Model Based Direction of Arrival Estimation Methods Applied to Experimental211 Antenna Data

摘要

This report compares some model based methods for DOA (Direction Of Arrival)211u001eestimation in radar applications. The report consists of three parts. The first 211u001epart gives the theory of some DOA estimation methods. The studied methods in this 211u001ereport are Conventional BeamForming (CBF), Capon's Multiple SIgnal Classification 211u001e(MUSIC), Min-Norm, Deterministic Maximum Likelihood (DML), Stochastic Maximum 211u001eLikelihood (SML), Signal Subspace Fitting (SSF), Weighted Subspace Fitting (WSF), 211u001eEstimation of Signal Parameters via Rotational Invariance Techniques (ESPRIT) and 211u001eRoot-MUSIC. These methods are implemented in MatLab. In the second part, 211u001eperformance of the DOA methods using simulated signals are studied. In the third 211u001epart, the methods are studied using real data. The real data is collected from a 211u001e12 channel digital beamforming array antenna. The antenna is not ideal and some 211u001ecorrection processing is done before evaluation of DOA. The standard deviation 211u001eand the mean value of the estimated DOA have been studied. The DOA estimation as 211u001ea function of the number samples and power of the signals has also been studied. 211u001eIn these algorithms, the noise has been assumed to be independent white Gaussian 211u001enoise. The noise characteristic has been studied and it agrees with the 211u001eassumptions in the algorithms.