A Computationally-Efficient Algorithm to Enable Joint Optimization of Connected Automated Vehicles’ Trajectories and Signal Phasing and Timing in Coordinated Arterials

Connected Automated Vehicles (CAVs) have the potential to improve traffic performance on urban roads, especially at signalized intersections. Several studies have developed signal control strategies for isolated intersections considering CAVs’ capabilities. However, few studies have been conducted on complex applications such as arterials. Therefore, there are still several questions that need to be addressed. To our knowledge, existing strategies reported in the literature require long computational times which are not feasible for field implementation. This study proposes a novel computation-efficient joint optimization algorithm for signalized arterials to facilitate deployment. This algorithm enables both trajectory and Signals Phasing and Timing (SPaT) optimization in a fully automated environment. In the first level of optimization (trajectory optimization), the vehicles’ optimal trajectories are determined to form platoons with vehicles following the pre-determined saturation headway. The trajectories are built in a piecewise fashion respecting safety and comfort-related parameters. In the second level of optimization, the SPaT is adjusted according to a novel heuristic that weighs the deviation between the ideal and actual travel time for all vehicles. Comparing the trajectory optimization and joint optimization methods, both perform similarly in low-demand scenarios. However, for moderate-to-high-demand scenarios, the joint optimization method outperforms, reducing average travel times by 3 to 16%, and improving the green interval utilization. Generally, both the trajectory and the joint optimization methodologies achieve higher performance than the baseline scenario with conventional vehicles and a semi-actuated SPaT control.