Linear System Identifiability from Distributional and Time Series Data
- Creators
-
Swaminathan, Anandh
-
Murray, Richard M.
Abstract
We consider identifiability of linear systems driven by white noise using a combination of distributional and time series measurements. Specifically, we assume that the system has no control inputs available and can only be observed at stationarity. The user is able to measure the full stationary state distribution as well as observe time correlations for small subsets of the state. We formulate theoretical conditions on identifiability of parameters from distributional information alone. We then give a sufficient condition and an effective necessary condition for identifiability using a combination of distributional and time series measurements. We illustrate the ideas with some simple examples as well as a biologically inspired example of a transcription and degradation process.
Additional Information
© 2016 AACC. This work was supported by AFOSR grant FA9550-14-1-0060. A. Swaminathan was also supported by the NSF GRFP. We acknowledge Professor Eduardo Sontag for helpful advice. We especially acknowledge Dr. Yutaka Hori for helpful advice as well as detailed comments on the manuscript. We also thank Reviewer 3 for very detailed feedback.Attached Files
Submitted - sm16-acc.pdf
Files
| Name | Size | Download all |
|---|---|---|
|
md5:31e4f84f85a725dc4f91ffa63e9f6a46
|
377.3 kB | Preview Download |
Additional details
- Eprint ID
- 69381
- Resolver ID
- CaltechAUTHORS:20160802-094940806
- Air Force Office of Scientific Research (AFOSR)
- FA9550-14-1-0060
- NSF Graduate Research Fellowship
- Created
-
2016-08-02Created from EPrint's datestamp field
- Updated
-
2021-11-11Created from EPrint's last_modified field