Study On Algorithm Of Signal Parameters Estimation Based On Polarization Array

The elements of polarization sensitive array are composed of many dipoles and loops pointing to different directions. Compared with traditional arrays, it can realize whole polarization receiving of electromagnetic signals, and then have stable and high gain, the research existed has proved the system advantage of this new kind of array.This passage has researched the algorithms of parameters estimation based on this kind of array to improve the algorithms.The main contents are concluded as follows:1. The rapid algorithms based on polarization array are studied. Aiming to solve the four dimension parameters search of polarization MUSIC, an algorithm called polarization-space joint spectrum MUSIC algorithm via metropolis-hastings sampler is presented, compared with original algorithm using traditional searching mode, the computation is effectively reduced while performance is not decreased. Aiming to solve multi-dimension parameters search of polarization WSF, an algorithm called polarization WSF based on differential evolution is presented. The simulations proved that the algorithm presents good performance and has rapid convergence, the success rate is high.2. Parameters estimation algorithms based on non-circular signals are studied. Making use of signals'non-circular character is an important way to impove algorithms'performance. This paper firstly introduces the polarization NC-MUSIC (Non-circular MUSIC) algorithm, then considers the real condition that circular and non-circular signals exist simultaneously, and presents the solvement. Aiming to reduce computation of the algorithm, this paper presents an algorithm called polarization propagator method based on non-circular signals. This algorithm doesn't need eigenvalue decomposition and then reduce the computation while having the same performance with the polarization NC-MUSIC.3. Parameters estimation based on single electromagnetic vector sensor is studied. The performan of algorithm is bad when the parameters of signal vary with time.This paper introduces an signal tracking algorithm based on electromagnetic vector sensor, aiming to the bad performance of the algorithm for using receiving data directly, presents an signal tracking algorithm based on projection approximation subspace updates, the algorithm doesn't need eigenvalue decomposition, so the amount of calculation is small. It converges fast and has better performance.