ZigBee-based Wireless Localization And Tracking Using Particle Filter

Wireless localization and tracking has wide application in navigation, automaticmonitoring, public security and many others. At present, most wireless localization andtracking systems can hardly meet the low-cost and high-accuracy demands in indoorenvironment. In this thesis, we discussed the application and improved methods of ParticleFilter in ZigBee-based wireless localization and tracking system, providing a feasiblesolution for locating and tracking indoor mobile target.In this thesis, we firstly, based on ZigBee wireless network, tested the characteristicsof Received Signal Strength distribution and wireless signal propagation. The test showedthat RSS (Received Signal Strength) fluctuated in a fixed location and presented anon-Gaussian distribution. In addition, observation model could not represent therelationship between RSS and distance due to the reflection, refraction and diffraction ofsignal propagation, which would lead large RSS observation errors. To solve thoseproblems, we proposed using Particle Filter estimation to solve the dynamic localizationand tracking problem.Based on the introduction of Particle Filter estimation, we simulated and analyzed theperformance of SIR (Sampling Importance Resampling) Particle Filter used in localizationand tracking. In the simulation, we used the random acceleration motion model and theRSS path loss model, analyzed how different number of reference nodes, number ofparticles and observation noise affected the performance of Particle Filter estimation, andcompared it with the Maximum Likelihood estimation and Extended Kalman Filterestimation. In the following chapter, we mainly discussed how to improve theperformance of Particle Filter in indoor environment. We proposed to utilize theinformation of environmental structures, such as deleting useless particles and exchangingthe state transition model, to improve the accuracy of the localization and tracking.The research in this thesis provided an implementation method for the application ofParticle Filter in ZigBee-based wireless localization and tracking. Furthermore, theimproved methods for the Particle, based on indoor environmental structures, had realisticmeanings to enhance the performance of indoor wireless localization and tracking.