Abstract:
The article addresses the problem of constructing time-optimal routes for one or several moving objects in a dynamic environment with obstacles. We propose an approach based on the optical-geometrical analogy and the principles of Fermat and Huygens, which reduces the problem to solving the eikonal equation. A series of algorithms (including modifications of the fast marching method) is developed. They allow one to construct collision-free trajectories and correct them when the situation changes. The effectiveness of the proposed methods is confirmed by computational experiments, demonstrating their superiority in route length and computation time compared to existing approaches such as CFM, PRIMAL, and DHC.