Research on the EEMD algorithm of penetration acceleration signal processing based on independent component analysis

Penetration acceleration signal is a typical non-stationary signal, which cannot be effectively and accurately analyzed by traditional time-frequency analysis methods. EEMD algorithm improved based on the core algorithm EMD of Hilbert-Huang transform, effectively solved the mode aliasing problem of EMD algorithm and adaptive analyze the non-stationary random signals. However, EEMD algorithm also has some flaws, namely the IMF decomposed does not have a physical meaning, while independent component analysis method can effectively restore the source signal. The paper combines ICA and EEMD algorithm together to effectively resolved acceleration signal penetration feature recognition and signal de-noising problem. At the same time, the algorithm provides a base and reference for precise control of initiation.