ABSTRACT
This chapter focuses on the basic theory and applications of modern adaptive signal processing. It introduces some of the basic ideas and applications of adaptive signal processing, because the subject is so broad. The chapter considers methods for searching the performance surface and tracking the movement of the minimum mean-squared error. The most general adaptive algorithms for searching performance surfaces under unknown signaling conditions are random search algorithms. The chapter focuses on the two most common methods for accomplishing such adjustments in real-time adaptive systems; namely, steepest descent and sequential regression. It discusses the process of searching the quadratic performance surface. The chapter examines how adaptive systems reach convergence under realistic conditions, with imperfect knowledge of the gradient and correlation properties of the performance surface, which may change from one iteration to the next. Recursive filters as well as nonlinear filter structures have been applied in adaptive signal processing.
