Ride Comfort-enhanced Optimal Decision and Planning Strategy at Urban Signalized Intersections Based on Hybrid Model Predictive Control

With the support of Vehicle-to-infrastructure (V2I) communications, autonomous driving technology has been developed to drive various traffic conditions. However, it remains challenging for autonomous vehicles to drive urban signalized intersections safely due to hazardous situations, such as abrupt stops by human drivers and tailbacks at junctions. In these complex mixed-traffic environments, autonomous vehicles' maneuver decision and motion planning modules should prioritize both driving safety and passenger ride comfort. This paper proposes a ride comfort-enhanced optimal decision and planning approach based on Hybrid Model Predictive Control (HMPC) that takes into consideration traffic conflicts at urban signalized intersections. The proposed method optimizes ride comfort while ensuring safe passing through intersections. After formulating this optimization problem as a Mixed Integer Quadratic Programming (MIQP), HMPC was implemented to calculate the optimal maneuver and desired motion considering constraints. The evaluation through computer simulations demonstrated that the proposed method improved ride comfort while maintaining driving safety. The results indicate the effectiveness of the HMPC-based decision and planning approach in improving ride comfort and safety at urban signalized intersections.