In this talk, we revisit multi-source localization from the perspective of the time-domain wave equation. Under a high-frequency approximation, the wave equation reduces to the eikonal equation, whose solutions describe travel-time fields. We show how numerical solutions of the forward problem governed by the eikonal equation can be used to address the inverse problem of localizing impulsive sources in topographically complex environments. We also consider classical free-field solvers and the problem of localizing multiple sources in the presence of clutter, timing uncertainty, incomplete data, and an unknown source count. To address these challenges, we introduce a combinatorial coalition-based model solved via set packing, enabling joint data association, source counting, and multi-source localization. Monte Carlo simulations and acoustic experiments in urban environments illustrate the behavior of the proposed methods. The presentation emphasizes the conceptual bridge between wave propagation and robust multi-source localization.

For further information about the seminar, please visit this webpage.