Design Of Sparse Structure Controllers For Distributed Networked-based Systems

Networked control systems(NCSs)are spatially distributed systems,in which the communication among sensors,actuators and controllers is carried out through a shared digital communication network.Amazing advantages of the network architecture include reduced system wiring,device plug and play,increased system flexibility,and ease of system diagnosis and maintenance.In the past two decades,modeling,analysis and control of NCSs have received extensive attention.Sparse optimization has broad research prospects,which can effectively save system resources,save construction costs,and effectively resist subsystem failures,network attacks and other network induced constraints.For NCSs,this paper studies the sparse optimization problem of structural controller,that is to say,the system can still maintain the given control performance with less underlying control elements(sensors,controllers,etc.)or communication links.The main contributions of this paper are as follows:(1)For distributed networked control systems with sensor failures,a column sparse structure controller optimization algorithm is designed.It is assumed that a certain number of subsystems have sensor failures,which is modeled as a mathematical model,which means that the same number of columns in the controller gain matrix are zero.This kind of cardinality constrained problem is usually NP hard because of its combinatorial nature.In this study,we also use a continuous event-triggered mechanism to save the transmission resources of the NCSs.Combined with Lyapunov function,the conditions of asymptotic stability and controller design are given.(2)A block sparse structure controller optimization algorithm is designed for distributed networked control systems with limited coupling communication links.Through this algorithm,the distributed control system can maintain the stable operation of the system with fewer controller communication links.This research is different from the existing norm based iterative methods,but proposes a design method based on mixed integer programming to deal with cardinality constraints.In order to save communication bandwidth,a periodic event-triggered mechanism is applied,and the effectiveness of the proposed method is proved by an example of a three machine power generation system.(3)For NCSs with limited coupling communication links,a sparse structure controller optimization algorithm based on alternating direction method of multipliers is designed.This method alternates between optimizing the sparsity of the controller and optimizing the closed-loop norm,and finally achieves the optimal trade-off between the sparsity level of the system and the control performance.Its special feature is that the minimization problem is decomposed into analytic subproblems by using the separability of the sparse penalty function.Then we use soft thresholding algorithm and Anderson-moore method to solve the subproblem.