Adaptive Fault Tolerant Control Of Multi-agent Systems With Actuator Saturation

Due to the autonomy,social competence and reaction speed of multi-agent systems,distributed control has become the new direction of study in control field.Currently,multi-agent systems has been widely applied to the fields like unmanned vehicle,grid system and unmanned aerial vehicle.Multi-agent systems is composed by several mutually independent units connected by communication topology and more complex in structure than single system and more demanding for the performance of system.Therefore,the study of multi-agent system is of great significance.Consistency is the most fundamental matter of the study in multi-agent systems,but current research conclusions pay little consideration to the consistency control of multi-agent systems with saturation fault in actuator.For that,the research contents of paper as shown as follow:Firstly,with the regard to the consistency of multi-agent systems with actuator saturation,when the input of actuator is saturated,a state feedback controller has been designed by relevant principles according to relative state information of neighbor agent,and the attraction domain of system is optimized by optimization process.Secondly,while actuator is involved in the fault of saturation,in view of the consistency of multi-agent systems,we have studied the situation of multi-agent systems without its leader.An adaptive fault-tolerant controller has been designed to evaluate an online fault and thereby solve the fault of system actuator.Adaptive law can update on real-time basis to compensate the influence on system.Optimization process expands attraction domain of system and data of F18 warcraft is introduced to simulated analysis.Finally,a consideration is given to the consistency of multi-agent systems given the saturation and fault of actuator and L2-disturbance,the fault of actuator can be solved by self-adaptive fault-tolerant controller;an adaptive H? index is given to describe the anti-disturbance capacity of system and the index is optimized by optimization process.At last,a simulated analysis is made based on the data of quadrotor.