Stochastic Hybrid Systems Estimation and Control Techniques

Abstract

quotThe Stochastic Hybrid System (SHS) allow the interaction among continuous newlinedynamics and discrete dynamics. Many applications in the real world problem can be newlinemodelled by way of SHS since it provides great versatility. In this research work, we study newlinesome state estimation techniques for linear and non-linear SHS with missing newlinemeasurements. The state estimation of SHS has different applications in communication newlinesystems, flying vehicle dynamics and control, stock costs, target tracking and so forth. The newlinemajority of the work on state estimation of SHS clustered around the deterministic models. newlineIt depicts the quality of the framework without taking into consideration any vulnerability. newlineIn reality, a few degrees of vulnerabilities are required to be considered in the system newlinemodel. Due to vulnerability like measurement loss and delay, state estimation of the newlinesystem will be influenced. newlineTo deal with such sort of circumstance, the researchers have broadened their newlineinvestigation to incorporate probabilistically or guard condition-based state transition for newlinevarious system models. For instance, in the flying object case, the dynamics of the flying newlineobject can be represented as an SHS, as discrete changes between flight modes and the newlinecontinuous physical dynamics for nonstop movement relating to a particular flight path. newlineFor effective state estimation of SHS, it is important to determine the accurate discrete newlinestate transition and continuous state. Several state estimation algorithms have been newlinedeveloped based on Kalman Filter and Particle filter for SHS without sufficient discussion newlineof missing measurements. Firstly, to study the state estimation of SHS, estimation methods newlinefor linear and non-linear system models is discussed. Then, the Data Loss Detection newlineKalman Filter is proposed to predict both the continuous and discrete states of the linear newlineSHS with missing measurements based on guard condition. Also, the performance of the newlineproposed algorithm is improved based on Chi-square statistics-based measurement loss newlinedetection for low pr