Control of robotic motion with neuron-like elements

Abstract

Using a controller made up of models of biological nerve cells, a controller network for a robotic limb was designed that incorporates pertinent environmental information into its motion directives, and does so without the need for conventional computers. Feedback information pertaining to the position, velocity, and acceleration of the robotic limb is encoded into compatible frequency-based signals and factored into an overall formula for the control of motion. The controller network was designed to be upwardly compatible with a virtually limitless amount of feedback data pertaining to its environment. This method is more comprehensive than conventional robotic controllers and with other applications of the controller network, such as prosthetics, the motion of electronically controlled limbs should become more lifelike.<>