Penetration signal adaptive cognitive filtering model based on wavelet analysis

Hard target penetration is a very complex dynamic problem, and the measured signals contain axial de-acceleration, vibration, and other weakly interference signals, so it belongs to the non-stationary random vibration process. The filtering method based on threshold is one of the effective methods on some extend to eliminate useless signals. However, the high-g acceleration signal components which related to vibration and interference in penetration processing usually cannot be completely filtering by using the Hard Threshold and Soft Threshold method. As such, we propose a new method to establish a unified adaptive cognitive wavelet decomposition filtering model base on the situation of the wavelet multi-scale transform. Further we establish similar signals with penetration processing and compare the filtering effects among LMS algorithms and Hard/Soft Threshold in mat lab simulation software. Finally we get the best choice for the type of signal filtering, which provide the best treatment for the measured penetration signals.