The session covered several example applications of the Kalman filter and parameter estimation via maximum likelihood. Participants presented their own problems and ideas. This was the final session of the workshop.
This session was the most hands-on of all. I have the impression the participants were all very motivated, and they enjoyed the workshop. I will request feedback soon.
The following examples were presented by the participants
-
Manuel Regueiro-Picallo from Eawag presented the problem of estimating temperature distributions between measurements at two points. He discretized the unsteady diffusion equation $$T_t = k T_{xx}$$ and built the Kalman filter with the resulting matrices.
This is something we had been exploring at Eawag a few years ago, and I think Manuel has very good chances of getting nice results.
-
Reto Christen from IET presented the tracking of rapid phase changes in signals form the electrical grid (very high noise!). He also discussed the implementation of the filter in embedded electronics. He used recursive regression with periodic functions of different frequencies.
His MATLAB/GNU Octave scripts can be found in the workshop repository, together with the slides he used to present the problem and solution.
-
Adrian Rohner and Florian Hammer from IET discussed the issue of smoothing speed wind velocity measurements from wind turbines, and the evaluation of uncertain speed-to-power curves.
Slides
To download a PDF file with the slides and notes click the image below!
