An algorithm on high resolution DOA estimation with unknown number of signal sources

A new algorithm to estimate the DOAs without knowing the number of signal sources is proposed in this paper. The new algorithm, using the linear prediction (LP) or Pisarenko method in conjunction with adaptive signal parameter, estimation and classification technique (ASPECT) to estimate the DOAs, is with high resolution and no need knowing the number of signal sources. The ASPECT is used to identify the spurious peaks in DOAs and the number of signal sources can be determined at the same time. Through the algorithm, computational complexity can be reduced and the spectrum resolution can be improved also. Computer simulation results approve the correctness and validity of the new algorithm.