Non-negative least-mean-square algorithm

Jie Chen (1, 2) , Cédric Richard (1) , José C. M. Bermudez (3) , Paul Honeine (2)

1 FIZEAU - Laboratoire Hippolyte Fizeau
2 LM2S - Laboratoire Modélisation et Sûreté des Systèmes
3 UFSC - Universidade Federal de Santa Catarina = Federal University of Santa Catarina [Florianópolis]

Dynamic system modeling plays a crucial role in the development of techniques for stationary and nonstationary signal processing. Due to the inherent physical characteristics of systems under investigation, nonnegativity is a desired constraint that can usually be imposed on the parameters to estimate. In this paper, we propose a general method for system identification under nonnegativity constraints. We derive the so-called nonnegative least-mean-square algorithm (NNLMS) based on stochastic gradient descent, and we analyze its convergence. Experiments are conducted to illustrate the performance of this approach and consistency with the analysis.

Mots clés: least mean squares methods; gradient methods; non-negativity; nonnegative constraints; transient analysis; adaptive filtering; parameter estimation; dynamic system modeling; signal processing; nonstationary signal processing; stochastic gradient descent; nonnegativity constraint; Equations; Prediction algorithms; Convergence; Algorithm design and analysis; least mean square algorithms; Facsimile; nonnegative least-mean-square algorithm; Adaptive filters; Mathematical model; adaptive signal processing; system identification; Least squares approximation

Domaines: Apprentissage [cs.LG]; Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]

Jie Chen, Cédric Richard, José C. M. Bermudez, Paul Honeine. Non-negative least-mean-square algorithm. IEEE Transactions on Signal Processing, 2011, 59 (11), pp.5225 - 5235. ⟨10.1109/TSP.2011.2162508⟩. ⟨hal-01965584⟩