From Wikipedia:

If the output signal slowly changes independent of the measured property, this is defined as drift

So my question is, if I have a gyroscope with sensor drift does that mean that $E\left\{f(x)\right\} \neq 0$ ? I have an excercise where it is written, that the sensor has drift but also that the $E\left\{f(x)\right\} = 0$. But to me one statement seems excludatory of the other one.


I assume you set the first momentum of your true measure to $\mu=0$, so your question actually reads

gyroscope with sensor drift does that mean that $E[f(x)] \ne \mu$?

No. It might still be the case that the sensor value, in the long term, drifts centred around the true value.

The thing is that $f(x)$ is a random variable with non-zero variance and a time dependency. That means: If there's drift, that random variable is not a stationary process.

  • $\begingroup$ Ah I see. So would an example of this be a thermometer that measures the temperature over a whole year. It has E[f(x)] = some_mean But it would have a drift in Summer and Winter respectively? Maybe my confusion stems from the contrieved example of a gyro having a drift. Or is that a realisitc case? $\endgroup$ – Hakaishin May 29 '17 at 12:13
  • $\begingroup$ Your thermometer example doesn't work, because the actual temperature does change between summer and winter. $\endgroup$ – Marcus Müller May 29 '17 at 12:37

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.