Research On Signal Processing Technology Of Cochlea Implant Based On Analog Wavelet Transform

Cochlear implant is an electronic device implanted through surgery and is currently the only medical device that can effectively help patients with severe deafness to regain hearing.Its working principle is to use the electrode implanted in the inner ear to bypass the damaged part of the inner ear and directly stimulate the residual auditory nerve with electric current,which can provide sound input for the deaf patient.In recent years,related scholars have been devoted to the research of the wavelet transform-based cochlear implant CIS(Continuous Interleaved Sampling)speech signal processing scheme,which overcomes the complex problem of adjusting the parameters of the band-pass filter bank.However,the digital wavelet transform requires A/D conversion,which increases the power consumption and volume of the system,which is not conducive to the development of fully implantable cochlear implant.In order to meet the needs of low-power,miniaturized cochlear implant chips,this paper proposes a technology of analog wavelet transform for cochlear implant CIS voice signal processing scheme.The main work and innovations of this article are as follows:(1)according to the frequency analysis characteristics of biological cochlea and the theoretical basis of simulated wavelet transform,an approximation method of simulated wavelet with analysis characteristics similar to actual cochlea is proposed.First,the minimum mean square error criterion is used to judge the approximation effect of the wavelet function,and the construction of the simulated wavelet mathematical model is realized.Then,for the constructed simulated wavelet optimization mathematical model,a wavelet approximation method based on hybrid simulated annealing algorithm is proposed.Simulation experiment results show that the simulation wavelet approximation method based on the hybrid simulated annealing algorithm is more superior in approximation accuracy and computational complexity.(2)the voice signal processing in the CIS voice signal processing scheme is studied,and a Gaussian analog wavelet filter bank is designed.A mixed simulated annealing algorithm is used to solve the optimized mathematical model of the constructed Gaussian wavelet basis function,and the simulated Gaussian waveletfilter bank is obtained according to the properties of the Laplace transform and the corresponding scale factors in each channel of the CIS speech signal processing scheme.The experimental results show that the designed analog Gaussian wavelet filter bank not only satisfies the stability of the system,but also has the excellent properties of the Gaussian wavelet filter,and can replace the digital wavelet transform to achieve the sub-band processing of speech signals.(3)in view of the shortcomings of the Hilbert transform method in the CIS speech signal processing scheme to extract the envelope,a speech signal envelope extraction method that simulates the complex Morlet wavelet transform is proposed.According to the principle of simulated wavelet transform,a mathematical model of the simulated complex Morlet wavelet basis function for CIS scheme speech signal envelope extraction is constructed,and the approximation function of the simulated complex Morlet wavelet is solved.Then,according to the properties of Laplace transform,a multiscale analog complex Morlet wavelet filter bank is obtained.Experimental results show that this method is feasible for Hilbert transform to realize signal envelope extraction in CIS scheme.