 (1).png){kind=spacer}
1 year ago
23
It's a problem that drives the parents of most teenagers mad.
But clothes piled up on the bedroom floor may soon be a problem of the past as scientists have developed a cleaning robot to pick up dirty laundry.
Scientists from UC Berkeley say their clothes-collecting bot solves the 'Teenagers Problem; of how to most efficiently pick up mess.
Using a combination of colour and depth-sensing cameras the robot collects laundry into piles before finding the best place to grab the clothes.
Professor Ken Goldberg, the study's lead author, says that this technology could become commercially available in the decade, meaning we may soon all have a robotic helper around the house.
Scientists have come to the rescue of teenagers and parents everywhere by designing a robot that can pick up dirty laundry from the floor
What is the Chinese Postman's Problem?
In 1962, a Chinese Mathematician named Kuan Mei-Ko started thinking about the way post was delivered.
He wondered how a postman, who had to deliver mail to a number of different houses, could find the quickest way to visit all their stops.
The paper he wrote was translated into English and sparked massive interest.
The problem may seem easy at first but as the number of stops and streets grows finding the solution becomes extremely difficult
Named after the Chinese Postman's Problem, a famous mathematical puzzle, the Teenager's Problem asks how you can clean a room with the least trips to the basket.
While a human might just be able to pick up all the items at once, a robot with a grabber might only be able to pick up a few at a time and so must be efficient.
To solve this problem, Professor Goldberg designed an AI that could pick up clothes by controlling a robotic arm.
Professor Goldberg told MailOnline: 'It self-teaches by repeatedly running a cycle where it picks up the garments from a flat surface, puts them into a basket, and then dumps the basket back onto the flat surface, and repeats.'
Though this might seem needlessly messy, Professor Goldberg says this allows the robot to 'learn' how to optimize the number of garments it can take each trip.
In the experiment, the AI repeated this cycle 200 times and grabbed more than 2,000 garments in the process.
The scientists first tried two different techniques, one using a standard colour camera and another using a special depth-sensing camera.
The colour camera could identify where clothes were but often failed to find the best spot to grab multiple pieces of clothing at once.
The depth-sensing camera, on the other hand, was great at seeing where lots of clothes piled up but struggled to make sense of individual garments.
Professor Goldberg and his co-authors say each of these methods reduced the number of trips needed to clear the room by about 20 per cent.
However, when combined together with the ability to stack clothes into piles the robot could be 67 per cent more efficient.
The robot collected and scattered a bundle of clothes 200 times to learn the most efficient technique for collecting laundry
This diagram shows how the AI selects the best place to grab, by working out how the clothes overlap
Asked if the technology could soon become commercially available, Professor Goldberg told MailOnline: 'We hope so. Millions of teenagers (and their parents) are counting on it!.
He adds: 'The ability to efficiently pick up garments could be very useful for senior citizens, in hospitals, hotels, and in retail clothing stores.'
However, Goldberg also says this won't be available right away since 'current mobile robots with arms attached are a bit too expensive to justify this for home use.'
This robot also won't be the perfect household helper just yet as it currently lacks the ability to sort garments according to colour.
However, the researchers do suggest that sorting ability would be a 'natural fit' for the techniques used to optimize grabbing.
Researchers found that they could make the robot almost 70 per cent more efficient by teaching it how to stack up clothes on the floor before picking them up
As robots become more advanced and more widely available, some experts believe we might be on the verge of a chore-less future.
A recent study published by Oxford University estimated that almost 40 per cent of household chores would be automated by 2023.
The researchers suggest that grocery shopping will the the area in which robots will take up most of the slack, with 59 per cent of the effort turned over to robots and algorithms.
Previous research has shown that Brits spend about 43 per cent of the total time they spend working and studying on unpaid labour, including household tasks like tidying and cleaning.
If this robotic cleaner were to come to market, it would join a growing number of robotic chefs, cleaners, carers, and labourers.
The creators of these household robots say the will soon be able to do any job a human can.
