Investigation on distributed Kalman filtering using diffusion strategies

Tutor: Shengdi Wang
Type of Thesis: Project (MSc)
date of end: 06/2018
Student: Lingrui Zhu
Status: finished

The main objective of this work is distributed state estimation in wireless sensor networks (WSNs). We assume that some sensors forming a connected network topology are deployed in a space to measure state variables in a common modeled process (e.g. a rotating object or a moving car). The most famous state estimation approach is called Kalman filter (KF). To avoid the centralized signal processing, there is no central node in our scenario. Each node will estimate the state variables distributedly based on its local knowledge and information from neighbors. Hence, a distributed Kalman filter (DKF) algorithm should be further investigated. 

In this project, the student will focus on a distributed Kalman filtering method using diffusion strategies [1]. According to the literature, an analysis on the performance of the algorithm should be studied. MATLAB simulations regarding to diffusion DKF compared with other DKF algorithms [2] also needs to be done.

[1]. Federico S. Carrivelli, '' Diffusion Strategies for Distributed Kalman Filtering and Smoothing''.

[2]. R. Olfati-Saber, ''Distributed Kalman Filtering for Sensor Networks''.

Last change on 29.06.2018 by S. Wang
AIT ieee GOC tzi ith Fachbereich 1
© Department of Communications Engineering - University of BremenImprint / Contact