Research On Location Method Of Whistling Vehicles Based On Microphone Array

With the rapid development of our country’s economy,more and more vehicles are used as means of transportation.While all kinds of vehicles bring convenience to people and create economic value,they also cause automobile noise problems that cannot be ignored.The whistling noise of vehicles is an important part of traffic noise and seriously affects people’s lives.In order to improve the urban environment,many cities in our country have formulated relevant laws and regulations to prohibit vehicles from whistling.However,due to the huge traffic volume and the moving state of vehicles,there are problems such as low efficiency and poor reliability in the identification and positioning of whistling vehicles.For this reason,based on the characteristics of sound propagation,this paper puts forward a more effective positioning method of the horn vehicles through the comparative analysis of multiple positioning algorithms of the microphone array,and verified by experiments.Firstly,for the positioning method based on time delay estimation,based on detailed analysis of the sound field characteristics and whistle signal characteristics,the comprehensive performance of different time delay estimation algorithms and positioning algorithms is compared and analyzed through simulation.The simulation analysis results show that the combination of the minimum mean square error adaptive filtering time delay estimation algorithm and the linear closed-form algorithm can take into account the stability and accuracy of sound source localization.Secondly,the beam scanning azimuth spectrum estimation algorithm is researched.The performance of different beamformers and their application effects in the beam scanning azimuth spectrum estimation algorithm are compared.The simulation results show that the MVDR(Minimum Variance Without Distortion Response)beamformer has a high gain.When it is applied to the beam scanning azimuth spectrum estimation,the azimuth estimation resolution is better.Thirdly,the high-resolution azimuth spectrum estimation algorithm is analyzed.Compared with the traditional multi-signal classification(MUSIC)algorithm in the array element domain,it is found that using the MVDR beamformer as a pre-processing tool in the beam-domain MUSIC algorithm can effectively improve the azimuth resolution ability.Then the particle swarm optimization algorithm is applied to the azimuth search,and the multiple search method of modifying the fitness function is used to solve the problem that the particle swarm optimization algorithm can only search for a single sound source,which effectively improves the positioning accuracy,and is verified by experimental data.Finally,a uniform ring microphone array is designed and a test system is built.An active correction algorithm is used to correct the amplitude and phase errors of the microphone array.The positioning effect of the above algorithms is verified and comparatively analyzed through horn test and actual vehicle test.The test results show that the combination of beam domain multiple signal classification algorithm and particle swarm optimization algorithm can effectively locate the whistling vehicle,which is suitable for the illegal whistling capture system and has a certain engineering application value.