Distributed Formation Control Of Networked Euler-Lagrange Systems

In recent years,the coordinated control of networked Euler-Lagrange(EL for short)systems has become one of the research frontiers for multi-agent systems in the field of nonlinear dynamics and control.This is mainly because EL systems can be used to describe the structure and motion characteristics of a large number of practical engineering systems,such as robot manipulator,mobile robot and autonomous vehicle,etc.Different from the single EL systems,mobility,flexibility,maneuverability and enforceability are the main features of the networked EL systems,such as rendezvous and docking of spacecrafts,attitude adjustment of satellites and coordination of multiple manipulators.Therefore,the collaborative control of networked EL systems has very important research value in both theoretical study and potential applications.Built on the current research results,this dissertation gives an in-depth study on the distributed formation control problem of networked EL systems.The main research tasks of the thesis are summarized as follows:1.Distributed formation cooperative control of networked EL systems based on virtual structure.A kind of distributed velocity observer is proposed based on the virtual structure under the directed network communication topology,which makes all the fol-lowers obtain the estimated velocity information of the dynamic leader indirectly.Then a fully distributed formation coordination controller is designed such that the formation error eventually converges to zero.Furthermore,by considering the communication lag in the process of information exchange,an improved adaptive formation control algorithm is constructed by introducing the time delay compensation.According to Lyapunov stability theory,the corresponding criteria for formation coordination control are developed to en-sure that the desired formation of networked EL systems can be achieved.The simulation results show that the proposed controller is robust to both bounded communication time delays and uncertain EL systems.2.Distributed formation cooperative control with collision avoidance of networked EL systems.In order to realize the multi-objective control task,including formation tracking,formation keeping and collision avoidance,two kinds of distributed formation protocol are proposed for both cases of with and without a dynamic leader based on the virtual structure and artificial potential function method.The virtual structure is used to design the desired rigid formation,and the potential energy function is employed to avoid collision.In the cases of leaderless,a new adaptive gain technology is proposed.Specifically,the developed adaptive gain is to adapt itself duly based on both the network communication topology and collision avoidance constraints,thereby realising formation tracking,formation keeping and collision avoidance.In the case of a dynamic leader,an improved distributed velocity observer for followers is designed to reach velocity matching with the leader,and some simple yet general algebraic criteria are derived to guarantee that the desired formation with collision avoidance for the networked EL systems can be achieved.Simulation results are provided for four two-link rigid robot manipulators to demonstrate the correctness of the theoretical results.3.Region-based formation cooperative control of networked EL systems.It is not conducive to implement large-scale formation control,since the geometric constraints between followers will become more complex with the increase of the number of followers in virtual structure.Based on this factor,the problem of region reaching formation control is investigated in this thesis.Firstly,by selecting the appropriate region potential function and using the classical PID control strategy,the single follower can enter the specified moving region instead of the traditional fixed-point tracking.In order to make multiple followers enter the target region and avoid collision between each other,an improved adaptive gain technique is proposed.In this case,the geometric shape of a swarm of the followers depends on the region potential function,and there are no definite geometric constraints between the followers in the objective region.Therefore,the region-based adap-tive formation controller is suitable for large-scale formation control.Finally,numerical simulations illustrate the feasibility of the proposed control scheme.