Seamless Overtaking Maneuvers for Automated Driving: Integrated Motion Planning Based on Hybrid Model Predictive Control

Motion planning for automated vehicles faces critical challenges to achieve safe and efficient maneuvers. Classical modular motion planning, constrained by error propagation and inconsistent optimization objectives across different functional modules, often resulting in overly conservative or irrational overtaking maneuvers. To overcome these limitations, this article attempts to integrate discrete logic decision-making and continuous motion planning in a single optimal problem, and presents a hybrid model predictive control (HMPC)-based motion planner framework. In particular, a novel asymmetric risk fields considering human drivers’ attention is incorporated to guide safe and timely intrinsic decision. The proposed framework eliminates the dependency on external semantic decision-making modules while ensuring efficient and safe maneuvers. Simulation and real vehicle test validate that the proposed HMPC framework is able to outperform the modular baseline, achieving safe and efficient overtaking maneuvers in an interpretable optimization approach.