Title :
A robotic system for neuronal control
Author :
Wolpert, S. ; Laffely, Andrew J. ; Hilton, Judy A.
Author_Institution :
Dept. of Electr. & Comput. Eng., Maine Univ., Orono, ME, USA
Abstract :
In order to investigate neuronally derived algorithms for feedback and control, one joint of a conventional robotic arm was fitted with sensory and motor devices consistent with biological memory-motor models. In this system, a single potentiometer senses angular position of a given joint. From this signal, the forward and reverse velocity are derived, as are the acceleration and deceleration. Controlling the stepper motor that drives this joint is an IC-based artifical neuron that has been specifically designed for this task. It possesses multiple excitatory and inhibitory input ports, and its output signals are amplified to a full 5-V level. The system has been constructed and successfully used to demonstrate empirical physiologically based models of inhibitory positional feedback and synaptic learning.
Keywords :
CMOS integrated circuits; learning (artificial intelligence); neural chips; position control; robots; 5 V; IC-based artifical neuron; acceleration; angular position; biological memory-motor models; conventional robotic arm joint; deceleration; empirical physiologically based models; feedback; forward velocity; inhibitory input ports; inhibitory positional feedback; motor devices; multiple excitatory input ports; neuronal control; neuronally derived algorithms; output signals; reverse velocity; robotic system; sensory devices; single potentiometer; stepper motor; synaptic learning; Biological system modeling; Biology computing; Circuits; Control systems; Motion control; Neurofeedback; Neurons; Robot control; Robot sensing systems; Velocity control;
Conference_Titel :
Bioengineering Conference, 1992., Proceedings of the 1992 Eighteenth IEEE Annual Northeast
Conference_Location :
Kingston, RI, USA
Print_ISBN :
0-7803-0902-2
DOI :
10.1109/NEBC.1992.285961
