University of Cincinnati (UC) engineers taught two robots to move a couch together without either robot directing the other.
The robots employ genetic fuzzy logic, which emulates human reasoning by substituting degrees of rightness or wrongness for simple binary classification (yes-no), while genetic algorithms learn from past results and optimize performance.
In simulations of hauling a virtual couch around two obstacles and through a narrow door, the machines successfully completed the task 95% of the time; they also completed the task when encountering two unfamiliar obstacles and a door in a different location 93% of the time.
UC's Andrew Barth said, "If you can train robots to work semi-independently with as little information as possible, then you made your system more robust to that failure and made it easier for large groups to collaborate."
From UC News
View Full Article
Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA
No entries found