0
$\begingroup$

I have 2 sensors, sensor A is sampling a signal at 1kHz, and sensor B is sampling the a signal at 100Hz.

Every timestamp that sensor B takes a measurement at is contained in sensor A. That is, sensor B takes a measurement essentially every 10th measurement taken by sensor A.

How should I downsample sensor A so that both data streams are 100Hz. Can I just pluck every 10th sample from sensor A, or should I low pass filter sensor A first? What frequency should the low pass filter be set at?

EDIT: It seems to me there should be no difference between taking every 10th sample or lowpass filtering if the signal I care about is less than 50Hz. Is that true?

$\endgroup$
1
$\begingroup$

If you can assure that the physical signal being measured by A has no detectable components above 50Hz, then you could take 1 every 10 samples.

However, if you have CPU power available, it is much better to filter the signal (lowpass, with stop-band starting below 50Hz) before downsampling. This way, components above 50Hz won't be aliased down to your working bandwidth. Also, this process will reduce quantization noise from the ADC, which is a good thing.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.