A Rapid Extraction Algorithm Of Wiener Kernel And The Application In Modeling Of OAE

In life, in order to carry out the identity authentication the technology of biometric identification has been widely applied, in which the fingerprint, face,iris and retina is identification are the most commonly used. But the fingerprint is easy to counterfeit, also some occupation population could be affected because of the fuzzy fingerprints; In face recognition, now the plastic surgery or cosmetic problems will cause great distress to the identification; in iris recognition, it depends the lens precision; in the retina recognition, laser will cause harm to the body of examiners. By contrast, because the origin of OAE(acoustic emission)are closely related to peripheral auditory system, it is difficult to imitate and has the characteristics of higher safety and reliability.This paper uses the Wiener kernel to the OAE signal modeling, but with the increasing of the order number and the sampling points number, the computation of kernel value grows exponentially. Therefore, fast algorithm of folding recursive is proposed, and in the kernel calculation of one to three order, one to five order, the ultiplication calculation of optimized ratio can reach more than90%. Then used in nonlinear analog circuits feature extraction, consistent with the core values acquired by existing Wiener kernel calculation method, the time of the fast algorithm simulation is shorter and its validity is verified.Otoacoustic emission model is a nonlinear systems the model samples 448 points in the same interval, which based on the signal of the stimulation sound and TEOAE output. And then using the proposed recursive folded Wiener extract ion kernel as fast algorithm model. By changing the stimulus frequency of the voice, compared with the TEOAE model in this paper and the output otoacoustic emission signal in the detection system, it has the better consistency, and verify the correctness of modeling.In order to obtain the otoacoustic emission signals used by TEOAEmodeling, this article designs a system of emission otoacoustic detection based on ATmega128 microcontroller, in which The hardware part includes stimulating sound unit, signal processing unit and acquisition unit and data storage unit and so on, the software part is mainly used to display the TEOAE signal waveform and realize the folded recursive algorithm.In the experimental part, the second order Wiener model is established by the signal of the fifty experimenters, then the second Wiener kernel of the fifty experimenters is measured repeatedly in random order and the correct rate can be96% by identify experimenters. It proves that the proposed approach in this paper is reliable which can identify authentication.