Extension of LMS stability condition over a wide set of signals

Abstract

SummaryA sufficient condition for least mean squares (LMS) algorithm stability with a small set of assumptions is derived in this paper. The derivation is not, contrary to the majority of currently known conditions, based on the independence assumption or other statistic properties of the input signals. Moreover, it does not make use of the small‐step‐size assumption, neither does it assume the input signals are stationary. Instead, it uses a theory of discrete systems and properties of a discrete state‐space matrix. Therefore, the result can be applied to a wide set of signals, including deterministic and nonstationary signals. The location of all eigenvalues of the matrix responsible for the LMS algorithm stability has been calculated. Simulation experiments, where the step size reaches a couple of hundreds without loss of stability, are shown to support the theory. On the other hand, simulation where the calculations based on the small‐step‐size theory provide a too large estimation of the upper bound for the step size, while the new condition gives a proper solution, is also presented. Therefore, the new condition may be used in cases where fast adaptation is necessary and when the independence theory or the small‐step‐size assumptions do not hold. Copyright © 2014 John Wiley & Sons, Ltd.

Research into dynamics of setting the threshold and a probability of ignition detection by self­adjusting fire detectors

“…In the latter case, SFD under consideration may appear incapable of providing guaranteed ignition detection. In this regard, various modifications of algorithms, in which a step size changes in the process of adaptation, continue to evolve and remain relevant [6][7][8][9][10][11].…”

“…However, such a number of adjustable parameters significantly limits the scope of its practical use, including for SFD. Certain simplification of the algorithm is possible by selecting the upper boundary of a step size, based on the conditions of algorithm stability [11].…”

Stability Conditions for the Leaky LMS Algorithm Based on Control Theory Analysis

“…A sufficient stability condition for the LMS algorithm, derived based on control system theory, without using either the independence, or small-step-size assumptions, has been recently presented in (Bismor, 2015). The same approach will be applied here to derive a sufficient stability condition for the LLMS algorithm.…”

“…This representation, which comes as a direct result of the optimization problem resulting in the LLMS algorithm, allows one to understand that the leakage and the step size are related, i.e., when one is increased the other should be decreased to keep the same cost function value. As in (Bismor, 2015), we will assume the input and the desired signals are real-valued and bounded, and that the adaptive filter is a linear (for frozen time), transversal filter of finite impulse response (FIR).…”

Nonlinear State–Space Predictive Control with On–Line Linearisation and State Estimation

“…In consequence, the MPC optimisation problems (5), (6) or (18) are nonlinear tasks which must be solved on-line in real time. Although in some applications, e.g., in active noise control (Bismor, 2015), on-line nonlinear optimisation is used, in general it may be time-consuming and very difficult from a computational point of view (e.g., non-convex or multi-modal problems). That is why two computationally efficient alternatives with on-line linearisation are considered, which result in quadratic optimisation problems.…”