Optimization and Empirical Analysis of Path Planning Algorithms for High-Speed Autonomous Driving Scenarios

Abstract

High-speed autonomous driving on highways demands precise and low-latency path planning to ensure vehicle safety and operational efficiency. This study proposes a novel high-speed path planning algorithm that integrates sampling-based optimization with behavior prediction mechanisms. The algorithm incorporates vehicle dynamics constraints and a real-time risk assessment model to enhance decision-making capabilities under high-speed and complex traffic conditions. Comprehensive experiments were conducted using the HighD dataset. Comparative analysis with baseline algorithms, including Rapidly-exploring Random Tree (RRT) and Hybrid A*, demonstrates that the proposed method significantly improves path safety, planning efficiency, and driving comfort. The results highlight the algorithm's potential for practical deployment in engineering applications of high-speed autonomous driving systems.