Stability Analysis And Controller Design Of Networked Control Systems With Aperiodic Sampling

Networked control systems(NCSs)in which feedback paths are implemented by a realtime network,were proposed in the 1980 s and have attracted wide attention over the past decade.NCSs break the space constraints in the traditional control systems and have advantages of sharing information resources,remote operation,high efficiency and flexibility.Accordingly,NCSs are widely used in various fields,such as complex industrial processes,spacecrafts,electrified vehicles,smart home systems and so on.Meanwhile,network also brings challenges to NCSs,such as time-varying transmission delays,packet dropouts and quantization error,which would lead to performance degradation,bring about destabilization and pose difficulties for researchers as well.Aperiodic sampling develops rapidly since it offers clear advantages in dealing with constraints of energy,computation and communication resources,while hardware device limitations and the environmental factors make it inevitable.Consequently,it is currently considered to be one of the key research fields of control engineering to establish models and to propose stability conditions of NCSs with aperiodic sampling.In this dissertation,NCSs with aperiodic sampling are considered and the time-varying delays can be larger than one sampling interval,which is generally applicable in real practice.With different networked control scenarios,model transformation approaches for aperiodic sampling intervals and time-varying delays,interconnected systems of the closed-loop system,new stability conditions and controller design methods are proposed,based on the scaled small gain(SSG)theorem and the integral quadratic constraint(IQC)framework.The primary work of this dissertation includes the following aspects:(1)Observer-based output feedback control and dynamic output feedback control are carried out for NCSs with both aperiodic sampling and time-varying transmission delays,under the traditional actuator strategy which means that the control input remains constant in one sampling interval.The closed-loop system is discretized as a discrete-time system with time-varying system matrices and time-varying delays.By model transformation of aperiodic sampling,the closed-loop system is converted to a discrete-time system with time-varying delays and norm-bounded uncertainties.It is then modeled as an interconnected system by virtue of model transformation of time-varying delays.Utilizing the SSG theorem,asymptotic stability conditions with lower computational complexity are derived for the closed-loop system.Design methods are also given to obtain the gain matrices of observer-based output feedback controller and dynamic output feedback controller,using a modified cone complementary linearization algorithm.(2)Under the actuator strategy in which more than one control inputs could be applied to the plant in one sampling interval,state feedback and output feedback control are addressed for NCSs with aperiodic sampling.By applying model transformation approaches of aperiodic sampling and time-varying delays as well as the lifting method,the time-varying discrete-time system is transformed into an interconnected system.The forward channel is delay-free and contains the norm-bounded uncertainties caused by aperiodic sampling,while uncertainties from time-varying delays are included in the feedback channel.Asymptotic stability criteria are proposed in terms of a single linear matrix inequality(LMI),which have lower computational complexity and conservatism.An iterative algorithm is employed to obtain the output feedback controller.(3)In the framework of integral quadratic constraints,the stability problem is investigated for NCSs with aperiodic sampling.The closed-loop system is modeled as a discretetime system with two kinds of time-varying terms,that is,multiple time-varying delays and norm-bounded uncertainties coming from the nonuniform sampling intervals.The IQC characterizations and corresponding IQC-induced systems are proposed for both aperiodic sampling uncertainty operator and the multiple-delay operator.The closedloop system is then modeled as an interconnected system in which the forward channel is linear time-invariant,while the feedback channel contains both aperiodic sampling uncertainty operator and multiple-delay operator.Under the IQC framework,a stability condition is given based on the IQC stability theorem.