Multiple snapshot matching pursuit for direction of arrival (DOA) estimation

The Matching Pursuit (MP) algorithm may be used to give a low complexity method of obtaining a solution to the directions of arrival (DOA) problem which arises in a large number of application areas. We present a novel estimation of the DOA in which we incorporate multiple snapshots of the received signal into each run of the MP algorithm. A tree based search is used in combination with MP to expand the search space. This new method, MSMPDOA, gives a substantial performance improvement over running the MP algorithm on each snapshot separately before the results are aggregated to give the DOA estimate. In a series of simulations, it is shown that the success rate in accurately determining arrival directions at low values of SNR is improved by 10-30% by utilizing the MSMPDOA algorithm.