Research On Sound Source Localization Method Based On Time Delay Estimation Technology

In recent years,sound source localization technology has been widely used in video conferencing,artificial intelligence,security monitoring,intelligent robots,etc.The traditional sound source localization technology has low accuracy under low SNR conditions.It's susceptible to reverberation interference,and is difficult to process nonstationary and non-linear sound signals.The practical application scenarios are limited.This paper takes indoor near-field sound source localization as the research background,conducts in-depth analysis and research from two aspects of delay estimation and location algorithm,and improves the traditional sound source location technology.Sound sources are classified according to the different types of sound sources and propagation properties.The corresponding geometric positional relationship between the near-field source and the far-field source is discussed in this paper.Finally,a fivedimensional cross space array model is constructed.The signal model is established,and the acoustic signal is pre-emphasized.Aiming at the problems of large error and poor positioning effect in current sound source localization technology,a time delay estimation and localization method based on improved MUSIC algorithm and objective function search method is presented.First estimate the delay of the sound source reaching each microphone,and then estimate the distance relationship between the sound source and the array.The weighting function is used to weight the MUSIC algorithm to identify the subspace of the sound source.Then,the objective function based on the least squares method is used to perform the iterative search in the subspace to obtain the optimal solution.At the same time,the array topology is more reasonable through the microphone array correction method.Such methods can effectively avoid errors caused by local optimal solutions.Aiming at the problem that the time delay estimation algorithm in the positioning technology has large calculation amount,poor anti-noise ability and low accuracy of estimation of non-periodic and nonlinear signals,an improved quadratic correlation delay algorithm is proposed.Based on the generalized cross-correlation delay algorithm,the algorithm combines EEMD and LMS algorithms to improve the signal-to-noise ratio and greatly reduce the noise interference.At the same time,a secondary correlation operation is performed to estimate the delay,and a weighting function index factor is introduced in the cross-correlation process.The simulation results showed that the improved time delay estimation algorithm can achieve better processing effect for both periodic and aperiodic signals compared with GCC and LMS.The ratio of outliers is about 1/2 of the previous algorithm.The performance of the algorithm was improved by about 2.5dB.The error of the improved positioning technology is also greatly reduced compared with the traditional algorithm,especially in the low SNR condition.