Memory improved proportionate affine projection sign algorithm

Abstract

A new proportionate affine projection sign algorithm is proposed for network echo cancellation. It uses a recursive procedure and takes into account the previously computed proportionate coefficients. It is shown that the proposed algorithm can obtain a lower steady-state misalignment than other affine projection sign algorithms for different echo paths, impulsive interferences and step sizes.Introduction: The family of affine projection algorithms (APAs) is known to provide a better convergence performance in comparison with the known normalised least mean square (NLMS) algorithm [1]. Their performance can be further improved by taking into account the sparsity of the echo paths [2] (e.g. in case of network echo cancellers (NEC)). A number of proportionate-type adaptive algorithms were proposed (e.g. improved proportionate APA (IPAPA) [2], and memory IPAPA (MIPAPA) [3] etc) Unfortunately, the above-mentioned algorithms are not robust to impulsive noise [4,5], and a user may experience a deteriorated performance during a telephone conversation in an impulse interference environment. It is known that the family of affine projection sign algorithms is robust against impulsive noise and they have lower complexity than the conventional affine projection algorithms [4]. One such algorithm was proposed in [4] and it was termed the affine projection sign algorithm (APSA). The real-coefficient improved proportionate affine projection sign algorithm (RIP-APSA) was proposed in [5]. In [5], the RIP-APSA's superiority over APA, APSA, and proportionate APA's in terms of robustness and fast convergence for both sparse and dispersive echo paths in impulsive environments was proved by extensive simulation results. However, its steady-state performance is inferior to that of the APSA especially for dispersive echo paths [5]. In this Letter, we propose a new algorithm called the memory improved proportionate affine projection sign algorithm (MIP-APSA), with similar complexity and performance but greatly improved steady-state misalignment as compared to the RIP-APSA in impulsive environments.

Proportionate Adaptive Filtering for Block-Sparse System Identification

“…Simulation results demonstrated that, the new BS-PNLMS algorithm outperforms the NLMS, PNLMS and IPNLMS algorithms for the block-sparse system, and the new BS-IPNLMS algorithm is more preferred for the dispersive system. This block-sparse proportionate idea proposed in this paper could be further extended to many other proportionate algorithms, including proportionate affine projection algorithm (PAPA) [23], proportionate affine projection sign algorithm (PAPSA) [24], and their corresponding low complexity implementations [25]- [26] etc. The proof of convergence for the proposed BS-PNLMS and BS-IPNLMS algorithms can also be part of the future work.…”

Improved proportionate-type sparse adaptive filtering under maximum correntropy criterion in impulsive noise environments

“…MIP−APSA [9] Normalized LHTAF with K=200 [10] Normalized LHTAF with K=50 [10] Normalized LVHTAF with K min =50 [10] IP−MCC [Proposed] Fig. 3.…”

Memory proportionate APSA with individual activation factors for highly sparse system identification in impulsive noise environment

“…Clearly, different strategies to calculate G(k) can lead to different proportionate-type APSAs, as shown in [14]- [15], and simulation results has shown that these algorithms converge faster than the APSA for sparse impulse responses. Particularly, the APSA is obtained when G(k) is equal to the identity matrix, i.e., all the filter weight coefficients are assigned the same step-size.…”

“…In [14], the proportionate ideas of the PAPA and IPAPA (or the PNLMS and IPNLMS) are directly extended to the APSA, respectively, and the resulting proportionate APSA (RP-APSA) and improved RP-APSA (RIP-APSA) achieve faster convergence rate than the APSA in sparse system. Afterwards, a memory RIP-APSA (MIP-APSA) [15] was proposed which obtains a lower steadystate misalignment than the RIP-APSA.…”