Robots are becoming more capable of performing household tasks, but folding laundry remains a major challenge for researchers. Although the task appears simple to humans, it requires machines to understand how different fabrics move, bend, stretch, and change shape during handling. Traditional robotic systems often use a method known as pick and place, which involves identifying a point on a piece of clothing, lifting it, and placing it at a predetermined location. However, this approach struggles when fabric becomes tangled, wrinkles unexpectedly, or differs in weight and texture. Researchers are developing more adaptable systems to improve performance. One example is AdaFold, a robotic folding method that monitors fabric movement and adjusts its actions during the folding process. In a 2024 study, AdaFold achieved an Intersection over Union score of 0.83, compared with 0.41 for the conventional approach when folding a rectangular piece of cloth in half. Human hands also have advantages because they can sense texture, flexibility, and weight. Robots must develop more advanced gripping capabilities and better reasoning to handle different garments reliably. Researchers are also improving training data and computer simulations to help machines understand clothing more accurately. Although these developments offer progress, fully automated laundry folding remains a difficult goal, and household robots may need further advances before they can manage the task consistently.
Read more here
https://www.smithsonianmag.com/innovation/when-will-robots-take-over-laundry-folding-180987801
