| Acoustic information acquisition technology based on sound pressure sensor array has been paid much attention by scholars.It is widely used in speech signal acquisition and military fields,but there is relatively little research on sound source location and recognition.In array signal processing,most of the researches focus on radar detection and other fields with high real-time requirements,and there are few researches on array signals with relatively low propagation speed.The research of spatial sound source signal location is of practical application value,which can locate specific sound sources in the environment and facilitate the quick detection of environmental noise and the collection of specified sound source information.Based on the study of a large number of literatures,this paper analyzes and compares the principle and advantages and disadvantages of array location of different structures,and proposes a 20-element hemispherical conformal array based on C60 structure.Based on the array structure,MUSIC algorithm and directional graph algorithm are used to study the array's positioning performance for multiple sound sources.Simulation results show that the performance of MUSIC algorithm is better than that of directional graph.When locating three sound sources,the directional graph method can only locate the position of one of them,with an error of 1?3°,while the MUSIC method can locate the position of three sound sources,with a stable error of 1?2°.An improved Multistage Gerschgorin disk estimator(MSGDE)was proposed for spatial multi-sound source number recognition,and the classical sound source number recognition algorithms in GDE,AIC and MDL were introduced to verify the accuracy of this algorithm for multi-sound source recognition.Simulation results show that MSGDE method is better than the other three methods in recognizing the number of multiple sources in spatial color noise and white noise environment,and based on the array structure in this paper,the theoretical limit of this method is 19.Finally,through the combination of software and hardware,the front-end data acquisition system with STM32 as the main control chip and the Lab VIEW data analysis and display platform,the actual sound source is located.The system can accurately locate three sound sources in space,and the error is controlled at1?2°.When studying the selection of adjustment factor,this paper used the algorithm to estimate the number of sound sources under the condition that the number of sound sources was known,and studied the distribution law of the multi-stage adjustment factor D(n),so asto obtain the optimal value.Simulation results show that the optimal adjustment factor selected by this method can adapt to the recognition of the number of multiple sources.MSGDE algorithm is actually similar to the multiple operations of traditional GDE algorithm.Based on the assumption,it carries out multiple classification operations on the same data,which greatly reduces the early estimation error generated by a single operation and improves the accuracy of the number of sound sources.Compared with traditional GDE,in the range of-20 d B ?10d B SNR given in the experiment,the recognition accuracy is generally improved by 10%?80%.Based on the research method in this paper and the proposed algorithm,the performance of sound source number recognition and location will be improved by more efficient computing methods. |
PDF Full Text Request