Moving Sound Source Localization in 3-D Space Using Time Delay and Intensity Ratio-Based Analytical Approaches

Moving sound source localization and tracking (MSSLT) has numerous applications in the military, machinery fault localization, hearing aids, teleconferencing, and robotics. However, MSSLT is extremely challenging due to the Doppler effect which alters the frequency and amplitude of the received source signal. Mitigating this effect requires pre-requisite information of the source’s trajectory, speed, initial position, and spectrum. In this paper, we propose two analytical MSSLT techniques, one based on time-delay intensity ratio (TD-IR) and another based on time difference of arrival (TDOA) that do not require any pre-requisite information about source and utilize just 6 and 5 microphones respectively. The techniques are validated through extensive simulations and real life experiments under diverse conditions. TD-IR accurately tracks and localizes moving sources in unrestricted search space but with low speed conditions (upto 15m/s). Whereas, TDOA has restricted source search space, but performs significantly better than TD-IR, and is shown to work until 67m/s source speeds.