Research On Extraction Of Auditory Brainstem Response Based On Kalman Filter

Hearing loss is the primary cause of hoarseness in children and one of the important problems that plague the elderly.Therefore,the early diagnosis of hearing loss has important research and practical significance.Auditory brainstem response(ABR)detection is an objective and non-invasive diagnostic technique for hearing loss.The accurate extraction of waveforms has an important impact on clinical diagnosis results.The existing ABR detection method is mainly based on the superposition average technique,which has serious problems such as severe signal distortion,long detection time,and high requirements for patients and test environments.Kalman Filter(KF)is a time domain filtering method that uses the recursive principle to estimate,and the amount of stored data is small,which can recover or approximate the signal in the observation with noise.Widely used in signal noise reduction and feature extraction.Therefore,this dissertation conducts KF-based ABR detection and analysis technology research.The experimental results show that the KF method can effectively reduce the number of experimental repetitions(down to 500 times)at rest.When the noise in the ABR changes,the process noise covariance Q and the measurement noise covariance R in the KF model need to be manually adjusted.When the two are inaccurate,the accuracy of the KF model may decrease or even diverge.In this regard,this dissertation studies the ABR detection and analysis method based on Adaptive Kalman Filter(AKF).The results show that the AKF is used to extract the ABR.Compared with the superimposed average and KF,AKF can significantly enhance the amplitude of each ABR wave,which is beneficial to the identification of ABR waveforms and the analysis of each wave.A clear ABR waveform can be obtained in the case of containing a large amount of myoelectric noise such as a mouth opening,chewing,etc.,and the extraction and analysis of each wave amplitude are completed.In summary,the KF-based ABR detection method can significantly reduce the number of experimental iterations when compared with the traditional superposition averaging method in a resting state or with a small amount of noise.The AKF-based ABR detection method realizes the automatic adjustment of the noise parameters.Compared with the KF and the superposition averaging method,it can easily and effectively filter out a large amount of noise and increase the ABR wave amplitude.The work of this dissertation is beneficial to reduce the difficulty of ABR detection and provide new ideas for clinical application.