Variants of Partial Update Augmented CLMS Algorithm and Their Performance Analysis

Abstract

Naturally complex-valued information or those presented in complex domain are effectively processed by an augmented complex least-mean-square (ACLMS) algorithm. In some applications, the ACLMS algorithm may be too computationallyand memory-intensive to implement. In this paper, a new algorithm, termed partial-update ACLMS (PU-ACLMS) algorithm is proposed, where only a fraction of the coefficient set is selected to update at each iteration. Doing so, two types of partialupdate schemes are presented referred to as the sequential and stochastic partial-updates, to reduce computational load and power consumption in the corresponding adaptive filter. The computational cost for full-update PU-ACLMS and its partialupdate implementations are discussed. Next, the steady-state mean and mean-square performance of PU-ACLMS for noncircular complex signals are analyzed and closed-form expressions of the steady-state excess mean-square error (EMSE) and mean-square deviation (MSD) are given. Then, employing the weighted energy-conservation relation, the EMSE and MSD learning curves are derived. The simulation results are verified and compared with those of theoretical predictions through numerical examples.

FRGAN: A Blind Face Restoration with Generative Adversarial Networks

“…Over the past few decades, many classical facial landmark detection methods have been proposed in the literature. Parameterized appearance models are represented by Active Appearance Models (AAMs) [21], Constrained Local Models (CLMs) [22], and Cascaded Regression [23].…”

State Estimation in Linear Dynamical Systems By Partial Update Kalman Filtering

“…It is shown that the PUKF algorithm is able to provide an acceptable estimation performance, while the computational complexity is kept low. It is noteworthy to mention that PUKF algorithm has not been addressed in any of references [11] and [16]. Indeed, algorithm in [16] uses the Kalman filter for adaptation and learning in the agents, where nodes are allowed to receive (diffuse) only a subset of their intermediate state estimate entries.…”

“…Indeed, algorithm in [16] uses the Kalman filter for adaptation and learning in the agents, where nodes are allowed to receive (diffuse) only a subset of their intermediate state estimate entries. Algorithm in [11] consider widely linear model, which in this paper we have assumed a linear state-space model. Moreover, in [11] the Augmented Complex LMS algorithm is employed as the learning rule; whereas in our algorithm we have used kalman filter algorithm for adaptation and learning.…”

“…Algorithm in [11] consider widely linear model, which in this paper we have assumed a linear state-space model. Moreover, in [11] the Augmented Complex LMS algorithm is employed as the learning rule; whereas in our algorithm we have used kalman filter algorithm for adaptation and learning.…”