An Introduction to Random Optimum and Adaptive Signal Processing下载

In this course we will consider the following key themes Optimisation—and in general we restrict our attention to quadratic cost functions. Adaption—we consider a number of algorithms for iteratively converging to and tracking the optimum solution. Statistical Signals—this necessitates some understanding of the theory of random processes. We will mainly consider discrete time linear filters—which implies real signals, however in many antenna systems we have quadrature receivers which we can treat as complex signals with the real and imaginary parts representing the inphase and quadrature components respectively. Although we probably won’t consider it in detail, the techniques here can readily be extended to linear combination of multichannel signals (e.g. beamforming). We will also consider spectrum analysis in the context of constrained optimisation and subspace methods. Because of its central role in estimation and as an introduction to lattice filters we will also consider in detail optimal linear prediction.