Estimation of Sound Direction for Multi-Sound Field Using Genetic Algorithm
It is a future goal of this research to develop the easy hearing aid with noise reduction and sound localization junction to hearing impaired. Though the detection of object sound signal and sound direction in the multiple sound fteld is the fundamental acoustic measurement, the high-performance measuring method has not been developed yet. This paper proposed the multi-microphone system using Genetic algorithm for estimating the multiple sound direction. Though several estimation methods of the sound direction have been reported until now, the estimable angle number is limited to at most 3 angles. This system is composed of three-dimensionatty spaced microphone array and FIR filter based on wavelength constant matrix that is linked for it. The wavelength constant matrix is formulated in order to accurately give the relationship between microphone output and individual sound wave at the origin. Under the condition of knowing the sound direction, this system can detect the individual sound wave in the multiple sound field. Detection error in this case is dependent on the accuracy of the sound direction. This fact indicates that the sound direction can be estimated, if the detection error is made to be correspondent to the fitness in Genetic algorithm. The computer simulation based on this idea showed that proposed method was fundamentally valid for the detection of sound direction up to the 6 angles of 3 sound sources until now.
J.Kawano T.Ogawa K.Kuroiwa O.Hoshino
Oita University 700 Dannoharu, Oita, 870-1192 Japan
国际会议
8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)
上海
英文
561-566
2001-11-14（万方平台首次上网日期，不代表论文的发表时间）