The Least mean squares filter solution converges to the Wiener filter solution, assuming that the unknown system is LTI and the noise is stationary. Both filters can be used to identify the impulse response of an unknown system, knowing only the original input signal and the output of the unknown system.By relaxing the error criterion to reduce current sample error instead of minimizing the total error over all of n, the LMS algorithm can be derived from the Wiener filter.
Contents
Derivation of the Wiener filter for system identification
Given a known input signal
where
The model system
where
The error between the model and the unknown system can be expressed as:
The total squared error
Use the Minimum mean-square error criterion over all of
Substitute the definition of
Distribute the partial derivative:
Using the definition of discrete cross-correlation:
Rearrange the terms:
This system of N equations with N unknowns can be determined.
Derivation of the LMS algorithm
By relaxing the infinite sum of the Wiener filter to just the error at time
The squared error can be expressed as:
Using the Minimum mean-square error criterion, take the gradient:
Apply chain rule and substitute definition of y[n]
Using gradient descent and a step size
which becomes, for i = 0, 1, ..., N-1,
This is the LMS update equation.