A Review of Path Planning Research Based on the Optimisation of the A* Algorithm Using Multi-Source Heterogeneous Data

Abstract

Path planning is a core technology in the fields of intelligent transport, autonomous driving and mobile robotics. The traditional A* algorithm is suited to static environments, but suffers from shortcomings in terms of real-time performance and adaptability in highly dynamic, highly constrained scenarios. With the development of multi-source, heterogeneous perception technologies, the deep integration of data with the A* algorithm has become a key area of research aimed at overcoming this bottleneck. Based on the end-to-end process of ‘data processing—fusion modelling—algorithm optimisation—scenario application’, this paper systematically reviews methods for fusing multi-source heterogeneous data. It constructs a four-dimensional analytical framework encompassing heuristics, cost functions, search strategies and dynamic updates, and provides an in-depth analysis of the optimisation mechanisms of the A* algorithm driven by data. By comparing typical application scenarios such as intelligent urban transport, drone obstacle avoidance and construction site navigation, this paper summarises existing bottlenecks—including low data fusion efficiency and weak cross-scenario transferability—and outlines future development directions such as lightweight data collection, adaptive intelligent decision-making and multi-agent collaboration. This research provides a systematic reference for the theoretical iteration and engineering implementation of path planning technologies.