Signal averaging

Signal averaging (engl. Signal averaging ) is a method by which the noise can be reduced in the repetitive signals. The useful signal is amplified by means of additive overlay, while the overlay of random noise does not lead to any amplification. This improves the overall signal-to-noise ratio . Requirements for signal averaging to be used are:

It is a repeated signal, the time- synchronization can be
Signal and noise are uncorrelated
The noise is random, so its arithmetic mean in the frequency range under consideration is zero ( white noise )

Examples

Biosignals

Electroenzophalograms (EEG) are often overlaid by various processes in the brain. So-called evoked potentials can be derived through targeted stimuli (e.g. optical or acoustic) and synchronized recording of the EEG . The number of stimulus responses necessary for a good result in practice depends on the signal-to-noise ratio and is different depending on the sensory modality and physical characteristics.

Several biosignals can be transmitted from a subject in a psychophysiological experiment, including e.g. B. the skin conductance or the pupil diameter can be measured in order to measure the state of arousal or the cognitive processing effort related to a certain stimulus. This stimulus is repeatedly presented to the test person. This results in signal curves that should be the same apart from some artifacts and errors in the measuring system caused by external influences. If these influences are not systematically caused and can therefore be perceived as noise, then signal averaging can be used sensibly.

Traffic monitoring

In traffic monitoring, signal averaging can be used to create a current background image if the camera remains in a fixed position. This background image automatically adapts to the changing light conditions, if z. B. the last 32 image signals are regularly summed and averaged, it can then be helpful for operations for object recognition . However, the disruptive component (the traffic) is not random and an interfering object can appear several times in a sequence of images, for example if it stops.

Individual evidence

↑ Noise and Lock-In Gain (accessed July 26, 2018)
↑ SIGNAL / NOISE IMPROVEMENT (accessed on July 26, 2018)
^ Rüdiger Kramme (ed.): Medical technology. Procedures - systems - information processing, 3rd edition, Springer Verlag, Berlin / Heidelberg, ISBN 978-3-540-34102-4 , pp. 220-223.
↑ G. Steinbeck (Ed.): Life-threatening ventricular cardiac arrhythmias. Steinkopff Verlag, Darmstadt 1987, ISBN 978-3-642-72406-0 , pp. 113-120.

[1] Noise and Lock-In Gain (accessed July 26, 2018)

[2] SIGNAL / NOISE IMPROVEMENT (accessed on July 26, 2018)

[3] Rüdiger Kramme (ed.): Medical technology. Procedures - systems - information processing, 3rd edition, Springer Verlag, Berlin / Heidelberg, ISBN 978-3-540-34102-4 , pp. 220-223.

[4] G. Steinbeck (Ed.): Life-threatening ventricular cardiac arrhythmias. Steinkopff Verlag, Darmstadt 1987, ISBN 978-3-642-72406-0 , pp. 113-120.