Research Of Real Time Sound Source Localization Technology By Microphone Array

In recent years with the popularity of intelligent life, researches on sound localization technology more and more, it is in the civilian and military aspects have a good prospect. However, the domestic research on this technology is mostly biased civilian aspect, less in military terms, and many are focusing on how to increase the accuracy of positioning algorithm for computing the amount of the decrease is less. Therefore, to defense needs, conduct research in real-time sound source localization algorithm has important practical value. In this paper, how to reduce the amount of localization algorithm calculation algorithm is improved, and set up a test system algorithm was validated.First, about the test at the beginning of the positioning speech activity, distinguished array near-field and far-field signals signal model model, analyzes the similarities and differences between the two signal model, focusing on the design elements of the microphone array.Secondly, it analyzes the current most popular SRP-PHAT algorithm, and for reducing the amount of calculation algorithm has been improved, using the estimated value of the delay time method based on GCC SRP-PHAT algorithm reduces the search area. In addition, the introduction of a space-based sampling to improve the SRP-PHAT algorithm, and improve before and after the three algorithms by simulation, positioning accuracy compared with calculated amount.Again, build a real-time sound source localization test systems. Detailed description of the hardware and software program designed to test the system. Write the PC software display interface so that you can clearly understand the results of observations sound source localization.Finally, the use of hybrid structure proposed algorithm built to run tests on a test system. The measured data show that the proposed algorithm hybrid structure while maintaining the original algorithm positioning accuracy and robustness while significantly reduces the computational algorithm for real-time positioning environment.