I have sensors placed in pipelines and I have amplitudes coming from it. I want to detect the type of activity taking place at the sensors(like digging,tunneling,clamping etc). So I have a signal which consists of 30 amplitudes, what features can i extract.

What are the assumptions that I have to make(like the signal should be stationary etc). Is my signal stationary in this case.If it is not can I apply frequency domain features.What are the differnt features I can use apart from mean,median,mode,skewness,kurtosis etc.I am trying t build a nueral network classifier.Is not my signal continous if it comes each second(amplitudes).Can I apply fourier tranform here.

Sorry I am a complete newbie in the field

  • 1
    $\begingroup$ Hm, sorry to say it so blatantly but chances anyone can help you with these questions are close to zero. Your signal may or may not be stationary (how should we know? we haven't seen your signal?). You can apply whatever transform you like and extract whatever features you like. Are they useful? We cannot know. We haven't seen your signals nor do we know what exactly you're looking for. Since you already have the sensors, why not take some measurements, try to induce known activities and see how they behave? If you have some more concrete examples to show, people might be able to help you. $\endgroup$
    – Florian
    Commented Jul 24, 2019 at 10:43
  • $\begingroup$ Oh, alright! I dont even know , what info will be helpful to you.I will come back with more info when I get $\endgroup$
    – Fasty
    Commented Jul 24, 2019 at 10:48


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.