Communications of the ACM
Machine Learning Helps Robot Swarms Coordinate
Engineers at Caltech have designed a new data-driven method to control the movement of multiple robots through cluttered, unmapped spaces, so they do not run into one another.
Credit: CalTech
California Institute of Technology (Caltech) engineers have designed a machine learning technique to govern the coordinated movement of flying robot swarms in order to avoid collisions.
The method incorporates the Global-to-Local Safe Autonomy Synthesis (GLAS) algorithm that mimics a complete information planner with local information, and Neural-Swarm, an enhanced swarm-tracking controller that learns complex aerodynamic interactions in close-proximity flight.
These components eliminate the need for robots to obtain a complete image of their environment, or of their fellow robots' intended path; instead, they learn on-the-fly navigation, incorporating new data as they enter a "learned model" for movement.
Tests with quadcopter swarms showed GLAS and Neural-Swarm outperformed state-of-the-art controllers.
Caltech's Yisong Yue said, "These projects demonstrate the potential of integrating modern machine-learning methods into multi-agent planning and control, and also reveal exciting new directions for machine-learning research."
From Caltech News
View Full Article
Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA
No entries found