Research On Phase Processing Methods Of Distributed Signal

The abilities to adapt complex environments and acquire reliabel and acurate information have enabled the development of modern information acquisition and processing techniques. With the development of these techniques, distributed system composed of a large number of sensors, high speed communication links and comprehensive processing center, is widely used in target detection, environment sense and information reconnaissance to provide collaborative operations and anti-jamming abilities.This dissertation discusses the phase processing methods of distributed signal. Starting with distributed signal magnitude vector model, the direction finding algorithms which are insensitive to phase noise and posse super-resolution ability are proposed through converting direction of arrival(DOA) estimation into angle difference of arrival(ADOA) estimation. Then the collaborative beamforming algorithms based on phase processing of distributed signal are discussed. Their phase requisition and capabilities of anti-jamming are also presented. The main contents are included as the following:1. With the analysis of distributed signal magnitude vector model, a distributed signal phaseless model which is insensitive to phase noise is presented. Based on this model, a distributed signal ADOA estimate algorithm is then deduced. This algorithm avoiding phase error estimation can be used for array signal processing not only with phase error but also without phase error. The Cramér-Rao Lower Bound(CRLB) of this algorithm is also provided as a benchmark.2. A subspace based ADOA estimation algorithm is then proposed utilizing distributed signal magnitude vector. Similar with MUSIC algorithm for DOA estimate, this algorithm is independent of array geometry by utilizing the orthogonality of array manifold and scanning vector in distributed signal magnitude vector model.3. For single snapshot situation, a distributed signal magnitude vector based ADOA estimation algorithm is proposed with sparse constraint. The strategy for choosing regular parameter is discussed for high precision estimation of ADOA under high signal-noise-ratio(SNR).4. For multiple snapshots, the ADOA estimation algorithms with distributed signal sparse constraint, including subspace algorithm and covariance matrix fitting algorithm, are proposed. The strategy for choosing regular parameter is also provided.5. A self-localization method with closed-form solution is presented utilizing the distributed ADOA estimates with unknown reference direction. A set of pseudo linear equations are first constructed to provide the closed-form of node position under a coordinate transformation. For improving the adaptability and availability of this algorithm, a method of eliminating geometric limitation is offered by Residual test.6. Utilizing linearly constrainted minimum power(LCMP) beamforming algorithm in array signal processing, a distributed signal collaborative beamforming algorithm is proposed based on linear constraint. By adding some constraints and robust cost function, this algorithm is much robust to the impulse noise and estimation error.7. Assuming that the node position distribution and the probability density function of phase error can be modeled with non-parameter kernels, the characteristic of distributed signal collaborative beamforming algorithm is analyzed with far field pattern. This method can overcome the flaws found in existing methods and can provide a comprehensive and effective benchmark for performance analysis and assess of distributed signal collaborative beamforming algorithm.