Fast and efficient DOA estimation method for signals with known waveforms using nonuniform linear arrays

In this paper, a new approach is proposed to estimate the direction of arrival (DOA) of multiple non-coherent source signals with known waveforms but unknown gains based on a nonuniform linear sensor array. Unlike some previous methods, which estimate the DOA using spatial signatures of the signals with known waveforms, the proposed method first uses the known waveforms and mutually independent sensor measurement noises to establish a maximum likelihood estimation problem corresponding to multiple linear regression models, each containing the DOA and the gain information of all the source signals. Then, regression analysis is performed to estimate the coefficients of each linear regression model, and the well-known generalized least squares is used to obtain the estimates of the angles and gains from the estimated regression coefficients. The proposed method does not require a search over a large region of the parameter space, which is normally needed in ML-based DOA estimation methods. The effect of correlated sources on the performance of the parameter estimation is also studied. It is shown that the DOA and gain estimates are asymptotically optimal as the sources tend to be uncorrelated. Finally, simulation results that demonstrate the estimation performance of the proposed method are given. HighlightsWe propose a new method to estimate the direction of arrival (DOA) of non-coherent source signals with known waveforms.We exploit nonuniform linear array in conjunction with linear regression model for DOA estimation.The effect of correlated sources on the performance of parameter estimation is also investigated.Our method estimates the parameters with linear operation and has lower computational cost compared to existing methods.