A study on echo feature extraction based on the modified relative spectra (RASTA) and perception linear prediction (PLP) auditory model

摘要

The more mature RASTA-PLP auditory mode in the field of speech recognition is presented to apply to the field of underwater target echo recognition. But also, According to the character of underwater target signal, The RASTA-PLP auditory mode is modified. Contrast to the PLP auditory model feature, the recognition ratio of the modified RASTA-PLP auditory model feature is higher. The recognition ratio of the test samples can arrive 90.32% by Fuzzy Adaptation Resonance Theory(FART) neural networks when the ratio of the sample numbers of training set to the testing set''s is 1∶10. After that, the Gauss white noise is convoluted with the signal. At the equal test condition, the recognition ratio of the modified RASTA-PLP auditory model feature is 18% higher than the PLP auditory model feature. It shows the modified RASTA-PLP auditory model feature is robust.