Development of a learning module using a virtual environment to demonstrate EMG and telerobotic control principles.

Abstract

A prototype system was developed for use as a teaching tool, allowing students to link EMG monitoring, data smoothing, robotic control, and neural network training within a rapid prototyping virtual environment (VE). The VE software allowed for the rapid development of scenarios and, when linked with EMG data input to a neural network, allowed the user to control an artificial world containing a virtual arm. Student teams then attempted to control the arm in the VE while performing the tasks by use of a neural network system they had specifically developed and trained using their own EMG signals. The results from their system were then translated into a form that enabled the control of a real robot. Students enjoyed the challenge and uniqueness of the module, and were enthusiastic about extending the concept to other areas of interest.