Communications of the ACM
Synergy Emergence in Deep Reinforcement Motor Learning
Tohoku University researchers have observed something similar to motor synergy in robotic agents using deep reinforcement learning algorithms.
Credit: Tohoku University (Japan)
Researchers at Tohoku University (TU) in Japan found robotic agents using deep reinforcement learning (DRL) algorithms act in a way similar to motor synergy, which enables a body's central nervous system to use a smaller set of variables to control a large group of muscles, simplifying control over coordinated and complex movements.
DRL can help solve complex robotic tasks while reducing manual operations and achieving peak performance.
The researchers used two DRL algorithms, the classical TD3 and the high-performing SAC, on two robotic agents that completed a total of 3 million steps.
Said TU's Mitsuhiro Hayashibe,"After employing deep learning, the robotic agents improved their motor performances while limiting energy consumption by employing motor synergy."
From Tohoku University (Japan)
View Full Article
Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA
No entries found