signalproc.algorithms

Class ALEProcessor

java.lang.Object

signalproc.algorithms.ALEProcessor

public class ALEProcessor
extends java.lang.Object

ALEProcessor performs adaptive noise cancelling on a series of chunks of data. It accounts for the endings and startings of the arrays of adjacent chunks and so treats the input chunks as one continuous input stream. This class is useful for processing continuous times series which maybe input as a series of sampled segments e.g. when receiving across a network. it saves the need to maintain buffers before processing can be applied.

This class is a computer model of :
 B. Widrow et al : Adaptive Noise Cancelling : Principles and Applications
 Proceedings of the IEEE, Vol 63, No. 12, December 1975.

Constructor Summary

Constructors

Constructor and Description
ALEProcessor(int numerOfTaps, int tapSpacing, java.lang.String stepSize, java.lang.String type, boolean useBIAS) Creates an ALE processor with a given number of taps (i.e.

Method Summary

Methods

Modifier and Type	Method and Description
void	allocate(double[] data) Allocates the arrays used if they are different sizes from currently set.
void	calcError()
double	calcMax(double[] data)
void	calcNewWeights()
void	calcOptStepSize()
void	calcStepSize()
void	calcYn()
void	normalise(double[] arr)
void	process(double[] data)
void	reset()
void	resetWeights()
void	setNumberOfTaps(int taps) Sets the number of taps (i.e.
void	setObject(Unit unit)
void	setOptimumStepSize(double[] data)
void	setStepSize(double stepSize) Sets the step size to given value
void	setStepSize(java.lang.String stepSize) Sets the step size to the double value contained within the string or to the reciprocal of the length of the input data if this string is set to auto on the next presentation of data.
void	setTapSpacing(int spacing) Sets the tap spacing (i.e.
void	setWeightUpdateType(java.lang.String type)
double	squareLength(double[] arr)
double	vectlength(double[] arr)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

ALEProcessor

public ALEProcessor(int numerOfTaps,
            int tapSpacing,
            java.lang.String stepSize,
            java.lang.String type,
            boolean useBIAS)

Creates an ALE processor with a given number of taps (i.e. weight vector size), tap spacing size, a step size (for the gradient descent), the type of algorithm used e.g. LMS or RLMS -> Least Mean Squared or Recursive Least Mean Squared and whether you want this ALE to use a BAIS input node.

Method Detail

setObject

public void setObject(Unit unit)

allocate

public void allocate(double[] data)

Allocates the arrays used if they are different sizes from currently set.

resetWeights

public void resetWeights()

process

public void process(double[] data)

calcYn

public void calcYn()

calcError

public void calcError()

calcNewWeights

public void calcNewWeights()

calcStepSize

public void calcStepSize()

calcOptStepSize

public void calcOptStepSize()

calcMax

public double calcMax(double[] data)

setOptimumStepSize

public void setOptimumStepSize(double[] data)

vectlength

public double vectlength(double[] arr)

squareLength

public double squareLength(double[] arr)

normalise

public void normalise(double[] arr)

reset

public void reset()

setTapSpacing

public void setTapSpacing(int spacing)

Sets the tap spacing (i.e. the increment of how to proceed along the input samples) to given value

setNumberOfTaps

public void setNumberOfTaps(int taps)

Sets the number of taps (i.e. the number of weights to use) to given value

setStepSize

public void setStepSize(double stepSize)

Sets the step size to given value

setStepSize

public void setStepSize(java.lang.String stepSize)

Sets the step size to the double value contained within the string or to the reciprocal of the length of the input data if this string is set to auto on the next presentation of data.

setWeightUpdateType

public void setWeightUpdateType(java.lang.String type)