Research Of Array Signal Processing Based On Non-circularity

In array signal processing, using the signal non-circularity to enhance the performance of DOA estimation, blind channel identification and blind source separation is a prosperous area. Most of existing array signals processing methods are based on the homogeneous circular or non-circular sources. However, mixed circular and non-circular sources are frequently encountered in increasingly complex situations. And how to use the circular difference of two kinds of signals to improve adaptive beamforming is a new problem need to be considered.In this thesis, several algorithms for mixed circular and noncircular signals are proposed to overcome the above knots. The main research work and contributions are detailed as follows.1.By analyzing the circularity of the complex random signal in detail, we give out the definition of the non-circular signal and some important property. Furthermore, we analyzed the CRLB of circular and non-circular source. We proved that non-circular Gaussian signal include more information than circular Gaussian signal, because its distribution is determined by both covariance matrix and eclipse covariance matrix.2.By analyzing the subspace property of the mixed circular and non-circular signal, four improved algorithms based on spectra MUSIC, root MUSIC, ESPRIT and Unitary ESPRIT are proposed under the mixed circular and non-circular circumstance. And four different identification methods for separate circular and non-circular source are also proposed. Simulation results showed that the new algorithm can estimate both DOAs of the circular and non-circular sources, and has better estimation performance that those regular algorithms. The new algorithm also can handle more mixed sources than array elements.3.A new adaptive beamforming criterion is proposed using the non-circularity of the signal, which is called Maximizing Non-circular to Circular Rate (MNCR) in mixed non-circular signal with circular interference and noise situation. And a new algorithm is derived based on MNCR. Compared by MVDR and MSINR, the new algorithm shows good performance which is nearly same as MSINR and better than MVDR. Since the new algorithm is only using the property of the signal, it is a blind beamforming algorithm and is robust.4.Since GPS signal is a non-circular signal, we extend the non-circular research to GPS anti-jamm application. Two new anti-jamm algorithms are proposed. Those algorithms based on non-circularity of GPS signal and circularity of jamm signal, respectively. Simulations have shown that both algorithms have better anti-jamm performance than the normally used algorithms.