Direction of arrival estimation using local series expansion evaluated on acoustic data
Gustaf Hendeby, Thomas Lunner, Fredrik Gustafsson · The Journal of the Acoustical Society of America · 2017
Direction of arrival (DOA) estimation using a local series expansion of the signal is evaluated on acoustic data. It was recently proposed that DOA estimation using a sensor array could be performed using a local temporal/spatial series expansion. That is, to use the local phase of the signal rather than its estimated second order moments or correlation as in classical methods. The approach should provide easier handling of general sensor configurations, ability to deal with very low sampling rates and few samples, and benefits from miniature sensor configurations. Additionally, the signal is not required to be narrow-banded, as sufficient smoothness is enough. Here, the local series expansion method is benchmarked to a standard beam-forming method using experimental data, recorded in an anechoic chamber at the Technical University of Denmark (DTU). The dataset was recorded using 52 high quality microphones distributed on a sphere with radius 50 mm, and contains several recordings from one or two well defined sources from several different directions, with different levels of reverberation. The dataset is used to evaluate important aspects of the new method and its potential to improve on standard DOA methods.