Denoising of weak radar signals using wavelet packet transform and genetic algorithm

Abstract

Radar constitutes one of the major application areas of communication systems. Until radar signal reaches from source to target, the noise may be mixed with the original signal because of environmental or human-induced reasons. Thus, it may be impossible to read the signal correctly. In this article, the purpose is to denoise the weak radar signals from noise with high-performance. For this purpose, Wavelet Packet Transform is used as different from the traditional methods. While these transforms are being performed, the selections of the appropriate wavelet family type, entropy type and level are quite important. Moreover, the suitable threshold function selection is important for high performance. In this article, the Genetic Algorithm structure as intelligent system for the optimization is proposed. With this structure, the best wavelet family type, entropy type and level number can be determined. Fuzzy s-function as thresholding function is preferred and variable parameters of this function are selected according to the best performance criteria. Then, this method, which is proposed for denoising the weak radar signals from noise, is compared with other algorithms available in the literature. The achievements are tested by using root mean square error and the correlation coefficient criteria. According to the obtained results, the achievement of proposed method is better than achievements of other methods.