Research Information System
5th International Palestra Scientific Research Congress, 4 - 05 May 2025, pp.161-166, (Full Text)
Sound source localization covers a very wide area of study. The need for the location of the sound source can have a wide variety of purposes. In some applications, it is aimed to isolate and strengthen the sound at a certain location and to suppress other sounds in the environment. Especially in crowded meeting rooms, sound location detection is used to detect the location of the speaker with a microphone array and to suppress the noise other than the speaker's voice with beam formation. In hearing aids, it is aimed to detect the direction of the sound with microphone arrays and to gain a hearing sense that feels more natural and higher quality in the user's daily life. In automobiles, microphone arrays are also used to suppress noises such as engine and traffic by focusing on the speaker's voice to enable hands-free phone calls. In some other applications, the location of unwanted sounds such as noise is the main purpose. In this study, it is aimed to evaluate the advantages and disadvantages by introducing the methods that can be used to detect the location of the source causing unwanted sounds coming from mechanical systems located in a relatively narrow volume. In this way, fault detection can be made depending on the source of the unwanted sound. On the other hand, in some cases, the cause of the sound may be due to design. In such cases, design changes can be made to prevent unwanted sounds. In order to determine the source of the sound, first of all, synchronous sound recording is performed with multiple microphone arrays placed in a certain order (linear, circular, 3D grid, etc.). The obtained sound data is filtered and purified from interfering sounds. Then, frequency analysis is performed to determine the frequencies of dominant or disturbing sounds. One of the methods used to determine the location of certain sounds is to analyze the difference in the time it takes for sound waves to reach different microphones. In the beamforming method, the signals received from the microphones can be combined and the direction and location of the sound source can be analyzed. 2D or 3D mapping of sound intensity can be done using acoustic cameras. Innovative approaches to detect the source of noise can also be used to train artificial intelligence models. Thanks to these models, the location of the sound source is mapped in a 2D or 3D environment. Then, the problem area or defective part is detected by evaluating the frequency structure and intensity of the sound.