Improved A*-based Path Planning for Mining Inspection Robots

[Purposes] Aiming at the unstructured environment and narrow local passable area of coal mine underground, and the problems of long search time, multiple search nodes, many redundant nodes in the path, and poor road strength smoothness in the traditional A* algorithm for planning the path, a path planning algorithm for mining inspection robot based on the improved A* algorithm is proposed. [Methods] First, the exponential function of predicted consumption and the sum of obstacle coverage were introduced into the heuristic function of traditional A* algorithm to improve the search efficiency and shorten the search time; Second, the traditional 8-neighborhood search was improved to 9-neighborhood search to avoid useless search and reduce the number of searching nodes; Then the redundant nodes in the path were rejected by the Floyd algorithm; And finally, the path smoothing task was accomplished by using the improved 3rd-order Bessel curve. [Results] The results show that compared with the traditional A* algorithm, the improved A* algorithm shortens the search time by 44.1%, 63.8%, and 84.8%, and reduces the search nodes by 31.6%, 47.9%, and 71% for specific 20×20, 30×30, and 40×40 raster maps, respectively. The proposed path smoothing algorithm is able to reduce the number of path nodes and improve the path smoothing, which makes it more suitable for the mining inspection robot path planning.