This work presents advances in the development of a real-time signal processing algorithm that estimates time-domain waveforms of multiple plane wave signals on a sample-by-sample basis from data measured by an acoustic array. Moreover, the algorithm provides sample-by-sample estimates of the direction-of-arrival (DOA) of the waveforms. In this case, sample-by-sample means that as each sample of data is measured by the array estimates of the waveforms and their directions-of-arrival are updated. Previous presentations (Spring 2023 ASA Ottawa, and Spring 2025 ASA New Orleans) described a sliding window-based approach using non-linear least-squares with a B-spline signal basis and an Unscented Kalman Filter (UKF) based approach with a Fourier signal basis. This presentation describes and presents results from a UKF approach using B-splines. The advantage of the UKF over the sliding window is less computations, and the advantage of the B-splines over a Fourier basis is fewer inter-sampleoscillations when estimating the predicted signals. The approach contrasts with blind source separation (BSS) methods that are based on non-Gaussian statistical assumptions and that do not assume a known array geometry nor a propagation model.
Garth Frazier (Wed,) studied this question.