Design Of Fuzzy Adaptive Controller For Heterogeneous Multi-agent System

In the gradual application of control theory,it is found that most of the uncertain systems in the industry are heterogeneous.However,the existing theoretical research focuses on homogeneous multi-agent systems(MASs),and the related control theories of heterogeneous MASs need to be continuously supplemented and improved.Through the analysis and summary of commonly used control theories,and by understanding its application scenarios,the fuzzy control is selected for the system model studied in this paper.Fuzzy logic system has a very powerful approximation ability,which is reflected in the ability to approximate any nonlinear function,so it is extremely effective for building nonlinear system models and has been favored by many scholars.This paper will use the superior approximation characteristics of the fuzzy logic model to proposed the type I and type II fuzzy adaptive control methods respectively for realizing the uniform control of nonlinear heterogeneous MASs.For a class of MASs with matched nonlinear functions,the leader-following consensus problem is mainly considered.The FLSs are adopted to approximate the unknown nonlinear functions.Unlike the existing results,the dimensions of each agent dynamic system are different in this paper.By defining similar property and introducing the similar parameters,a distributed fuzzy adaptive controller with similar parameters is designed to guarantee that the state of each follower can stably track the state of the leader system.It is assumed that in the network connection graph,there is at least one node that is associated with other nodes and can exchange information.Then based on the graph theory and Lyapunov function approach,it is proved that all signals in the closed-loop system are uniformly ultimately bounded and the consensus tacking error converges a compact set that can be designed as small as possible.Finally,the feasibility of the proposed controller is demonstrated by analyzing the simulation results of MASs with different dimensions.Aiming at a class of nonlinear heterogeneous MASs with external disturbances,it mainly solves the problem that the state trajectory of the follower can track the state trajectory of the leader stably with a small error.First,the similar structure of a class of heterogeneous MASs is introduced,and some parameters between each agent are defined.Secondly,a robust distributed type-2 fuzzy adaptive control with similar parameters is constructed to achieve the consistency of heterogeneous MASs with external interference.According to the knowledge of graph theory and the Lyapunov function method,it can be ensured that all signals in the closed-loop system converge to a small zero domain.Finally,by designing some simulation examples with the same and different dimensions,the advantages of this control scheme are highlighted by comparing with the results of the existing literature.For a class of multi-agent robot manipulator systems with unstructured models,a novel distributed feedback adaptive control is designed to enable it to track the state of the leader system stably.It needs to be emphasized that the dimensions of each agent robot manipulator system can be allowed to be different,some manipulators have 2 degrees of freedom,and some manipulators have 3 degrees of freedom,which is the main advantage compared with other distributed control schemes.Aiming at the problem of different dimensions,the definition of similar nodes is proposed,and similar parameters are calculated according to the proposed similar conditions and some information of the system itself.Based on these similar parameters,a distributed type 2 fuzzy control scheme with coupled weight adaptive law and fuzzy weight adaptive law is proposed to ensure that the state of each agent manipulator system can track the dynamic behavior of the leader system with a good performance,and the tracking error converges to a small range.This method extends the work of the existing multi-agent follower system and leader system with the same dimensions to MASs with different dimensions.Finally,the good stabilization performances of the proposed control strategy are demonstrated in the simulation example by employing the relationship of directed connected communication graph for every multi-agent robot manipulator system and the leader system.