Improving DOA estimation methods using a priori knowledge about the sources location

The linear prediction technique applied to a sensor array can provide useful information about the source´s location. A simple relation between the minimum-norm solution for the optimum filter and the source´s direction is derived. Direction of arrival (DOA) estimation methods can use this information to improve its resolution capability. An algorithm applied to the two closely spaced sources case is given to show the improvement achieved. In addition, a subspace-based approach to the use of a priori information about the source´s direction is presented