This paper presents a new algorithm for realtime collision avoidance of two mobile robots based on a heuristic risk evaluation. The main idea of this algorithm is to express the collision avoidance problem of mobile robots in terms of the velocity vector modification. A distance function obtained from the state information of two neighboring robots is introduced to evaluate the risk of collision. According to the encounter situation of robots resulting from the risk evaluation, the desired velocity vectors are determined for each robot. The robots can avoid collision from each other by following their desired velocity vectors.
In a simulation study, the proposed algorithm is applied to various colliding situations of two robots and achieves acceptable performance. The feasibility and the effectiveness of our method are also discussed through the simulation results.
