LMS Adaptive Filter (Obsolete)
Compute filter estimates for input using LMS adaptive filter algorithm
Library
dspobslib
Description
Note
The LMS Adaptive Filter block is still supported but is likely to be obsoleted in a future release. We strongly recommend replacing this block with the LMS Filter block.
The LMS Adaptive Filter block implements an adaptive FIR filter using the stochastic gradient algorithm known as the normalized least mean-square (LMS) algorithm.
$$\begin{array}{c}y(n)={\widehat{w}}^{H}\left(n-1\right)u(n)\\ e(n)=d(n)-y(n)\\ \widehat{w}\left(n\right)=\widehat{w}\left(n-1\right)+\frac{u(n)}{a+{u}^{H}(n)u(n)}\mu {e}^{*}(n)\end{array}$$
The variables are as follows.
Variable | Description |
---|---|
n | The current algorithm iteration |
u(n) | The buffered input samples at step n |
$$\widehat{w}(n)$$ | The vector of filter-tap estimates at step n |
y(n) | The filtered output at step n |
e(n) | The estimation error at step n |
d(n) | The desired response at step n |
µ | The adaptation step size |
To overcome potential numerical instability in the tap-weight update, a small positive constant (a = 1e-10) has been added in the denominator.
To turn off normalization, clear the Use normalization check box in the parameter dialog. The block then computes the filter-tap estimate as
$$\widehat{w}(n)=\widehat{w}(n-1)+u(n)\mu {e}^{*}(n)$$
The block icon has port labels corresponding to the inputs and outputs of the LMS algorithm. Note that inputs to the In and Err ports must be sample-based scalars. The signal at the Out port is a scalar, while the signal at the Taps port is a sample-based vector.
Block Ports | Corresponding Variables |
---|---|
| u, the scalar input, which is internally buffered into the vector u(n) |
| y(n), the filtered scalar output |
| e(n), the scalar estimation error |
| $$\widehat{w}(n)$$, the vector of filter-tap estimates |
An optional Adapt
input port is added when you select the
Adapt input check box in the dialog. When this port is enabled,
the block continuously adapts the filter coefficients while the Adapt
input is nonzero. A zero-valued input to the Adapt port causes the block to stop
adapting, and to hold the filter coefficients at their current values until the next
nonzero Adapt
input.
The FIR filter length parameter specifies the length of the filter that the LMS algorithm estimates. The Step size parameter corresponds to µ in the equations. Typically, for convergence in the mean square, µ must be greater than 0 and less than 2. The Initial value of filter taps specifies the initial value $$\widehat{w}\left(0\right)$$ as a vector, or as a scalar to be repeated for all vector elements. The Leakage factor specifies the value of the leakage factor, 1 –μ α, in the leaky LMS algorithm below. This parameter must be between 0 and 1.
$$\widehat{w}(n+1)=(1-\mu \alpha )\widehat{w}(n)+\frac{u(n)}{{u}^{H}(n)u(n)}\mu {e}^{*}(n)$$
Examples
See the lmsadtde
demo.
Parameters
- FIR filter length
The length of the FIR filter.
- Step-size
The step-size, usually in the range (0, 2). Tunable (Simulink).
- Initial value of filter taps
The initial FIR filter coefficients.
- Leakage factor
The leakage factor, in the range [0, 1]. Tunable (Simulink).
- Use normalization
Select this check box to compute the filter-tap estimate using the normalized equations.
- Adapt input
Enables the Adapt port when selected.
References
Haykin, S. Adaptive Filter Theory. 3rd ed. Englewood Cliffs, NJ: Prentice Hall, 1996.
Supported Data Types
Double-precision floating point
Single-precision floating point
See Also
Kalman Adaptive Filter (Obsolete) | DSP System Toolbox |
RLS Adaptive Filter (Obsolete) | DSP System Toolbox |
See Noise Cancellation in Simulink Using Normalized LMS Adaptive Filter for related information.