基于群集思想的多智能体编队避障算法研究

Obstacle Avoidance and Formation Control for Multi-agent Based on Swarming

基于生物群集理论,针对多智能体系统的避障问题和编队问题,在传统势场函数的基础上,提出了静态障碍环境下的多智能体编队避障策略,其中包括基于群集理论的分布式群组运动模型、基于极限环理论的智能体避障控制算法以及基于一致性理论的多智能体编队控制算法。提出了清晰的多智能体系统的协调控制框架,通过算法方面的改进和优化,有效地克服了人工势场法在智能体避障控制中易陷入局部极值,以及障碍物周围易发生振荡等缺陷,为多智能体的编队控制问题提出了新思路,同时通过算法稳定性分析和仿真实验验证了该运动模型的可行性和有效性。

On the basis of the biological swarming theory, the research of obstacle avoidance and formation control of multi-agent system was focused on. Based on the traditional potential field function, a novel method of obstacle avoidance and formation control strategy were proposed in the presence of static obstacles, in which a distributed group mobility model was constructed, a new obstacle avoidance algorithm based on the limit cycle was proposed and a consensus-based formation control algorithm was introduced. A novel coordination control scheme of multi-agent system was proposed. By improving and optimizing the algorithm, the defects of artificial potential field method in obstacle avoidance, such as the oscillation around the barrier, falling into the local extreme value etc. were overcome and a new strategy for the formation control of multi-agent system was provided as well. Finally, the feasibility and effectiveness of the mobility model were verified by means of stability analysis and simulation experiments.