Home > Research > Research Areas  
Sound Source Localization for Robot Auditory System
Robot Auditory System
- It is very important function for Human-Robot Interaction. It can make Robot to recognize what and where someone said.

- It mainly consists of the speech/speaker recognition and the sound source localization

Research Objectives
- Develop the Sound Source Localization method to satisfy some conditions for robot auditory system

- Localize the source in 3D spaces

- Use the minimum number of microphones for reducing the hardware requirements.

The Summed GCC method
- This is based on TDOA (Time Delay of Arrival) for localization.

- The cross correlation function to estimate the time delay is converted to the mapped GCC function to estimate the source position.

- When multi-microphones are used for localization, estimation results of all microphone pairs can be easily combined in the global coordinate.

- It is possible to localize the source in 3D spaces with only three microphones to utilize the robot platform effect, although the conventional method should have the minimum four microphones to do.

- This shows the better performance than the conventional one in the worse environment conditions, that is a bad SNR condition.