Direction-of-Arrival Estimation Using Signal Processing on Graphs

In this paper, we describe a new approach to single-snapshot direction-of-arrival (DOA) estimation using signal processing on graphs (SPG). Our approach uses SPG principles and tools to translate and generalize the conventional beamformer (CBF) on to a graph domain and give the CBF more processing flexibility. The benefits of this flexibility were analyzed through simulations of our method with inputs from different array sizes and sound source directions. We found that, depending on the choice of graph structure that our SPG-based method models and processes data on, we can produce outputs with more improved DOA estimation accuracy or outputs equivalent to conventional signal processing.