2
$\begingroup$

I am new to Kalman filter. Is it possible to use a Kalman filter with just the position information of the target?

I am so confused with states and measurement.

$\endgroup$
2
  • $\begingroup$ so the position information refers to measurement right ? $\endgroup$
    – Fat32
    Commented Sep 28, 2017 at 13:53
  • $\begingroup$ yes....I have only the position of the targets $\endgroup$ Commented Sep 28, 2017 at 14:32

1 Answer 1

2
$\begingroup$

from the generality of your question then yes you can design a Kalman filter which would accept the target position as the only measurement possibly corrupted with noise.

Then the Kalman filter will try to estimate the true position of the target based on the noisy position measurement and the assumed motion model. Hence in effect the Kalman filter would be reducing the measurement noise of the target position measuring device.

The states of the filter would be defined based on the assumed motion model: a constant acceleration model would use for example the position, velocity and acceleration as the internal states which are being estimated (when the acceleration is unknown) or another model would use a deterministic input for the known acceleration and use two states instead.

$\endgroup$

Your Answer

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

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