Filtres adaptatifs

Filtres LMS, RLS, de domaine fréquentiel, de projection affine et adaptatifs en treillis

DSP System Toolbox™ propose plusieurs variantes des algorithmes de filtres adaptatifs à réponse impulsionnelle finie (FIR) LMS ET RLS. Bien qu’ils diffèrent par leurs détails, ces algorithmes utilisent tous la même approche opérationnelle consistant à minimiser la différence d’erreur entre la sortie du filtre adaptatif et le signal souhaité. La métrique la plus souvent utilisée pour quantifier cette erreur est l’erreur quadratique moyenne (MSE). Les filtres adaptatifs sont largement utilisés dans diverses applications comme la suppression du bruit acoustique ou des échos, le beamforming, l’identification de systèmes, l’amélioration des signaux biomédicaux, l’égalisation des canaux de communication, etc. Pour des exemples illustrant certaines de ces applications, veuillez consulter System Identification of FIR Filter Using LMS Algorithm, Noise Cancellation Using Sign-Data LMS Algorithm et Inverse System Identification Using RLS Algorithm.

Lorsque les entrées sont colorées, les algorithmes de filtres adaptatifs de projection affine disponibles avec l’objet dsp.AffineProjectionFilter accélèrent considérablement la convergence par rapport aux variantes LMS. Pour un coût de calcul plus élevé, l’algorithme de filtre adaptatif en treillis disponible avec l’objet dsp.AdaptiveLatticeFilter peut offrir une meilleure convergence que ses homologues LMS et RLS. Il est également possible d’implémenter un filtre FIR adaptatif dans le domaine fréquentiel avec l’objet dsp.FrequencyDomainAdaptiveFilter.

La performance de la convergence dépend de la trajectoire de la MSE déterminée par msesim ainsi que de la façon dont elle converge vers la MSE prédite, déterminée par msepred.

Objets

`dsp.BlockLMSFilter`	Compute output, error, and weights using block least mean squares (LMS) adaptive algorithm
`dsp.LMSFilter`	Compute output, error, and weights of least mean squares (LMS) adaptive filter
`dsp.RLSFilter`	Compute output, error and coefficients using recursive least squares (RLS) algorithm
`dsp.AffineProjectionFilter`	Compute output, error and coefficients using affine projection (AP) Algorithm
`dsp.AdaptiveLatticeFilter`	Adaptive lattice filter
`dsp.FastTransversalFilter`	Fast transversal least-squares FIR adaptive filter
`dsp.FilteredXLMSFilter`	Filtered XLMS filter
`dsp.FrequencyDomainAdaptiveFilter`	Compute output, error, and coefficients using frequency-domain FIR adaptive filter

Blocs

Block LMS Filter	Compute output, error, and weights using LMS adaptive algorithm
Fast Block LMS Filter	Compute output, error, and weights using least mean squares (LMS) adaptive algorithm
Frequency-Domain Adaptive Filter	Compute output, error, and coefficients using frequency domain FIR adaptive filter
Kalman Filter	Predict or estimate states of dynamic systems
LMS Filter	Compute output, error, and weights using least mean squares (LMS) adaptive algorithm
LMS Update	Estimate weights of least mean squares (LMS) adaptive filter
RLS Filter	Compute filtered output, filter error, and filter weights for given input and desired signal using RLS adaptive filter algorithm

Rubriques

Overview of Adaptive Filters and Applications
General discussion on how adaptive filters work, list of adaptive filter algorithms in DSP System Toolbox, convergence performance, and details on few common applications.
System Identification of FIR Filter Using LMS Algorithm
Identify an unknown system using LMS algorithm.
System Identification of FIR Filter Using Normalized LMS Algorithm
Identify an unknown system using normalized LMS algorithm.
Compare Convergence Performance Between LMS Algorithm and Normalized LMS Algorithm
Compare the speed with which the adaptive filter algorithms converge.
Signal Enhancement Using LMS and NLMS Algorithms
Introduces adaptive filters through a signal enhancement application.
Noise Cancellation Using Sign-Data LMS Algorithm
Perform noise cancellation using sign-data LMS algorithm.
System Identification Using RLS Adaptive Filtering
This example shows how to use a recursive least-squares (RLS) filter to identify an unknown system modeled with a lowpass FIR filter.
Inverse System Identification Using RLS Algorithm
Perform inverse system identification using dsp.RLSFilter.
Compare RLS and LMS Adaptive Filter Algorithms
Comparison of RLS and LMS adaptive filter algorithms.
Adapt Multiple Filters Using LMS Update Block
Adapt multiple filters independently using the same LMS Update block. Use the Adapt port of the LMS Update block to selectively enable or disable the filters from being adapted.
Model Adaptive Linear Combiner using LMS Update Block
Use LMS Update block as an adaptive linear combiner.
Éliminer le bruit à basse fréquence dans Simulink en utilisant un filtre adaptatif normalisé LMS
Concevez un filtre adaptatif normalisé LMS et utilisez-le pour éliminer le bruit à basse fréquence dans Simulink^®.
Noise Cancellation in Simulink Using Normalized LMS Adaptive Filter
Remove colored noise generated from an acoustic environment, using a normalized LMS adaptive filter.
Variable-Size Signal Support DSP System Objects
List of System objects that support variable-sized signals in DSP System Toolbox.

Informations connexes

Contrôle actif du bruit : de la modélisation au prototypage temps réel

Exemples présentés

Adaptive Line Enhancer (ALE)

Apply adaptive filters to signal separation using a structure called an adaptive line enhancer (ALE). In adaptive line enhancement, a measured signal x(n) contains two signals, an unknown signal of interest v(n), and a nearly-periodic noise signal eta(n).

Ouvrir le live script

Adaptive Noise Cancellation Using RLS Adaptive Filtering

Use an RLS filter to extract useful information from a noisy signal. The information bearing signal is a sine wave that is corrupted by additive white Gaussian noise.

Ouvrir le live script

Acoustic Noise Cancellation (LMS)

Use the least mean square (LMS) algorithm to subtract noise from an input signal. The example uses a preconfigured Simulink® model. The LMS Filter block in the dspanc model models an adaptive filter that uses the reference signal at its Input port and the desired signal at the Desired port to automatically match the filter response. The LMS Filter block subtracts the filtered noise from the original signal. As the filter converges, the resultant error signal contains only the original signal.

Ouvrir le modèle

Active Noise Control with Simulink Real-Time

Design a real-time active noise control system using a Speedgoat® Simulink® Real-Time™ target.

Adaptive Filter Convergence

Compares the rate of convergence for adaptive filters using different LMS algorithms.

Ouvrir le script

Noise Canceler (RLS)

Subtract noise from an input signal using the Recursive Least Squares (RLS) algorithm. The RLS adaptive filter uses the reference signal on the Input port and the desired signal on the Desired port to automatically match the filter response in the Noise Filter block. As it converges to the correct filter, the filtered noise should be completely subtracted from the "Signal+Noise" signal, and the "Error Signal" should contain only the original signal.

Ouvrir le modèle

Time-Delay Channel Estimation Through Adaptive Filtering

Adaptively estimate the time delay for a noisy input signal using the LMS adaptive FIR algorithm.

Ouvrir le modèle