(Samah El_Tantawy, Baher Abdulhai, Hossam Abdelgawad (AEA-Scholars

Samah El_Tantawy, Baher Abdulhai, Hossam Abdelgawad

Association of Egyptian-American Scholars

Abstract:

The population is steadily increasing worldwide resulting in intractable traffic congestion in dense urban areas. Adaptive Traffic Signal Control (ATSC) has shown strong potentialto effectively alleviate urban traffic congestion by adjusting the signal timing plans in real-time in response to traffic fluctuations to achieve the desired objectives (e.g., minimal delay). Efficient and robust ATSC can be designed a multi-agent reinforcement learning (MARL) approach in which each controller (agent) responsible for the control of traffic lights around a signal traffic junction. Applying MARL approaches to ATSC problem is associated with a few challenges as agents typically react to changes in the environment at the individual level but the overall behavior of all agents may not be optimal. This presentation introduces the development and evaluation of a novel system of Multi-Agent Reinforcement Learning for Integrated Network of Adaptive Traffic Signal Controllers (MARLIN-ATSC). The MARLIN-ATSC control system is developed to provide a self-learning ATSC using a synergetic combination of reinforcement learning approaches and game theory. MARLIN-ATSC operates in two modes: (1) independent mode, i.e. each intersection controller operates independently of other agents; and (2) integrated mode, where each controller coordinates the signal control actions with the neighboring intersections. The system was tested on three networks (i.e., small, medium, large-scale) to ensure seamless transferability of the system design and results. The large-scale application was conducted on a computerized testbed network of 60 intersections in the lower downtown core of the City of Toronto for the morning rush hour handling 25,000 trips. The results show unprecedented reduction in the average intersection delay ranging from 27% in mode 1 to 39% in mode 2 at the network level; and travel time savings of 15% in mode 1 and 26% in mode 2, along the busiest routes in downtown Toronto. MARLIN has been integrated with an off-the-shelf ATC 1000 controller and tested in the hardware-in-the-loop setup. MARLIN is currently under further development for field implementation and commercialization.

This article was published in 3alamaltanmya

sponsored by Future Builders International Academy

Led by Dr.Maha Fouad