Research on Obstacle Avoidance Path Planning for UGV in Complex Environments Based on Fusion Algorithm
Zhenpeng Jiang, Qing-Quan Liu, Di Zheng, Ronghua Shi, Yu Wang · 2024
The traditional A* algorithm is widely used in global path planning for unmanned vehicles due to its advantages such as simple computational process and short path planning. However, the algorithm has some deficiencies in search efficiency and inflection point redundancy. Therefore, we improve the traditional A* algorithm and combine it with the dynamic window method to enhance the path planning capability of unmanned vehicles. In the process of optimising the A* algorithm, firstly, the evaluation function is improved, and secondly, the key inflection points are screened out and the redundant inflection points are removed, so as to improve the retrieval speed and path smoothing degree of the A* algorithm. Between each pair of optimised key inflection points, the dynamic window method, which takes into account both speed and safety, is used for local obstacle avoidance, so that the unmanned vehicle achieves real-time obstacle avoidance on the basis of ensuring the optimal global path. The experimental results show that the fused algorithm can achieve local optimality while ensuring global path optimality, and successfully realise real-time obstacle avoidance.