Ability of adaptive filters to track carrier offsets and channel nonstationarities

Tracking Analysis of Affine Projection Filters to Carrier Frequency Offset

“…Then, based on the energy conservation relation, the tracking performance of the APA to the carrier frequency offset (CFO) is analyzed. This result is compared with the tracking performance of NLMS in [4][5] and the performance of the APA without CFO in [7]. Simulation results show that the analysis and the derived performance expressions are close to the simulation curves.…”

“…The variations of the system or channel can be classified as random system nonstationarities and cyclic system nonstationarities [2]. For the last decades, the ability of adaptive filters to track random system nonstationarities has received considerable attention [3] [4][5] [6]. The cyclic system nonstationarities, which is generated by the mismatches between the transmitter and receiver carrier generators, degrade the performance of adaptive filters seriously, even for ever small carrier frequency offsets.…”

“…The ability of adaptive filtering algorithms to track such cyclic system variations is not yet fully understood. A recent contribution of Yousef in this regard [4], who presents a framework for the tracking analysis of adaptive algorithms that handles simultaneously both cyclic and random system nonstationarities. But this framework is restricted to scalar adaptive algorithms, such as Least Mean-Square (LMS) algorithm, Normalized LMS (NLMS) algorithm.…”

“…In general, in the case of a fading channel, the perturbation θ(k) can be modeled to a reasonable extent as an autoregressive (AR) process of order 1 [4]:…”

Least Mean-Phase Adaptive Filters With Application to Communications Systems

“…Overall, this results in a lower symbol error rate in fast fading wireless channels. Note that the LMS algorithm alone has a limited tracking performance [3], and the step-size parameter cannot be chosen unboundedly large in order to track fast channel variations, since it will make the filter unstable. For such a scenario, the phaseerror term enhances the phase tracking performance of the filter (and, consequently, the symbol-error-rate), without sacrificing the stability of the filter.…”