An AI tool: Generating paths for racing game

In this paper, we propose an AI tool for generating plausible paths of racers based on the A* algorithm. User can define the race by providing a race course of 3D model and weights of the devised turn and heuristic functions in our system. The search space for path-finding is represented by a grid. Then, the proposed cost map generator automatically generates necessary information of the race course including cost value for each cell in the grid, feature cells, etc. Different from the traditional A* problem, in our research the obstacles are dynamic and there are multiple sources and destinations. Our approach generates the path on the basis of time slots and smoothes it by Gaussian filter. By the devised heuristic function we have a natural racer behavior. Also, our AI tool for path finding has been verified in a horse racing game, demonstrating realistic and exciting racing.