A Specific Sound Source Localization Method Based on Improved TDOA
Sound sources are Filtered and located in complex scenarios,there are the shortages of low signal-to-noise ratio,un-practical linearity,and multipath effects,often requiring high hardware requirements for sound source localization equipment.To solve this problem,this paper proposes a method for using the fuzzy set filtering to process the collected raw data and optimizing the intermediate variables of the time difference of arrival(TDOA)algorithm.The original audio signal is collected through a six-element microphone array,and the membership function is used to project the data into the fuzzy set.According to the size of the membership value,the data are weighted to obtain the ideal audio data.Then,the processed data are fed into the grid iterative search method to obtain the localization result of the sound source.To verify the algorithm,an embedded sound source localization system is built using STM32 as the main control chip.Under the same conditions,the proposed method is compared with the TDOA algorithm with gener-alized cross-correlation to obtain the time difference.The results show that with a locating audio of same frequency,this method has better anti-interference and localization accuracy,meeting the goal of improving the anti-interference ability of sound source localiza-tion through algorithm improvement,and reducing hardware collection devices in practical applications.
万方数据
维普
