Multi-robot formation control is an important issue in the multi-robot cooperation field. It is a hot and difficult problem to achieve multi-robot dynamic formation while making them move toward the same target. Concerning this problem, a new biological inspired neural network based approach for multi-robot formation was proposed in this paper. In the proposed approach, a leader-referenced formation model was used to calculate the virtual target location for each robot in real-time, and a biological neural network was used to realize multi-robot navigation. Finally, some simulation experiments were carried out. The experimental results show that the proposed approach has some good performances, such as the real-time obstacle avoidance, keeping formation and moving toward the same target. Furthermore, multi-robots can change the formation quickly, which proves the real-time and intelligence of the proposed approach.%多机器人编队控制是多机器人协作领域的重要研究内容之一,如何实现多机器人朝同一目标移动的同时保持队形是多机器人编队的一个热点和难点问题.针对这一问题,提出一种新的基于生物刺激神经网络的多机器人动态编队方法,采用基于leader-referenced编队模型实时计算各机器人的虚拟目标位置,利用生物刺激神经网络进行机器人导航.最后进行仿真实验,实验结果表明该方法在实现多机器人实时避障并保持队形的同时,朝同一目标移动,而且可以很快实现队形变换,具有较好的实时性和灵活性.
展开▼