Open panel discussion

The first discussion originated with a question from the floor concerning self-adaptive controls. In particular, the panel members were asked if they could provide more information about a device in which a parameter was not varied to improve control but where a simple model of the desired system, operating in parallel with the actual system, was used to determine the desired output. Then, this output, and possibly its higher derivatives, was compared with the actual output to provide a correction signal to the actual system. The second topic from the floor also concerned adaptive control of the brain and nervous system. A description was given of neurophysiological experiments on cats. Electrodes were inserted into a eat´s nervous system along an auditory pathway. Audible clicks were presented to the cat and the electrical activity at the electrodes was observed. After a few treatments, the magnitude of the electrical signals would decrease at each of the electrodes, which were essentially detecting signals in anatomical feedback loop. However, after this habituation or conditioning to sound, the electrical signals could be increased again if the cats were given electrical shocks simultaneously with the audible clicks. This example was used to illustrate an actual nonlinear control system,where unimportant input signals can be reduced to a level of no response, but not below a monitoring level. This is similar to reducing one input to allow the reception of more important signals until that input again becomes important. Then, a new topic of discussion emerged concerning the training of engineers and the proportion that should be maintained between pure and applied mathematics in an engineering curriculum. The panel proceedings and the problems of nonlinear controls are then summarized by the panel members.