Experiences in Data Processing and Bayesian Filtering Applied to Localization and Tracking in Wireless Sensor Networks

Abstract

The authors discuss algorithms and solutions for signal processing and filtering for localization and tracking applications in Wireless Sensor Networks. Their focus is on the experiences gained from implementation and deployment of several such systems. In particular, they comment on the data processing solutions found appropriate for commonly used sensor types, and discuss at some length the use of Bayesian filtering for solving the tracking problem. They specifically recommend the use of particle filters as a flexible solution appropriate for tracking in non-linear systems with non-Gaussian measurement errors. They also discuss in detail the design of some of the indoor and outdoor tracking systems they have implemented, highlighting major design decisions and experiences gained from test deployments.