Localization And Tracking Technology Based On Particle Filter For Wireless Sensor And Actuator Networks

Target location and tracking technology have played an important role in wireless sensor network. Wireless sensor and actuator network is an extension and expansion of wireless sensor networks, how to take full advantage of this new network to timely, effectively locate and track the moving targets is a key technical issue to be solved. Traditional target tracking technology in wireless sensor network is observing and positioning the nodes in every moment in order to achieve continuous tracking of the target. For the lack of predicting to model, the tracking performance on accuracy and real-time is unsatisfactory. For the tracking of maneuvering target, mostly involving non-linear, non-Gaussian scenes, needs to use the high performance of particle filter. Therefore, this thesis focuses on the Monte Carlo localization algorithm and interacting multiple model particle filter tracking technology, the author has improved the existing algorithms and applied them into the wireless sensor and actuator network to locate and track the moving targets.First of all, the research status of positioning and tracking technology based on particle filtering techniques in WSN and WSAN are analyzed and summarized in detail, the particle filter algorithm and its improved algorithm are introduced, which lay a theoretical foundation for the study and improvement of particle filter.Secondly, the basic principles of Monte Carlo localization algorithm based on wireless sensing and actuator network are introduced, and the specific implementation steps of MCL algorithm are elaborated. Considering the requirements of positioning accuracy, an improved Monte Carlo algorithm was proposed. Simulation and analysis verify the effectiveness of the improved algorithm.Thirdly, the interacting multiple model algorithm based on Kalman filter and interacting multiple model algorithm based on particle filter are analyzed. As two Inadequacies existed in IMMPF algorithm, the author has researched a new IMMPF algorithm based on variable rate model and genetic algorithm, and verified the effectiveness of improved algorithm based on simulation.Then, the target tracking technology are discussed and tracking model based on WSAN are established, the improved algorithm proposed in Chapter IV is applied into the target tracking in WSAN and a distributed target tracking algorithm is researched, the effectiveness of this algorithm is verified by simulation and performance analysis.Finally, the work carried out in this thesis is summarized, and the content and direction for further research are discussed.