Enhancing and Evaluating a Decentralized Cycle-Free Game-Theoretic Adaptive Traffic Signal Controller on an Isolated Signalized Intersection

• Published in: 2024 IEEE International Conference on Smart Mobility (SM) pp. 216 - 221
• Main Authors: Shafik, Amr K., Rakha, Hesham A.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 16-09-2024
• Subjects: adaptive signal control; cycle length optimization; Delays; Fuel economy; Game theory; intelligent transportation systems; Length measurement; Nash bargaining; Optimization methods; Performance analysis; Phase measurement; signal timing; Switches; Traffic congestion; Traffic control; traffic signal controller; Transportation
• DOI: 10.1109/SM63044.2024.10733500

Effective traffic signal control strategies are essential for optimizing urban traffic flow and adapting to dynamic traffic patterns. This research evaluates the performance of the Laguna-Du-Rakha (LDR) developed cycle length optimization strategy and enhances a decentralized cycle-free Nash bargaining (DNB) adaptive traffic signal controller. The enhanced DNB controller utilizes traffic measures, such as queue lengths and approach densities, to optimize phase sequences and splits, considering each signal phase as a player in a game-theoretic framework. Both methods are implemented at a simulated isolated intersection in Toronto, Canada. Comparative performance analysis demonstrates superior performance of both optimization methods over the state-of-practice, the Webster method. Specifically, the optimized cycle length method achieved a 25.6% reduction in delay and a 5.9% reduction in fuel consumption. The DNB controller produced a 37.7% reduction in delay and a 7.4% improvement in fuel economy, alongside significantly decreased queue lengths at intersection approaches for both methods. The optimized cycle length method mitigates the overestimated cycle lengths produced by the Webster method, leading to enhanced control performance. Furthermore, the application of game theory to traffic signal control provides dynamic switching behavior that adapts to fluctuating traffic conditions, thereby reducing traffic congestion and improving fuel economy in urban networks. As such, this research contributes to the development of smarter and more responsive urban traffic management systems.