Robotic Skill Acquisition Based on Biological Principles

Abstract

As humans acquire motor skills, slow, stiff, and cautious movements give way to smooth ballistic trajectories requiring much less mental concentration. Complex, yet efficient, sensorimotor responses can be learned by an individual if given verbal explanations of how to accomplish a task, examples of typical motions involved, and time to practice. The aim of designers of robot control systems is to emulate behavioral characteristics of biological systems in order to maximize ultimate robot capability while minimizing the amount of design effort required to obtain it. This chapter presents an approach to robotic skill acquisition that attempts to parallel the training of an athlete (the robot) by a coach (the designer), whereby the robot learns through experience how to perfect tasks initially specified in a high-level task language. Knowledge-based system components encode neural network learning strategies, and skill acquisition is associated with the shift from predominantly feedback-oriented, rule-based representation of control to predominantly feedforward, network-based form. To demonstrate its utility, the technique is applied to the problem of learning how to control the longitudinal dynamics of an airplane during approach and landing. Conclusions are drawn regarding how a biologically inspired control technique can address issues related to adaptive system design and man-machine interfacing.