Melvin's digital garden

Urban traffic modelling and intelligent intersection control

[2016-09-20 Tue 19:13:20] speaker: Yang Bo, IHPC event: Science in the Cafe

complex systems

  • many components that interact with one another
  • simple, decentralized interactions
  • eg. flock of birds, bees, stampede in crowded area, landslide,
  • spreading of disease, occurrence of traffic jam

next generation urban mobility at IHPC

  • modelling of traffic flow

one dimensional driven system

  • nearest neighbour
  • anisotropic (only consider vehicle in front)
  • non-linear
    • dissipative media

empirical observations

  • density (veh/km) vs flow (veh/hour)
  • free flowing -> traffic jam beyond a critical density
  • time vs distance vs flow
  • phantom jam due to human reaction time

microscopic models

  • a_n(t + T) = c v_n(t + T)^l h_n^-m dv_n
  • … many other possible models
  • a_n: acceleration of nth vehicle
  • h_n: headway
  • v_n: velocity of nth vehicle
  • dv_n: relative velocity of vehicle in front
  • difficult to compare model and empirical results

video-taping of traffic flow with high frame rate camera

  • along Queensway

traffic light optimisation

  • GLIDE system in Singapore

  • adaptive to real time demand

  • creation of green wave

  • difficult to adapt to fluctuating demand

  • minimize average stational time per car

lightless intersection control

  • traffic lights is centralized control
  • feet-off -> hands/feet-off -> brain-off
  • lightless control for feet-off vehicles
    • beacon that collects information of approaching vehicles
    • broadcast this to oncoming vehicles

Links to this note