Research And Application Of Robust Consensus Tracking Control For Multi-Agent Systems

Compared with the simple superposition of function and efficiency of multiple single agents,the distributed cooperative control of multi-agent systems can effectively improve work efficiency,reduce energy consumption and improve system reliability.Therefore,the distributed cooperative control problem of multi-agent systems has received intensive attention and obtained fruitful results.However,most of distributed control strategies rely heavily on accurate modeling,ideal communication environment and accurate actuator input.In practical applications,external disturbances,communication capacity limitation and actuator failures inevitably affect the performance of the controller,which reduces the effectiveness and practicability of the control strategy.In order to improve the performances of cooperative control for multi-agent systems in practical applications,this thesis considers external disturbances,communication network limitation,actuator failures and convergence speed into the cooperative control problem of multi-agent systems.The following problems are addressed in the study:(1)A robust consensus tracking sliding mode control strategy and a consensus tracking control strategy based on extended state observer are proposed for second-order leader-follower multi-agent systems with external disturbances such that leader and followers can achieve consensus within a fixed time and improve the anti-interference ability.Firstly,a fixed-time distributed observer is designed for each follower to estimate the position and velocity of the leader when some followers cannot obtain the information from the leader.Then,based on the observed information,a novel terminal sliding mode surface with fixed-time convergence property and sliding mode controller are proposed to ensure that the system errors can converge to the sliding mode surface and converge to zero along with the surface within a fixed time.In order to further improve the anti-interference ability,the fixed-time extended state observer is proposed to estimate the external disturbances and compensate the effects of disturbances into the controller,such that multi-agent systems still can accurately achieve consensus tracking control despite the effects of disturbances.Finally,two sets of simulation results demonstrate the effectiveness of the proposed robust control strategy.(2)Considering the energy consumption of communication and calculation for multi-agent systems,a distributed continuous-time consensus sliding mode tracking control protocol and a distributed event-triggered consensus sliding mode tracking control protocol are proposed.The convergence time of multi-agent systems and the update frequency of control input are compared under two control strategies.It is proved that the distributed event-triggered control strategy can effectively reduce the update frequency of control input and save energy consumption.Firstly,a terminal integral sliding mode surface with fixed-time convergence is proposed for multi-agent systems with external disturbances.Then,the distributed continuous-time sliding mode controller is designed such that multi-agent systems can achieve consensus tracking control.In order to reduce the communication and computing resources consumption,the event-triggered mechanism is applied to the sliding mode control algorithm.An event triggering condition which is fairly suitable for systems with uncertain model and perturbations is put forward such that the update frequency of control law can be considerably reduced and Zeno behavior can be removed.Finally,two proposed sliding mode control strategies are applied in numerical simulations.The relationship between convergence time and event triggering times is analyzed in detail.Moreover,the event-triggered sliding mode control algorithm is applied to the nonholonomic robot systems,which reduces the update frequency of control input and energy consumption effectively.(3)In view of the multiple actuator failures,the fault-tolerant cooperative output regulation problem of multi-agent systems is addressed.The proposed adaptive fault-tolerant output regulation control strategy can effectively compensate the impact of faults and improve the reliability of multi-agent systems.Firstly,a distributed fixed-time observer is designed to estimate the system matrix and state of the exosystem.Compared with the asymptotic observer and the finite-time observer,the proposed fixed-time observer can not only accurately estimate the information of the exosystem,but also the estimated time is globally bounded for any initial conditions.Secondly,the adaptive control gains are designed to compensate the influence of partial loss of effectiveness fault,outage fault and bias fault such that the multi-agent systems can achieve output regulation control despite the impact of faults.Finally,the proposed control algorithm and the existing method are applied to the robot manipulators,and simulation results show the superiority of the proposed adaptive fault-tolerant output regulation control algorithm.(4)Apply the consensus tracking control idea to the cooperative hunting problem of multiple UAVs.Considering obstacles and disturbances in the environment,the robust distributed cooperative hunting control strategy is designed for UAVs to improve the robustness and reduce the required-time for encircling the target.Firstly,a distributed fixed-time observer is designed to estimate the target’s information.Then,the potential field function and repulsive force are designed to guarantee that the repulsive force can make the UAV stay away from obstacles and other UAVs.In addition,a fixed-time extended state observer is designed to suppress the influence of disturbances on the multi-UAVs system.The sliding mode encirclement position controller with collision avoidance and sliding mode attitude tracking controller are presented for the position and attitude subsystems of UAVs such that UAVs can avoid collision and encircle the target successfully within a fixed time.Finally,the results of simulation comparisons show that the proposed cooperative hunting control strategy has a faster convergence speed and stronger robustness than the traditional asymptotic control method.