The Research Of Target Tracking In Three-Dimensional Wireless Sensor Networks

Wireless sensor is regarded as one of ten key technologies of the 21st century due to its favorable sensing ability, fast information processing capacity, low-cost and so on. These characteristics provide extensive and great application value for wireless sensor networks, such as military defense, environmental monitoring, medical care, smart home and so on. Target tracking is considered as one of these important applications of wireless sensor networks, and has been aroused the interest of research scholars at home and abroad.Target tracking mainly consists of the target detection, positioning, prediction and information transfer. It is necessary for achieving highly accurate target tracking effectiveness to design a high-precision target tracking protocol. At the same time, in order to prolong the networks life, an energy-efficient tracking protocol is also essential due to the sensor nodes are mainly battery-powered by limited energy and not easily replaced. Therefore, how to achieve high precision and energy-efficient target tracking in the wireless sensor networks has a deep theoretical study and practical significance. At present, due to the fewer targets tracking research based on the three-dimension wireless sensor networks, in order to strengthen the three-dimension space study and exploration, all the target tracking protocols in the paper are proposed in the 3D wireless sensor networks. In this paper, the works we mainly do are as follows:(1) The issue of the target location. To deal with the major positioning challenges caused by the complexity and variability of the physical environment, a target localization algorithm with grid-based probability is proposed in this paper. In the algorithm, the signal decay model with Gaussian noise is adapted to simulate and calculate the spatial distribution probability of the target position. The proposed method can effectively avoid the distortion of locating caused by the dynamic change of environment. Furthermore, this paper has proposed an adaptive learning method to correct the relevant parameter of the signal decay model, which can further improve the accuracy of positioning.(2) The issue of the target tracking based on energy-efficiency. Due to the limited energy of nodes and the node-failure will shorten the entire life of the network, so the adaptive sensor activation for target tracking strategy is proposed in the paper. Under the different sensor nodes density, the proposed tracking policy is always to meet the required tracking performance with minimum energy consumption.In addition, a method is proposed in the paper to remove outliers through the normal distribution function model, doing can effectively eliminate unreliable measurements to further improve tracking accuracy.