Updates of Matlab toolboxes based on latest Academic Research findings (Sys Id Toolbox)
1 view (last 30 days)
I was recently dealing with System Identification toolbox, and I ended up in a very interesting presentation of professor Ljung here
As you can see, he propose an improved strategy for finding the impulse response of a dynamical system. Furthermore, I have found this paper that exploit machine learning techniques for System Identification.
Those facts triggered me a question: how to verify if these techniques are already implemented in the latest System Identification toolbox release?
More in general: for each Toolbox release is it possible to see which techniques are implemented, and what future Toolbox releases will include?
Walter Roberson on 27 Nov 2015
Most of the documentation pages have a section at the end that gives reference to the algorithms used.
The Release Notes may describe new commands or changes to the approach, but sometimes the changes are not noted.
Mathworks does not publicly discuss what future releases will include; you would need to talk to Mathworks Sales to get information about that (you might need to sign a Non-Disclosure Agreement)
More Answers (1)
Arkadiy Turevskiy on 5 Jan 2016
Hi, As pointed out in the previous answer, each release we publish information about new features added to each of our products. For System identification Toolbox, the page with this information is located here. The page lists main new features for each of the last several releases and also includes links to more detailed release notes in the documentation.
As for regularization capabilities mention in prof. Ljung's talk, they were implemented in R2012b of the toolbox, you will find more detailed info on the Release Notes page , under R2012b (first feature bullet).
About future capabilities: as mentioned in the other answer, we do not disclose this information until the release actually ships. If you have specific suggestions or request, feel free to post them in MATLAB Answers, or contact me directly through my MATLAB Central profile.