[From Bill Powers (2006.07.09.1910 MDT)]
Marc Abrams (2006.07.09.1831) --
Bill, is there any way of hooking this block diagram up with the tracking task?
That's basically what I do when matching a model to the experimental data. The disturbance is the same pattern of disturbances that is used during the experiment with the human subject (actually the sum of two disturbances, one affecting the target position and one affecting the cursor position). The program can automatically adjust the input delay, the reference signal, the output gain, and the output time constant to get the best fit between the model's output quantity and the handle movements produced by the human subject.
You can also, during the analysis, manually alter the main parameters while the tracking task is repetitively simulated at many times the real speed, with a plot overlaid on the experimental data plot. The parameters are adjusted by spinners so you can see the effects immediately, just as you do with sliders in the LiveBlock program. The main difference is that there's no diagram of the system displayed, and not all the system variables are plotted.
The same is done in other kinds of control experiments not involving tracking: control of shape, color, orientation, and other aspects of physical situations that can be represented on a computer screen. Tracking is just one kind of control. There will be some of that in the new book. Tracking, of course, is the easiest to implement and for people to understand -- and it's sure easier on my limited graphics programming skills. In principle any control task can be analyzed the same way.
Best,
Bill P.
···
--
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.1.394 / Virus Database: 268.9.10/383 - Release Date: 7/7/2006