In the last chapter, we reviewed the basics of applying DSP techniques for Fi-Wi link distortion compensation. In this chapter, we review adaptive filtering techniques. Higher-order adaptive filters can be used to model Fi-Wi systems. Adaptive filters can also be used to inverse model the Fi-Wi link and, hence, compensate the distortions resulting from the transmission over Fi-Wi links.
Adaptive filters offer significant improvement over fixed filter solutions in environments of unknown statistics . This nature can be very well exploited by using adaptive filters for modeling and compensation of Fi-Wi links. Typically, the exact nonlinear characteristics of a Fi-Wi system are unknown since the nonlinearity arises from a multitude of devices and processes.
In this chapter, we consider an additive white Gaussian noise (AWGN) wireless channel and a memoryless nonlinear fiber channel. The nonlinearity compensation becomes fairly straightforward when the multipath effects of the wireless channel can be ignored. The AWGN wireless channel assumption is valid when the radio cell is small with a strong line of sight reception. This could be the case in microcell scenarios.
An adaptive filter is a filter that self-adjusts its transfer function according to an optimization algorithm to minimize an error signal. Widely used adaptive algorithms such as least mean squares (LMS) or recursive ...