Research On Localization And Target Tracking In WSNs

The accurate localization of sensor nodes is the premise of monitoring events normally. It has important theory and practical significance to sensors' self-localization and target tracking. The localization algorithm with a mobile beacon reduces the cost of the whole network and the influence of barriers. But the mobile beacon's movement and sending positions influence the position coverage rate, the position time, the position accuracy directly. If the target tracking algorithm has a good prediction mechanism, it can improve the tracking performance greatly.The localization mechanism with a mobile beacon and target's prediction mechanism of Wireless Sensor Networks were researched in depth in this thesis. It mainly embodied in following aspects:(1) Paying a great effort on the algorithms of localization with a mobile beacon, and mainly focusing on the analysis of the mobile beacon's optimal sending positions and movement. Virtual force (BOALR) algorithm impacted by static sensors, can't attain global positions optimization, and has low position coverage rate. According to the optimal path principles of mobile beacon, this thesis proposes a self-organizing technique for enhancing the position coverage rate, which is called mobile beacon trajectory algorithm based on particle swarm optimization (PMBT). PMBT can be successfully used in WSNs and effectively achieve global searching for optimal positions of mobile beacon. Then presents an algorithm called mobile beacon trajectory algorithm based on virtual force-guided particle swarm optimization (VPMBT). It is to use virtual force to guide the evolution of all particles and accelerate the updating of PSO for improving the convergence speed to get the optimal movement of mobile beacon. And then locate the sensors with trilateration.(2) In order to reduce the tracking error and accurately predict the target trajectory in the target monitoring and tracking process, an improved distributed algorithm for single-target tracking based on Kalman filter model was provided, and the value on the target's position as a Kalman filter model of observations on the estimated value of target motion to amend. And make it strategy for the nodes of the network's work and sleep mode. And also proposes a target tracking failure recovery mechanism.(3) At last the performance of all the solutions proposed in this thesis are verified by simulation. The simulations prove that, the two beacon optimal movement patterns of this thesis can attain the global positions optimization with higher position coverage rate of 7.3%,13.5% respectively than BOLAR. And the single object tracking algorithm based on Kalman Filter can greatly reduce the tracking error, predict the goal's trajectory more accurately under the premise of improving the robustness and reducing the energy of failure recovery. And this algorithm improves tracking accuracy by 8.2% than AMKF algorithm(Adaptive Algorithm based Multi-Kalman Filter).