Exploitation of Modulation Properties For Blind Adaptive Signal Processing

正在研究开发和评估利用随机调制的正弦波和脉冲序列信号的特性来完成检测和到达时间差(TDOA)估计、相控阵到达方向(DOA)估计和波束/零点定向信号提取以及周期性时变信号校正的算法。目前正在考虑单平台和多平台接收。正在开发的特性包括由循环平稳性产生的光谱自相干，以及周期或恒模特性。正在研究的算法执行盲自适应信号处理，其绕过了对训练信号、导频信号、前同步码、DOA知识等的需要，这些通常是常规自适应处理所需要的。正在研究的检测和TDOA算法使用基于FFT的自相关和互谱相关的测量来执行信号选择性检测和TDOA估计。相控阵算法利用循环相关矩阵的奇异值分解和投影来完成信号选择性波达方向估计，并通过谱自相干恢复来完成波束/零点定向信号的提取。信号校正算法使用载波模数和基带模数恢复来适配周期性时变(频移)滤波器，以抑制同频道干扰，并校正多径和其他失真。一些算法同时使用了自相干恢复和模数恢复。被调查的所有算法所需的先验信息是至少一个循环平稳频率的知识--通常用于感兴趣的信号--例如波特率或载波频率，并且在某些情况下是关于感兴趣的信号和干扰信号两者的此类频率的知识。正在进行的研究基于理论分析和计算机模拟。这项研究聚焦于重大问题，由一位合格的研究人员进行。强烈建议您提供支持。

Research is being conducted to develop and evaluate algorithms that exploit the properties of randomly modulated sine-wave and pulse-train signals to accomplish detection and time-difference-of-arrival (TDOA) estimation, phase-array direction-of-arrival (DOA) estimation and beam/null-steering signal extraction, and periodically time-variant signal correction. Both single-platform and multiple-platform reception is being considered. The properties being exploited include the spectral self-coherence that results from cyclostationarity, and periodic or constant modulus properties. The algorithms being studied perform blind adaptive signal processing which circumvents the need for training signals, pilot signals, preambles, knowledge of DOA, and so on, which are typically required for conventional adaptive processing. The detection and TDOA algorithms under investigation use FFT-based measurements of auto- and cross-spectral correlation to perform signal- selective detection and TDOA estimation. The phase-array algorithms use singular-value-decomposition and projection of cyclic correlation matrices to accomplish signal-selective DOA estimation and to accomplish beam/null-steering signal extraction by performing spectral self-coherence restoral. The signal correction algorithms use both carrier-modulus and base-band-modulus restoral to adapt periodically time-variant (frequency-shifting) filters for cochannel interference rejection, and correction of multipath and other distortion. Some algorithms employ both self-coherence restoral and modulus restoral. The a priori information required by all the algorithms being investigated is knowledge of at least one frequency of cyclostationarity - typically for the signal of interest - such as a baud rate or carrier frequency and, in some cases, knowledge of such frequencies for both the signal of interest and interfering signals. The research being conducted is based on both theoretical analysis and computer simulations. This research focusses on significant problems and is being conducted by a well qualified researcher. Support is highly recommended.