Robust State Estimation For Discrete Linear Uncertain Time-delay Systems With Packet Los

In the modeling process,due to the need to balance the complexity of the actual control object and the simplicity of mathematical processing,the model error is inevitable.The combination of network and control system results in time delay and measurement data loss,which are very common in networked control systems.The existence of model error,time delay and measurement data loss will lead to the degradation of estimation performance and stability,and bring new challenges to the study of state estimation problems.Therefore,in order to obtain state estimation algorithm with better performance,it is very meaningful to study the robust state estimation problem of a class of discrete linear uncertain systems with time delay,measurement data loss and deterministic control input.In this thesis,the robust state estimation problem for discrete linear uncertain systems with time delay,measurement data random loss and deterministic control input is studied from the following aspects:1.For a class of discrete linear systems with deterministic control input,the robust state estimation problem is studied under the condition of considering the influence of state delay and model error.Firstly,the state augmentation method is used to obtain a non-delay system.Secondly,based on the relationship between the standard Kalman filter and the regularized least squares problem,the innovation process sensitivity penalty function is introduced to improve the performance index function of the regularized least squares problem,a robust state estimation algorithm with a similar iterative form and comparable computational complexity to the standard Kalman filter is proposed.Thirdly,the convergence of the state estimator and the boundedness of the estimation error covariance matrix are discussed.Finally,several sets of numerical simulation examples are used to verify the effectiveness of the algorithm.2.For a class of discrete linear systems with deterministic control input,the robust state estimation problem is studied under the condition of considering the influence of observation delay and model error.Firstly,the system model transformation is completed by using the state augmentation method.Secondly,a robust state estimation algorithm is proposed based on the method of innovation process sensitivity penalty.Thirdly,steadystate properties such as convergence and boundedness are given under certain assumptions.Finally,the estimation performance of the state estimator is analyzed and compared through numerical simulation examples.3.Based on the above research results,the robust state estimation problem for uncertain system with observation delay,measurement data random loss and deterministic control input is discussed.Firstly,the state augmentation method is used to deal with the observation time delay,and the original system is transformed into an augmented system with non-delay in form.Secondly,a random variable obeying Bernoulli distribution is introduced to describe the phenomenon of system measurement data random loss.Thirdly,based on the method of innovation process sensitivity penalty,a robust state estimation algorithm is proposed.Finally,the simulation results show that the state estimator has better estimation performance.