Event triggered rolling horizon based systematical trajectory planning for merging platoons at mainline-ramp intersection

Abstract

Abstract Two traffic steams merging at a mainline-ramp intersection often leads to recurrent traffic accidents, delays, oscillations and pollution. Existing literature shows that we still lack efficient methods to address this issue considering stream-wide traffic optimality, due to the model complexity and computation challenge. To make up these gaps, this study developed an event-triggered rolling horizon based systematical trajectory planning (RSTP), which expects to safely and smoothly merge two platoons (pure or mixed CAV) in a well-designed trajectory control zone at a mainline-ramp intersection. Specifically, the trajectory control zones on both mainline and ramp are divided into pre-merging, virtual-merging and post-merging subzones. The RSTP is only activated and then generates an optimal trajectory plan for the CAV platoons in the two pre-merging subzones by a mixed integer non-linear program (MINLP), when a new CAV approaches to the virtual merging subzone either on the mainline or ramp. The RSTP combined with a real-time correction mechanism enables the two platoons to smoothly merge at the end of the virtual merging subzone against certain level of disturbance, aiming to maximize stream-wide traffic performance. Using the unique features of the problem, a heuristic algorithm (MIA) is developed to efficiently solve the MINLPs. The numerical experiments built upon the NGSIM data show that the RSTP achieves expected performance under both pure CAV and CAV-HDV mixed traffic flow. The MIA can solve the MINLPs more quickly with a competitive solution quality than the existing solvers do in GAMS. The RSTP performs better under a traffic flow with higher CAV penetration. The real time re-planning mechanism works effectively to correct mild trajectory deviation occurring to a few CAVs.