Design of the Self-organizing Sensor Networks Based on Neural Networks

Multisensor target tracking is an important part of the sensor data fusion. Although a lot of fusion algorithms have been put forward, very few research is taken on the configuration of the sensor network, which is necessary for a successful UGS network system. We designed an adaptive Hopfield network with adjustable neural thresholds, which can select suitable sensors to form a target tracker. The system maintained high enough tracking precision with sensors as few as possible. Simulation results proved the effectiveness of the algorithm.