Application of wavelet, logic and neural network in multi-target tracking data processing

This article describes the method of eliminating original data noise base on wavelet multi-scale decomposing and recombination. It can keep the instantaneous characters of original signal and track the multi-target with logic method, which is able to reduce the untrue warning rate for route efficiently. It could also put RBF neural network into obtaining a time alignment between multi-sensor data, then calculate a same goal of route by similarity in a high precision. The simulation results have shown the advantages of the method.