org.openuat.authentication.accelerometer
Class MarinPerianuComparison

java.lang.Object
  org.openuat.authentication.accelerometer.MarinPerianuComparison

public class MarinPerianuComparison
extends java.lang.Object

This class implements comparison of accelerometer time series based on their correlation coefficient. It uses the parameters suggested in R. Marin-Perianu, M. Marin-Perianu, P. Havinga, and H. Scholten, “Movement-based group awareness with wireless sensor networks,” in Proc. Pervasive 2007. Springer-Verlag, May 2007, pp. 298–315. Inputs, i.e. the time series to be compared, are acceleration magnitudes sampled at 8Hz.

Version:

1.0

Author:

Rene Mayrhofer

Field Summary

static int	bitsPerSample Samples are represented by just 1 Byte.
static float	correlationCoefficientThreshold For correlation coefficients > 0.5, the devices are assumed to be together, for correlation coefficients <= 0.5, they are assumed to be separate.
static int	sampleRate With a sample rate of 8Hz.
static int	stepSize But compute correlation coefficients every 16 samples (i.e. every 2s).
static int	windowSize Use the last 128 samples (i.e. the last 16s).

Constructor Summary

MarinPerianuComparison()

Method Summary

Methods inherited from class java.lang.Object

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

Field Detail

windowSize

public static final int windowSize

Use the last 128 samples (i.e. the last 16s).

See Also:

Constant Field Values

stepSize

public static final int stepSize

But compute correlation coefficients every 16 samples (i.e. every 2s).

See Also:

Constant Field Values

sampleRate

public static final int sampleRate

With a sample rate of 8Hz.

See Also:

Constant Field Values

bitsPerSample

public static final int bitsPerSample

Samples are represented by just 1 Byte.

See Also:

Constant Field Values

correlationCoefficientThreshold

public static final float correlationCoefficientThreshold

For correlation coefficients > 0.5, the devices are assumed to be together, for correlation coefficients <= 0.5, they are assumed to be separate.

See Also:

Constant Field Values

Constructor Detail

MarinPerianuComparison

public MarinPerianuComparison()