Target Detection Technology Based On Self Organizing Network

Target tracking technology in self-organizing sensor networks has great practical value in both military and civilian. Its main task is to estimate the position, orbit, velocity and acceleration of the target in the noisy environment, according to the data collected by mobile sensors or static sensors. Self-organizing sensor technology includes communication technology, sensor technology, distributed processing technology,embedded computing technology, and is widely to implement regional monitoring and tracking tasks in dangerous environment not suitable for manual operation. It's a hot topic how to carry out effective sensor management, so that the network sensing area is able to cover the entire monitoring area, and the monitoring system can choose the appropriate sensor for accurate tracking at the right time.This paper is presented according to the functional implementation sequence of ad hoc networks. First of all, scientific deployment algorithms are needed to mobilize the sensors in the network. Given some constraint conditions, let the network monitoring area contains the entire area, and drive or allocate multiple sensors to track targets properly.After completing the sensor deployment task, suitable algorithms should be designed to estimate the velocity and acceleration information of the target according to the data transmited by mobile sensors or static sensors with spatial position dispersion. In some difficult target tracking problems, such as the track where the abrupt change of the trajectory happens, suitable tracking algorithms are desired to be designed to keep the system robust. The main work of this paper is as follows:1.Three distributed self propagating deployment algorithms for mobile sensor are presented. Each node determines whether to perform a peer-to-peer schema or a local cluster model according to the residual energy and local node density in this algorithm.This algorithm can make the sensor network to achieve the maximum coverage of the monitoring area and prolong the service life of the system.2. Improvements are made on the pseudo linear estimation. A Auxiliary variable weighting matrix is constructed based on the pseudo linear error compensation estimation algorithm to reduce the error covariance matrix of the original algorithm and ensure the convergence of the algorithm at the same time.3.A distributed variational algorithm is presented. This algorithm compress the data exchanged between nodes with variational formulation to reduce the amount of communication data and communication energy consumption. Besides, this kind of variational algorithm has good characteristics of independent parameters to ensure the robustness of the data processing, which makes it easier to track the target with a track mutation.