Research On Using Single Motion Microphone To Recognize Sound Source Based On Time Delay Estimation

The control methods of noise are mainly divided into three types: suppressing noise at the sound source,controlling in the sound transmission path and protecting the receiver.Noise suppression at the sound source is a more effective noise control method.Only by accurately identifying the location of the noise and studying and analyzing the parameter information of the noise can the noise be more effectively suppressed.At present,the commonly used noise recognition methods all need to use a microphone array.The microphone array consists of multiple microphones,ranging from a dozen to hundreds.Moreover,the microphone array has many problems.Before measurement,matching and calibration work is required,and the measurement cost is relatively large.In this paper,the use of a single moving microphone for sound source identification will be studied based on the microphone array combined with the time-of-flight method to identify the sound source.The basic principle of its identification is to use the time delay estimation method to calculate the time difference between the sound signal reaching the corresponding microphone pair.Then using these time differences calculate the position of the sound source,combined with the spatial geometric relationship of the microphone array.Explore the use of a single-moving microphone combined with the time-of-flight method to identify sound sources.Derive the formula of sound pressure radiation from a stationary point sound source to a mobile microphone.Build a model of mobile microphone to collect the point source signal in the plane and in the space.Firstly,use the microphone to collect the sound pressure signal at rest,and do spectrum analysis to identify the frequency of the sound source.Let the microphone move at a low speed according to the preset track,and continuously collect signals during the movement.Because the signal collected by a single microphone is different from the signal received by the microphone array,the method of the array cannot be directly applied.Segment the signal at equal intervals.Use the time delay estimation method to calculate the time difference between each signal and the first signal,and integrate them into vectors in sequence.Construct a virtual point sound source.Using the sound pressure radiation function of the stationary point sound source to the moving microphone,the signal obtained by collecting the virtual sound source is calculated,and the same processing as the actual sound source signal is performed to obtain the time difference vector corresponding to each virtual point sound source.The position of the actual point sound source is the virtual point sound source position,which corresponds to the vector closest to the actual point sound source.This method can identify the location of the sound source and has a certain noise resistance through simulation calculations.Investigate the effects of different segment time lengths and different sound source signal frequencies,as well as the speed of the microphone movement speed and the deviation of the movement speed on the sound source recognition results.Analyzing the results of simulation calculations,it is found that the sound source recognition result is related to the length of the signal segment.When the length of the segmentation time is less than 5 signal cycles,the recognition result has a large error.Analysis found that when the length of the segmentation time is an integer multiple of the signal period,the position of the sound source can be calculated using the geometric method.This reduces the amount of calculation and improves the calculation efficiency.Analysis of the movement speed of the microphone found that the movement speed of the microphone should not be too fast.When the speed is very high,a larger recognition error will occur.Small speed deviation has little effect on the result of sound source recognition.In order to reduce the influence of reverberation and environmental noise on the experiment,do the experiment of sound source recognition in semi-anechoic room.Identify the object is a non-directional spherical sound source.Use a high-sensitivity sensor to move uniformly along a preset track.Acquire sound pressure signals continuously during exercise.Import the acquired signal into the preset program for calculation.This method can effectively identify the sound source though the experiment.Measure sound sources with different frequencies and different depths.