0
$\begingroup$

I've got a following signal: Radar signal

It's a discrete signal from two radar devices from one day. Signal is sampled every 333 ms. So at axis x there is a timeline, axis y is a current radar value. We use radar for object detection. The peaks are indicating that there is an object in front of radar and the rest is noise or a background. As you can see blue one is a better then red one (peaks are higher and background is lower).

Problem is that we need to find technique how to measure quality of that signal (is there many peaks? noise is low? or signal is all noise and we can not identify signal?). I'm novice in this area so any help will be welcome. We've started with standard deviaton of signal but i'm not sure if it's a right way.

Thank you.

New contributor
jakubM is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
$\endgroup$
  • $\begingroup$ no, I can't see that! How do you decide what's "better"? We don't even know what the axes of your plot are! We can't tell you how to assess the quality of something if we don't know what it's used for (e.g. Camembert makes for terrible building material, but excellent dinner; also: an estimator with high resolution but bad variance might be good for one, but bad for a different problem) $\endgroup$ – Marcus Müller Jan 13 at 14:08
  • $\begingroup$ You need to define "quality". Is it signal to noise ratio? Is quality a continuous thing or binary like good or bad? Edit this stuff into your question to help people understand $\endgroup$ – Engineer Jan 13 at 14:36
  • $\begingroup$ thanks for reply, i've added some more information. $\endgroup$ – jakubM Jan 13 at 15:14
  • $\begingroup$ What's a "radar value"? What are the units of either axis? We still can't tell you which curve is closer to reality, and we can't tell you about which aspects your system cares about, so I'm afraid we can't help you. Quality is, as said, always relative to purpose! $\endgroup$ – Marcus Müller Jan 13 at 16:23

Browse other questions tagged or ask your own question.