In the areas of acoustic research or applications that deal with not-precisely-known or variable conditions, a method of adaptation to the uncertainness or changes is usually necessary. When searching for an adaptation algorithm, it is hard to overlook the least mean squares (LMS) algorithm. Its simplicity, speed of computation, and robustness has won it a wide area of applications: from telecommunication, through acoustics and vibration, to seismology. The algorithm, however, still lacks a full theoretical analysis. This is probabely the cause of its main drawback: the need of a careful choice of the step size - which is the reason why so many variable step size flavors of the LMS algorithm has been developed.
This paper contributes to both the above mentioned characteristics of the LMS algorithm. First, it shows a derivation of a new necessary condition for the LMS algorithm convergence. The condition, although weak, proved useful in developing a new variable step size LMS algorithm which appeared to be quite different from the algorithms known from the literature. Moreover, the algorithm proved to be effective in both simulations and laboratory experiments, covering two possible applications: adaptive line enhancement and active noise control.
Speech enhancement objective is to improve the noisy speech signals for human perception. The intention of speech enhancement algorithm is to improve the performance of the communication, when the signal is occluded by noise. The quality and intelligibility of speech is reduced because of the presence of background noise. There are various adaptive filtering algorithms for speech enhancement. The existing least mean square and normalised least mean square algorithms have the problem of choosing the step size that guarantees the stability of the algorithm. To overcome this problem, we focus on speech enhancement by amended adaptive filtering. The proposed algorithm follows blind source separation strategy using adaptive filtering. Comparison of existing adaptive filtering algorithms with proposed algorithm justifies the amendment incorporated in this paper. Taking the objective criteria into account the algorithms has been tested for segmental signal to noise ratio (SegSNR), segmental mean square error (SegMSE), signal to noise ratio and mean square error. The proposed algorithm can be used for hand-free cell phone, hearing aids and teleconferencing systems.
This paper presents a new simple and accurate frequency estimator of a sinusoidal signal based on the signal autocorrelation function (ACF). Such an estimator was termed as the reformed covariance for half-length autocorrelation (RC-HLA). The designed estimator was compared with frequency estimators well-known from the literature, such as the modified covariance for half-length autocorrelation (MC-HLA), reformed Pisarenko harmonic decomposition for half-length autocorrelation(RPHD-HLA), modified Pisarenko harmonic decomposition for half-length autocorrelation (MPHD-HLA), zero-crossing (ZC), and iterative interpolated DFT (IpDFT-IR) estimators. We determined the samples of the ACF of a sinusoidal signal disturbed by Gaussian noise (simulations studies) and the samples of the ACF of a sinusoidal voltage(experimental studies), calculated estimators based on the obtained samples, and computed the mean squared error(MSE) to compare the estimators. The errorswere juxtaposed with the Cramér–Rao lower bound (CRLB). The research results have shown that the proposed estimator is one of the most accurate, especially for SNR > 25 dB. Then the RC-HLA estimator errors are comparable to the MPHD-HLA estimator errors. However, the biggest advantage of the developed estimator is the ability to quickly and accurately determine the frequency based on samples collected from no more than five signal periods. In this case, the RC-HLA estimator is the most accurate of the estimators tested.
The zero attraction affine projection algorithm (ZA-APA) achieves better performance in terms of convergence rate and steady state error than standard APA when the system is sparse. It uses l1 norm penalty to exploit sparsity of the channel. The performance of ZA-APA depends on the value of zero attractor controller. Moreover a fixed attractor controller is not suitable for varying sparsity environment. This paper proposes an optimal adaptive zero attractor controller based on Mean Square Deviation (MSD) error to work in variable sparsity environment. Experiments were conducted to prove the suitability of the proposed algorithm for identification of unknown variable sparse system.