Timeline for How does sampling jitter affect state estimation?
Current License: CC BY-SA 4.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 10, 2022 at 20:41 | comment | added | TimWescott | I think you'd find that you'd just need to change the state evolution matrix ($\mathbf A$) and possibly the input matrix, but not the output matrix. And, depending on the eigenvalues of $\mathbf A_c$, you may be able to use $\mathbf A \simeq I + \Delta t \mathbf A$ instead of the matrix exponential. | |
| Jan 10, 2022 at 19:25 | comment | added | fpf3 | @TimWescott Maybe I could take Δt as a parameter to my model. If I model it as e^(A_c Δt) instead of just A, perhaps it just "goes away". I think I would have to also modify my input and output matrices... hmm... probably too convoluted | |
| Jan 10, 2022 at 19:06 | comment | added | fpf3 | I'm curious about both cases, but the real-life problem that led me to post this is one where I can measure the current Δt_k, yes. | |
| Jan 10, 2022 at 19:01 | comment | added | TimWescott | I just realized something -- is the sampling error unknown, or is your sampling interval randomly varying but known at the point of sampling? If the latter, then you could do much better than by modeling it as an unknown variation. | |
| Jan 10, 2022 at 18:42 | vote | accept | fpf3 | ||
| Jan 10, 2022 at 18:36 | comment | added | fpf3 | thanks, marcus :) | |
| Jan 10, 2022 at 9:19 | comment | added | Marcus Müller | I do like your user logo :) | |
| Jan 9, 2022 at 3:20 | history | edited | fpf3 | CC BY-SA 4.0 |
added 11 characters in body
|
| Jan 9, 2022 at 1:36 | answer | added | TimWescott | timeline score: 2 | |
| S Jan 9, 2022 at 0:22 | review | First questions | |||
| Jan 9, 2022 at 0:42 | |||||
| S Jan 9, 2022 at 0:22 | history | asked | fpf3 | CC BY-SA 4.0 |