Heuristic Decision Planning for Intelligent Vehicles Considering Multi-Vehicle Interaction Risks

In highly dynamic and complex traffic environments, diverse multi-vehicle interactions significantly impact the accuracy and efficiency of intelligent vehicle decisionmaking. To address this, a planning framework that incorporates interaction risk prediction and heuristic decision-making is proposed. An Interacting Multiple Model risk model with a preview mechanism is designed to compensate for perception delays and accurately predict interaction risks with surrounding vehicles. A temporal decision tree, combined with heuristic rules, multi-objective cost functions, and a periodic replanning scheme, is used to generate smooth, dynamically feasible trajectories that balance decision accuracy and efficiency. Simulation results show that the proposed method improves planning performance under complex and rapidly changing traffic conditions.