Cooperative avoidance control for multiagent systems

The objective of this paper is to present a methodology for designing cooperative control laws for individual agents that guarantee collision avoidance in multiagent systems. The proposed avoidance control laws are easy to design and implement, and may be directly appended to the optimal control laws of the individual agents within the cooperation framework. The avoidance control laws are computed using value functions that resemble the behavior of barrier functions in the static optimization theory. The most attractive feature of the proposed optimization scheme is the fact that the avoidance laws are active only in the bounded sensing regions of each individual agent, and they do not interfere with the agents' individual optimal control laws outside of these regions.