Control of a Robot Arm with Artificial and Biological Neural Networks

Shultz, Abraham Michael (University of Massachusetts Lowell) | Lee, Sangmook (University of Massachusetts Lowell) | Shea, Thomas B. (University of Massachusetts Lowell) | Yanco, Holly A. (University of Massachusetts Lowell)

AAAI Conferences 

To perform research on learning in cultures of mouse neurons, a hardware and software system for interfacing a biological neuronal culture to a robot arm has been constructed. The software architecture is modular, which permits simulated neurons to be used in place of biological neurons. In both cases, the activity of the culture over time is represented as an activation vector that captures recent spatiotemporal patterns of neuron firing. The activation vector is converted into control signals for the arm in a manner that can be generalized to multiple degrees of freedom. Preliminary results from the system with both simulated and biological cultures are presented.

Duplicate Docs Excel Report

None found

Similar Docs  Excel Report  more

None found