Direction of Arrival Estimation of Wideband Acoustic Wavefields in a Passive Sensing Environment

Abstract: Direction Of Arrival (DOA) estimation has been an active research area for many decades because of its usefulness in many fields, such as wireless communication, navigation, astronomy, surveillance, medicine etc. In DOA estimation, emitted wavefields from unknown sources are measured with the aid of an array of sensors, and this information is then processed to determine various parameters of the sources including the angle between each source and the array. The emitted wavefields could be electromagnetic, seismic or acoustic.Many of the existing methods for DOA estimation were developed in the communication area for dealing with electromagnetic radiation. In this work, we are investigating the problem of DOA estimation for wideband acoustic sources using a microphone array. Since microphones are getting popular in wireless ground sensor networks, this research direction could potentially offer a cost efficient and passive way for detecting, tracking and classifying moving targets. In this thesis, several existing spectral based wideband DOA methods suitable for this purpose are surveyed. Further, a new method based on local polynomial expansion (LPE) is proposed. The performance of these methods is compared by testing them on both simulated and real data, where the real data come from practical field tests with different motorized vehicles. The LPE method is found to be quite competitive for this kind of wideband acoustic sources in real environments.