Control Of Networked Control Systems Based On Event-Triggering Mechanisms

The development of computer science and technology and network technology meets the requirements of industrial control systems for remote control,real-time control,easy installation and maintenance,and promotes the emergence of networked control systems（NCSs）and its development and application in a variety of fields,such as smart power grids,industrial automation,remote diagnosis,space exploration,and so on.At the early stage,NCSs use time-triggered control scheme,in which way,a great amount of redundant sampled data will be released into the communication network,leading to network congestion.To solve this problem,NCSs based on event-triggering scheme have been studied.Although the introduction of network improves the efficiency of the control system and reduces the installation cost of the control system,it also brings some network constraints,uncertainties and security problems to the control system,which deteriorates the control performance and threatens the security.At present,researches on NCSs based on event-triggering scheme still have some shortcomings.For example,most of the researches are based on linear systems,and researches on nonlinear NCSs still need to be further improved.For the researches of NCSs with packet loss,disturbance and input constraints,there is room for innovation in control methods and research ideas.In addition,most of the exsiting researches on NCSs only consider a single type of disturbance,and the researches on NCSs with multiple disturbances need to be improved.Last but not the least,the problem of NCS with security from the control system’s perspective is still in infancy and how to design the resilient control strategy to defend the malicious attacks is worth further study.This thesis is focused on introducing some new ideas and new methods for NCSs with nonlinearities,random packet loss,input constraint,disturbance,fault or denial-of-service attack based on event-triggering scheme.The main ideas,novelties and contributions of this thesis are presented as follows,1.For the linear NCSs with disturbance and actuator fault,the problem of fault estimation and tolerant control based on the event-triggered model predictive control approach.Different from other studies,the NCS is modeled in delta domain in this study.The unknown input observer is designed to simultaneously estimate the actuator fault and system state and is robust to the disturbance.Then the event-triggered fault-tolerant model predictive control is studied based on the estimation of the state and fault,where the actuator fault can be compensated.The data transmission from the sensor to the control is via the communication network and an event-triggering scheme is constructed after the sensor,such that the network resource can be reduced.In addition,the robust₂and_∞performances are considered in the stability analysis and controller design.Simulation results are presented to show the effectiveness of the unknown input observer and controller.2.For the nonlinear NCSs with packet loss and input constraints,the stability analysis and event-triggered model predictive control are studied.The system is described by the interval type-2 Takagi-Sugeno fuzzy model,where the uncertainties of the system parameter are captured by the upper and lower membership functions.The random packet loss is modeled by the Bernoulli distribution process.To save the communication resource,the event-triggering scheme is designed to guarantee that only the sampled data that meet the threshold requirements could be released to the network.As for the input constraint,it can be solved explicitly in the model predictive control optimization problem.The control method is designed to be robust to the disturbance and the₂performance index is regarded as the goal of the optimization problem,while the_∞one is the constraint.The resulting closed-loop system is proven to be stochastically stable and the effectiveness of the proposed methods are verified by the simulation results.3.For the nonlinear NCSs with multiple disturbance,the problem of disturbances attenuation and rejection based on event-triggering scheme is studied.The system is described by interval type-2 fuzzy model and then is approximated to be a set of linear subsystems.Based on the fuzzy systemm model,the fuzzy disturbance observer and fuzzy controller are designed.Different from above studies,there are multiple disturbances in this system,including norm bounded disturbance one and harmonic one.The robust_∞control method is used to make the system be robust to this kind of disturbance.As for the harmonic disturbance,the disturbance observer is designed to estimate it and the estimation can be used as the disturbance compensation,such that this kind of disturbance can be rejected.The data from the sensor is transmitted to the controller via the network and an event-triggering scheme is designed to reduce the frequency of control data update and data transmission.The system under the composite_∞control and disturbance observer based control is stable,robust to the norm bounded disturbance,and can reject the harmonic disturbance.Simulation results are given to show the effectiveness of the disturbance observer and even-triggered controller.4.The event-triggered control problem of NCSs under denial of service attack is studied.The system is affected by multiple disturbances and its state is unmeasured.For the unmeasured state,a state observer is designed,where the data transmission between the sensor and the observer is realized through a communication network,which is affected by denial of service attack.The attack is described by its frequency and duration and the control gains are set in switching modes based on whether the attack is active or not.A disturbance observer is designed to estimate the harmonic noise and the estimation is used in the controller design as the noise compensation.The limited bandwidth of network transmission is considered in the design of controller based on both the state observer and disturbance observer and an event-triggering scheme is constructed to reduce the released data into the network.Then the event-triggered resilient switching controller is designed,which makes the resulting closed-loop system exponentially stable,be robust to the norm bounded noise and eliminate the harmonic noise.Simulation is carried out on the assumption of attack frequency and attack duration.Results under different event trigger parameters are given to show the effectiveness of the state observer,disturbance observer and controller.