0
$\begingroup$

I have use case, where there are lot of sensors placed under ground.The sensors give amplitudes each second.So my aim is to detect if some anomalous activity takes place at that point.

IF an activity such as digging takes place, there will be variation in amplitudes, so can that data be fed to a machine learning model and find the anomalous activity taking place?

I have signatures of activities like drilling,digging,augering etc.But my question is amplitude a unique feature here or do I need to consider things like frequency?

$\endgroup$
1
$\begingroup$

Taking FFT and studying the frequency of each activity would be helpful. Each activity should have its own unique frequency. You can use combination of both if any two activities have the same frequency or amplitude.

| improve this answer | |
$\endgroup$
0
$\begingroup$

The different phenomena that you describe involve time and eventually periodicity. In this sense it is better to consider also frequency domain features combined with intensity.

| 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.