Cross-correlation function based two-sensor auditory localization unit for chat robots

This paper mainly focus on providing an auditory localization unit for robots. Due to auditory localization ability, human beings can turn their face to face the speaker when they are communicating with each other. This action can make communication more efficient and have a better mutual understanding. For robots, this ability will improve quality of their imaging processing and enhance voice recognition. In this paper, we designed a two-sensor auditory localization unit for robots. Firstly, we analyzed the mathematical model of different sound field, the far field and near field. Based on that, we analyzed features of received signal, for far sound field, we designed cross-correlation function algorithm to localize direction of the sound source using phase delay information, for near sound field, we designed cross-correlation function with amplitude attenuation information algorithm to get exact two-dimension location information. According to the results of simulation we made our prototype and based on the test results we got our actual performance data.