Schedule (click here to subscribe to the webcal file)
| Date | Time | Room | Presenter(s) | Topic | PDF(s) |
| 16/02/2007 | 13:15 | E115 | Kristof | The early work, Kohonen Maps |
|
| 23/02/2007 | 13:00 | E302 | Tam, Ulf | Work on Low- and Highlevel Activities |
|
| 02/03/2007 | 13:00 | E302 | Maja | Recognition of ADLs, RFID + accelerometers |
|
| 16/03/2007 | X | X | MIS retreat |
MIS retreat |
|
| 23/03/2007 | 11:00 | E115 | André | Long-term, fine-grained actigraphy |
|
| 06/08/2007 | 16:00 | E115 | Maja, Tam | A Long-Term Evaluation of Sensing Modalities for Activity Recognition, and more |
|
| 24/08/2007 | 14:30 | E115 |
Andreas |
Activity recognition on the workfloor |
|
| 24/01/2008 | 14:30 | B111 |
All |
Annotation Methods |
Notes
The early work...
ISWC 1999 and 2000 had some of the first papers that used wearable accelerometers for activity recognition in general. One of the very first (Golding & Lesh, 1999) mentions actigraphy and monitoring patients with mood disorders as related work.. seems that the field hasn't progressed much?
Unresolved questions: (how) did the compass drift? --> update by kristof: Doh! It was the accelerometers, not the compass.. and the drift was about them shifting in the toolbelt
(Randell & Muller, 2000) is very short and proposes to recognise basic activities from accelerometer data.
The "What shall we teach our pants?" (Van Laerhoven & Cakmakci, 2000) paper illustrates the same, but stressing on the need for online learning with some user involvement, using Kohonen maps.
Annotation Methods
We had a meeting about annotation methods on Jan. 24, 2008. The notes of this meeting are here.