216 12 15 Adaptive Ramp Control The Gardiner Expressway Case Study Kasra Rezaee Department of Civil Engineering Congestion Solutions Intelligence for managing both supply and demand More Supply Less Demand 1
216 12 15 Control Approach Characteristics 1. Pacing Beats Rushing 2. Real Time: Closed Loop Optimal Control 3. Role for Artificial Intelligence 3 1. Pacing Beats Rushing 4 2
216 12 15 2. Closed Loop Optimal Control Optimal Control Actions Traffic Network Optimal Control Law/Policy Real Time Monitoring and Measurements Always measuring and mapping system state to optimal actions Better chance to be on top of changes But: How to get the optimal control policy Possibly let the controller learn it: Machine Learning and AI 5 3. Artificial Intelligence: Self-Learning the Optimal Control Law 3
216 12 15 AI: Reinforcement Learning Environment Action Agent State Reward RL: Illustrative Demo Source: https://www.youtube.com/watch?v=dcjbk4m1g6i Baher 4
216 12 15 RL: Another Illustrative Demo Baher Ramp Metering Why Ramp Metering? 1. Pacing demand: avoid congestion due to demand exceeding capacity and resulting capacity breakdown 2. Avoid blockage of exit ramps 3. Influence route choice behavior 4. Enhance traffic safety: less congestion safer merging 5
216 12 15 Congestion Avoidance Off Ramp Blockage q in q cap d q cap (q 1 cap d) d q in q cap γ q in d 6
216 12 15 Impacting Route Choice Summary of Today s Status Quo Under-utilization of freeway corridors and networks due to recurrent and non-recurrent congestion: high demand creates flow breakdown and congestion causing loss of capacity downstream of congestion (empty stretch ahead) off-ramp blockage (stuck on the freeway) suboptimal utilization parallel arterial reduced safety increased pollution 7
216 12 15 Ramp Metering Benefits in the Literature 3% more vehicles served during rush hours Improved service for all users Reduced urban network load > 5% reduction of total time spent Efficient response to incidents Increased traffic safety Decreased fuel consumption and environmental pollution Why NOT RM? Fallacy: RM benefits people on the freeway at the expense of those entering from the ramp RM causes traffic to use the surface street, i.e. benefit freeway at the expense of surface streets Fact: ramp queues do not mean dis-benefit to surface streets. As the overall throughput of the freeway is improved, more surface street traffic can now use the freeway. i.e. benefit both. Coordinated RM does not penalize later entries (see Gardiner case) 8
216 12 15 Ramp Metering Categories Fixed Time Metering: Mainly off line Based on historical demand Not responsive to real time traffic dynamics Traffic Responsive: Realtime Regulator Approach: e.g. ALINEA Optimal Control Metering e.g. RL Control of Multiple On-ramps 18 9
216 12 15 Ramp Metering the Gardiner Scenarios Base Case RLRM-I : Isolated RLRM-IwQO : Isolated with queue override RLRM-C : Coordianted ALINEAwLC : ALINEA with linked control Gardiner Expressway westbound in Toronto from DVP to Humber Bay Traffic from Don Valley Parkway Humber Bay 19 Network Total Travel Time Total time spent on network (Veh.hr) 12 1 8 6 4 2 48% 45% 5% 35% RLRM- ALINEAw BaseCase RLRM-I RLRM-C IwQO LC TTT 1276 536 5665 514 666 TTTml 6998 4147 4768 4249 461 2 1
216 12 15 Downtown Origins Travel Times Travel time from different origins to west end of network 14.34 16.65 16.27 12.13 11.79 9.32 8.81 8.75 9.47 8.68 6.34 6.19 5.47 6.7 6.57 6.75 BASECASE RLRM-I RLRM-IWQO RLRM-C DVP Jarvis York Spadina 21 Local vs. Coordinated Base Case (Jameson Closed) RLRM-I traffic speed RLRM-IwQO traffic speed RLRM-C traffic speed Westbound Direction p ( ) 1 8 6 4 2 1 8 6 4 2 1 8 6 4 2 1 8 6 4 2 14 16 18 2 Time of day (hour) 14 16 18 2 Time of day (hour) 14 16 18 2 Time of day (hour) 14 16 18 2 Time of day (hour) 3 25 BaseCase (Jameson Closed) Spadina York RLRM-I RLRM-IwQO Spadina York RLRM-C (coordinated) Total Vehicles in Queue 2 15 1 5 13 14 15 16 17 18 19 2 2113 14 15 16 17 18 19 2 2113 14 15 16 17 18 19 2 21 13 14 15 16 17 18 19 2 21 Time of day (hr) Time of day (hr) Time of day (hr) Time of day (hr) 22 11
216 12 15 Conclusion Least overall TTT Best overall travel times from all ramps Reduced and balanced queueing 23 Thank You! 12