The pattern recognition of radar signal by self-adapted wavelet

Radar signal is unbalanced and Usually mixed by noises. Wavelet transform is a Power tool for analyzing unbalanced signal and multiresolution is the normal way. Also the choice of the wavelet bases is important. A new method for band limited wavelet construction is developed in the radar signal processing, which automatically locks on the frequency band that energy is concentrated. By the algorithm, an optimal scale function is calculated according to the power spectrum of the learning radar signal, the corresponding wavelet can be calculated by two scale relation. With the scale and wavelet filter, the processing result will be the recognized pattern which is similar with, the learning radar, signal. The experimental results demonstrate the method is effective. Radar signal can be processed in real time.