Compliance-constrained resilient system optimal trajectory planning for CAVs at on-ramp intersection with multiple lanes

This paper addresses traffic merging at highway intersections (labeled as OMM intersections) where mainline traffic with multiple lanes and on-ramp traffic converges, which often represent traffic bottlenecks causing severe traffic congestion and safety issues. To do that, we developed a Compliance-constrained Resilient System Optimal Trajectory Planning (CR-SOTP), which is devised as an event-triggered rolling-horizon system-optimal trajectory planning and replanning coupled scheme combined with a compliance incentive instrument. Specifically, we developed a bi-level optimization model (C-SOTP-BL) with the upper level (SOTP-MINLP) formulated as a mixed integer nonlinear program to generate a system optimal trajectory plan for all CAVs within a trajectory planning area around an OMM intersection, subject to the compliance constraints from the lower level (C-MINLP) optimal incentive model built upon game theory. To adapt to real-time implementation, we developed a parallel computing-aided relax-enforcement-refinement (P-RER) solution method to efficiently solve the C-SOTP-BL. Thanks to the advanced computing performance of the P-RER, the CR-SOTP approach gains resilience through the coupled rolling horizon framework, which recursively plans and replans CAVs’ trajectory by solving the C-SOTP-BL with a descending scale to respond to random disturbances. Numerical experiments employing NGSIM data confirmed the effectiveness of the solution algorithm P-RER in real-time implementation, outperforming existing commercial solvers. The CR-SOTP can generate efficient and resilient trajectory plans to ensure stream-wide safe and smooth traffic merging at an OMM intersection subject to random disturbances. The experiments noticed notable enhancements in social welfare, demonstrating benefits outweigh the associated incentive cost for ensuring CAV compliance. Overall, we claim that the CR-SOTP strategy significantly improves stream-wide traffic safety, efficiency, and sustainability at multi-lane on-ramp intersections.