Answers are grounded in this paper and cite their sources.
An object recognition system using the state-of-the-art YoloV5 algorithm has been trained to recognize different mechanical parts randomly placed on the work plane of a collaborative robot. The parts are classified according to their technological characteristics and then precisely localized using a wrist-mounted camera. The aim is to improve the performance of collaborative human-robot stations by freeing them from the constraint of having to place the parts in a fixed, well-defined position and in the same order as the assembly sequence. The result is the full exploitation of human-robot collaboration for assembly purposes.
No takes yet. Share an insight, caveat, or question.
Antonelli et al. (2024) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: