Balanced Data Focusing: Direction of Arrival Estimation with Maximum LikelihoodPerformance

摘要

The performance of a direction of arrival estimation procedure at low signal- to-noise ratios and limited data samples is an important characteristic. The approach based on maximum likelihood (ML) estimation is considered to be among the best for this problem as long as the underlying signal model is properly chosen. Unfortunately, in most cases, there is no closed-form solution so fast search procedures are employed. Given to a priori knowledge, selecting the initial parameter values for these search procedures can be a difficult problem, especially under low signal-to-noise conditions. In this paper, a new method for uniform linear sensor arrays which overcomes the initial value problem is introduced. This method is called Balanced Data Focusing (BDF). Simulation results are included comparing the performance of this new method to that of the ML approach using the Alternating Projection Maximization search procedure and another popular estimation approach, the root-MUSiC method.