Overview 1
Congestion Reduction 2
Related Publications 4
Demos 5

Our research group has pro­posed an agent based-mod­el for Intel­li­gent Trans­porta­tion Sys­tems that we call Agent-ITS or ATS. ATS is based on the fol­low­ing premis­es:

  1. The traffic phys­i­cal space is par­ti­tioned into areas called zones which are man­aged by spe­cial­ized agents called zone man­agers. A zone man­ag­er is respon­si­ble for a) gath­er­ing and ana­lyz­ing traffic data from its zone and extract­ing use­ful traffic infor­ma­tion, b) inform­ing vehi­cle and traffic devices of the cur­rent traffic con­di­tion, c) noti­fy­ing oth­er man­agers of changes that may affect their zones, and d) iden­ti­fy­ing appro­pri­ate glob­al traffic man­age­ment strate­gies to ensure that micro-lev­el behav­iors and inter­ac­tions are con­sis­tent with the glob­al sys­tem behav­ior.
  2. Con­text-Aware Intel­li­gent (CAI) vehi­cles are equipped with agent-based sys­tems and sen­sors that allow them to a) mon­i­tor the driver’s behav­ior, b) com­mu­ni­cate with oth­er vehi­cles, c) com­mu­ni­cate with smart traffic con­trol devices, and 4) inter­act with zone man­agers to obtain traffic infor­ma­tion and guid­ance in real time.
  3. Nov­el dig­i­tal traffic devices con­trolled by a traffic con­trol agent which deter­mines the traffic sign to be dis­played based on traffic con­di­tions.

Through the devel­op­ment of advanced traf­fic man­age­ment algo­rithms for var­i­ous con­fig­u­ra­tions of ATS, our research has shown that decen­tral­ized, coor­di­nat­ed solu­tions improve on the state-or-the-art tech­nolo­gies for traf­fic reduc­tion and urban evac­u­a­tions.

Our approach to con­ges­tion reduc­tion is based on mul­ti-agent col­lab­o­ra­tive algo­rithms for coor­di­nat­ed traffic sys­tems. Inter­sec­tion con­trollers are equipped with agents, i.e., autonomous soft­ware sys­tems which are capa­ble of com­mu­ni­cat­ing and coop­er­at­ing with one anoth­er to achieve an indi­vid­ual or glob­al goal. Our approach is based on real-world traffic para­me­ters and con­straints, and is meant to be imple­ment­ed in exist­ing traffic sys­tems with min­i­mal changes. By default, agents exe­cute a stan­dard tim­ing strat­e­gy. At the same time, they observe and ana­lyze traffic at their inter­sec­tions. At any giv­en time, if an agent deter­mines that its inter­sec­tion is con­gest­ed, it delib­er­ates and defines a tim­ing plan to alle­vi­ate con­ges­tion.

Exper­i­men­tal results on a traf­fic net­work con­sist­ing of 384 road seg­ments, 133 nodes and 40 sig­nal­ized inter­sec­tions show that our agent-based approach out­per­forms the tra­di­tion­al pre-timed and actu­at­ed modes when traffic is heavy.

Exe­cu­tion of a DALI agent
  • Behnam Tora­bi, Rym Z. Wenkstern, and Robert Say­lor. A Self-Adap­tive Col­lab­o­ra­tive Mul­ti-Agent based Traf­fic Sig­nal Tim­ing Sys­tem. In Pro­ceed­ings of the 4th IEEE Inter­na­tion­al Smart Cities Con­fer­ence, ISC2 2018, Kansas City, Mis­souri, USA, Sep­tem­ber 2018.
  • Behnam Tora­bi, Rym Z. Wenkstern, and Robert Say­lor. A Col­lab­o­ra­tive Agent-Based Traf­fic Sig­nal Sys­tem For High­ly Dynam­ic Traf­fic Con­di­tions. In Pro­ceed­ings of the 21st IEEE Inter­na­tion­al Con­fer­ence on Intel­li­gent Trans­porta­tion Sys­tems, IEEE ITSC 2018, Maui, Hawaii, USA, Novem­ber 2018.
  • Behnam Tora­bi, Rym Z. Wenkstern, and Robert Say­lor. A Mul­ti-Hop Agent-Based Traf­fic Sig­nal Tim­ing Sys­tem for the City of Richard­son. In Pro­ceed­ings of the 16th Inter­na­tion­al Con­fer­ence on Autonomous Agent and Mul­ti­a­gent Sys­tems, AAMAS 2018, page 2094–2096, Stock­holm, Swe­den, July 2018.
  • Behnam Tora­bi, Rym Z. Wenkstern, and Robert Say­lor. Agent-based decen­tral­ized traf­fic sig­nal tim­ing. In Pro­ceed­ings of the 21st Inter­na­tion­al Sym­po­sium on Dis­trib­uted Sim­u­la­tion and Real Time Appli­ca­tions, DS-RT 17, page 123–126, Rome, Italy, Octo­ber 2017.
  • Moham­mad Al-Zinati and Rym Wenkstern. Sim­u­la­tion of traf­fic net­work re-orga­ni­za­tion oper­a­tions. In Pro­ceed­ings of the 20th IEEE/ACM Inter­na­tion­al Sym­po­sium on Dis­trib­uted Sim­u­la­tion and Real Time Appli­ca­tions, DS-RT 16, pages 178–186, Sep­tem­ber 2016.
  • Moham­mad Al-Zinati and Rym Wenkstern. Matisse 2.0: a large-scale mul­ti-agent sim­u­la­tion sys­tem for agent-based its. In Pro­ceed­ings of the 2015 ieee/wiciacm inter­na­tion­al con­fer­ence on intel­li­gent agent tech­nol­o­gy, lAT’ 15, pages 328–335, Decem­ber 2015.
  • Moham­mad Al-Zinati and Rym Wenkstern. A self-orga­niz­ing vir­tu­al envi­ron­ment for agent-based sim­u­la­tions. In Pro­ceed­ings of the 2015 inter­na­tion­al con­fer­ence on autonomous agents and mul­ti­a­gent sys­tems, AAMAS ’15, pages 1031–1039, May 2015.

More pub­li­ca­tions avail­able here

Deploy­ment of DALI

DALI Agents Run­ning in the Lab

Eval­u­a­tion of DALI through sim­u­la­tion