A path-planning algorithm integrated improved A* and D* Lite with bat clustering

In disaster rescue scenarios, the high complexity and level of danger to the environment necessitate the efficient utilization of robotic technology to facilitate rapid search and rescue operations. Addressing this critical issue, this study focuses on optimizing global path-planning strategies. Specifically, the study has improved the classic A* algorithm by incorporating a bidirectional search mechanism and conducted simulation validation on a grid map to assess its performance. Considering the uncertainty and dynamic obstacles in the rescue environment, the study introduces the D* Lite algorithm integrated into the improved A* algorithm to enhance the robot’s adaptability. Concurrently, a movement prediction method based on the historical trajectories of dynamic obstacles has been proposed, aiming to effectively address the challenges presented by dynamic environments. To address the issue of time efficiency in multi-robot systems with multiple rescue points, this study employs the bat clustering algorithm to optimize the clustering of multiple rescue target points. Simulation experiments with the integrated algorithm have verified the effectiveness in shortening rescue paths and time in Matlab. Ultimately, the study utilizes Gazebo simulation software to construct a simulation experiment environment, further confirming the significant improvements in robotic rescue performance achieved by the proposed algorithm.