0
$\begingroup$

I have a 1D Time-of-Flight based range finder that returns distance in mm. I am trying to implement a Kalman filter to get outlier-free estimation. The sensor measures the distance to the ground below it while the sensor assembly is mounted on a mobile platform that moves orthogonal to the direction of measurement. The following diagram tries its best to clarify: enter image description here

I am unable to come up with a motion model for the filter as the measurements are independent of the motion of the assembly. Is Kalman Filter still applicable here ?

Improve this question
$\endgroup$

1 Answer 1

Reset to default
0
$\begingroup$

How are you expecting the surface to appear? Sinusoidal, with random amplitudes? Flat, with random deviations? The technical example on Wikipedia has an example that implements a generic motion model; this would be what I'd start with if I had no knowledge of what the surface was supposed to be.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Hi Chuck, the wiki example tries to use the Newton's motion equations (using v and acceleration) for estimating the state (x) while I am trying to estimate the state/range of the range-finder which is independent of the assembly/buggy's motion. $\endgroup$
    – Pe Dro
    Aug 5, 2022 at 1:14
  • $\begingroup$ @PeDro Do you have a model of what your expect the surface to do? You can assume the same model if you're assuming there will be smooth curves on the surface - the Wikipedia example is not actuating the cart, it's just applying the motion model to an object "buffeted by random forces." This would result in random accelerations, which would draw random curves if you plotted position. In the absence of any better model for surface position, this is what I would start with. $\endgroup$
    – Chuck
    Aug 5, 2022 at 11:39

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.