一种面向多智能体群集的避障算法

多智能体群集中的避障问题是研究的难点问题，每个智能体需要安全避开障碍物并朝着目标点前进.根据现有的基于人工势场函数的群集算法，提出一种改进的具有避障能力的群集算法.在该算法中，将障碍物等效成虚拟智能体进行避障.智能体感知到障碍物后，不是立即采取避障措施，而是将智能体的速度方向和目标点考虑在内，根据智能体不同的速度方向和目标点的位置，采取不同的避障措施.经理论分析与实验验证，表明所提出的算法能够有效地躲避障碍，并且在避开障碍物后更快地达到群集.

A Multi agent Flocking Oriented Obstacle Avoidance Algorithm

The problem of obstacle avoidance is important in multi agent flocking. Each agent should avoid obstacles safely， and then moves toward the target. Based on the existing artificial potential field flocking algorithm， an improved algorithm with obstacle avoidance capability was presented. In this algorithm， the obstacle was equivalent to a virtual agent for obstacle avoidance. Obstacle avoidance were not taken immediately when the agent perceived obstacles， but take the speed direction of the agent and the target point into consideration. According to different speed directions and position of the target， different obstacle avoidance measures will be taken. Through theoretical analysis and experimental verification， obstacles could be avoided efficiently based on the proposed algorithm， which can make the flocking faster.