Cooperative Control of Multi-Agent Systems Using Fuzzy Approximation of Neighbors Dynamics

Cooperative control of multi-agent systems (MAS) is complicated due to uncertainties such as heterogeneous/unknown agents' dynamics, limited access to other agents' and the leader's information, and external disturbances. A cooperative control structure is proposed here for affine nonlinear heterogeneous MAS with partially unknown dynamics to deal with such challenges. The artificial potential functions (APFs) model the agents' interactions. A fuzzy system is employed to approximate the agent's unknown dynamics, while a second fuzzy system is devised to estimate a part of the control input related to its interaction with neighbors, thus eliminating the need to access the nth-order derivative of the neighbor's state. The overall stability and the update rules for the weights of the fuzzy systems are derived using the Lyapunov stability theory. Simulations are carried out on cooperative control of third-order MAS, and the results demonstrate the proposed method's efficacy in tracking the leader's state compared to a competing control strategy.