Consensus based hybrid state estimation for networked control systems

[July 2016 - Dec 2017]

This project aimed at extending the idea of consensus based state estimation (such as consensus-Kalman filters) to hybrid state estimators (such as IMM, GPB-I, GPB-II, etc.). The key idea is to allow consensus at each agent of the NCS and include it in a manner that allows for time-varying topologies of the connecting network.

We consider the distributed state estimation problem of a stochastic linear hybrid system (SLHS) observed over a sensor network. The SLHS is a dynamical system with interacting continuous state dynamics described by stochastic linear difference equations and discrete state (or mode) transitions governed by a Markovian process with a constant transition matrix. Most existing hybrid estimation algorithms are based on a centralised architecture which is not suitable for distributed sensor network applications. Further, the existing distributed hybrid estimation algorithms are restrictive in sensor network topology, or approximate the consensus process among connected sensor agents. This study proposes a distributed hybrid state estimation algorithm based on the multiple model based approach augmented with the optimal consensus estimation algorithm which can locally process the state estimation and share the estimation information with the neighborhood of each sensor agent. This shared information comprises local mode-conditioned state estimates and edge-error covariances, and is used to bring about an agreement or a consensus across the network.

A schematic of the state estimate update mechanism is outlined below.

Further, the proposed estimate was shown to outperform existing hybrid estimators, with at least as much complexity at each node, at the cost of a marginally higher covariance mixing steps at each node.

Principal Investigators list

Related Publications

R. Deshmukh, O. Thapliyal, C. Kwon and I. Hwang, “Distributed State Estimation for a Stochastic Linear Hybrid System over a Sensor Network”, IET Control Theory and Applications, Vol.12(10), pp. 1456-1464, June 2018, DOI: 10.1049/iet-cta.2017.1208