Speaker localization for human-machine interaction

• Main Author: Muhammad Azlan , Ahmad
• Format: Undergraduates Project Papers
• Language: English
• Published: 2012
• Subjects: TA Engineering (General). Civil engineering (General)
• Online Access: http://umpir.ump.edu.my/id/eprint/7606/1/CD6724.pdf; http://umpir.ump.edu.my/id/eprint/7606/
• Institution: Universiti Malaysia Pahang

Speaker localization is the study direction of sound detection. A listener can identify the location or origin of sound. Human auditory system mechanisms of speaker localization have been widely studied. To adapt the ability of the human auditory system, the researcher uses several methods such as time and level differences between the ears, spectral information, timing analysis, correlation analysis, and pattern matching. The application of the system is implemented to the robot as a model human auditory system. The robot now can interact naturally when human speak to them as it can determine the localization of the speaker. In this paper is to build speaker localization where to determine the sound source in front side of 180 degrees. The azimuth, distance, and also height will be fixed. The sensor is built from the three microphones to detect sound. The speaker source that came will be comparing its intensity at the right and the left side. So, if the source is from the right side, then the servo motor will rotate in the right until the different intensity of sound is same to the both sides. Therefore, the servo motor will stop here. Whereas, this will show the direction of the sound source. The result in the experiment shows that it’s quite satisfying with error accuracy as this technique need a more sensitive instrument to compare the sound source.