .dot.gov/toolbox/s crits _man ua l.pdf JJ
APPENDIX A BSP Spreadsheet (Disk Attached) A-1
bus service d urin g off peak? bus stops? saturated cross streets over 1 .0 v/s ratio? vol ume intersectio n s in tho Mtv,:o rk? o f two transit vehicles approaching one i n t ersection? have AVL tech n ology i nstalled? IP1rio1rity recommended? per weekday on priority ro!)tes average bus speed on arterials (mph ) o f bus trsvel time attributable to sig.na l delay % reduction in signal delay from p re-em ption minutes per mil e for buses without prio rity minu tes per mile for buses v.Ath priotity bus speed priOlwiiiZ