Research On Motion Planning for Intelligent Driving

Abstract

With the ongoing maturation of automated vehicle (AV) technology, motion planning has become a fundamental component to ensure its safe, efficient, and comfortable operation. This paper presents a comprehensive survey of motion planning techniques for intelligent driving systems. It commences with a summary of the theoretical objectives and their theoretical underpinnings in the perception-decision-control architecture. Additionally, it elucidates how motion planning connects high-level behavioural decisions with low-level control commands. The paper examines the following principal algorithms: graph-search methods, sampling-based planners, and spline-based path generation, as well as model predictive control (MPC) under dynamic constraints for trajectory synthesis. Each class of method is analysed concerning obstacle avoidance, compliance with traffic regulations, and dynamic feasibility. On the other hand, it highlights emerging trends in the integration of real-time perception and reactive safety machines. It explores prospective directions for human-inspired planning paradigms, vehicle-to-vehicle (V2V) cooperation, and ethical considerations. Based on the research, this survey lays the groundwork for the development of motion planners capable of tackling the complexities of real-world traffic. Intelligent motion planning thus remains a key driver in advancing automated driving toward a safer, more human-centric mobility paradigm.