| Polarization Sensitive Array(PSA) can acquire the polarization information of electromagatic signals. Compared to the common array, PSA has many advantages such as resisting interference effectively, detecting signals robustly, sharpness of separation and polarization Multiple Access. So it has a good prospect of application in radar, communication, navigation and so on. The parallel factor (PARAFAC) was first introduced as a data analysis tool in the physiology. And most of its application is in chemometrics, spectroscopy and chromatographic scopes. It is a measure of multidimensional data analyses. Recently, PARAFAC technique is becoming widely used in signal processing and communication.The received signal of the Polarization Sensitive Array is proved to have trilinear model characteristics, so the blind PARAFAC signal detection and separation algorithm is proposed and researched in this paper, which relies on the uniqueness of low-rank three-way array decomposition and the trilinear/quadrilinear alternating least square(TALS/ QALS)algorithm. TALS/ QALS is used to obtain the source matrix, then the matrix is decided. The PSA includes Uniform-Linear-Array(ULA), Uniform-Circular- Array(UCA), L-array and Uniform-Square-Array(USA). Lastly, the simulations are done to evaluate the performance of this algorithm. In the simulations, Complex parallel Factor analysis (COMFAC) algorithm is used, which is a kind of TALS algorithms with fast speed. The performance of COMFAC algorithm is very close to TALS algorithm, but the former has better convergence. The simulation results reveal that the performance of blind PARAFAC signal detection algorithm for polarization sensitive array is close to non-blind MMSE method. And the performance is fairly good in low snapshots. Besides, this algorithm works well in array error condition. The algorithm doesn't require DOA (direction of arrival) information and polarization information, so it has blind and robust characteristics. |
PDF Full Text Request