The Research On The Target Localization And Tracking Based On WSN

As a new information capturing and proeessing technology, wireless sensor networks can be widely applied in many applieation areas, such as battlefield surveillance,environment monitoring, space exploration, indoor user tracking and others. It can realize the functions of locating the person, tracking of the trapped person and so on in a new huge building rescue system because of its unique technology advantages. Target localization tracking is one of its applications which has universal military and civil applications. In this thesis, target localization tracking based on wireless sensor networks is researched.For the trapped person in huge building, particle swarm optimization algorithm is applied to the problem of sound source localization in wireless sensor networks. A model of sound source localization is established which applys the location algorithm based on time difference of arrival method and maximum likelihood estimation method using the different propagation speed between RF signal and sound signal, sample data through wireless sensor network node and realize sound source localization with particle swarm optimization algorithm. The simulation results illustrate that the algorithm has high estimation accuracy and low computational complexity.Classical target tracking algorithms are researched and used for wireless sensor networks target tracking. Tracking precision and robustness of different target tracking algorithm are investigated through changing the non-linear degree of wireless sensor networks system. The simulation results demonstrate the particle filter algorithm has higher tracking precision and stronger robustness in wireless sensor networks target tracking.A new particle filter algorithm is presented. Combining particle filter with extended kalman filter since the disadvantage of general particle filter algorithm, the linear error of extended kalman filter and the current measured value are considered comprehensively which make the distribution of particles closer to the true state of posterior probability distribution. The simulation results show that the algorithm has better tracking effect.