Today, the emergence of autonomous robots is becoming increasingly significant in different industrial and residential applications. To be able to truly fulfill their mission without external intervention, autonomous robots are required to be able to perceive their environment adequately and with sufficient detail. It is possible to detect and process several parameters of the environment, which lets the robot to function properly. An environmental parameter used in certain cases by these robots is the incoming sound.
The incoming sounds must be processed by the robot to understand the inherent information. The first steps in the processing are to detect the sound sources in the incoming mixture and then separate those sources from each other. After the separation it is possible to make further processing steps for example speech recognition. Another important step after the separation may be the localization of the detected sources.
In this paper I present multiple audio separation algorithms and several methods for the source localization. Then, by modeling a real physical layout, I present observations and tests for several audio separation algorithms with sound recordings made by me. Then I present a localization algorithm I developed and the tests of this algorithm. After processing the test results, it can be determined that the methods which are described in this paper provide a solution for separation and localization problems.