WSEAS Transactions on Systems and Control, ISSN 1991-8763, 2224-2856, Volume 13, 2018 art. # 6, pp. 44-53, Issue published: February 8, 2018

Link: http://www.wseas.org/multimedia/journals/control/2018/a125903-037.php

A research area in control studies the human capabilities to control dynamical systems, because the brain could be the most powerful control centre. However, there are many limitations for the application of a human in the loop. For instance, the time response of a system could be too slow or too fast for a human. Thus, losing attention or not having enough time to make decisions becomes the main challenge in this control field. This paper proposes the use of time scaling plus the learning in a Neural Network to overcome those time constraints. This new control strategy starts by scaling the system in time until a comfortable value for a human, then a Neural Network learns the control actions from the human, and finally that Network runs at different time rates, which will be applied to control a robotic differential platform. The new control procedure improves the control performance carried out by a human by properly changing the time constant of the robot model. We also consider the problem of possible variations of the robot platform after the training stage by using a dynamic version of the back propagation algorithm.