VNS-BA*: An Improved Bidirectional A* Path Planning Algorithm Based on Variable Neighborhood Search

The A* algorithm is an effective method for path planning; however, it has certain drawbacks, such as a high number of turns, low planning efficiency, and redundant searches. To address these issues, this paper proposes an improved bidirectional A* global path planning algorithm based on a variable neighborhood search strategy, named VNS-BA*. The new algorithm first employs an 8-11-13 neighborhood search method for node expansion. Then, the bidirectional search strategy is optimized by using the current nodes of the opposite path and the global target point, enabling the two paths to meet in the middle of the map. Finally, redundant turns are removed from the path, and cubic spline interpolation is applied to achieve local smoothing at the turns. The effectiveness of the improved algorithm was validated on different maps and compared with A* and its three derived improved versions. The simulation results indicate that VNS-BA* shows significant improvements in terms of the number of path turns, turn angles, and planning efficiency.