[ad_1]
They call me the master of robots – or at least they should. I grab a flat pack, hold his barcode under a red laser dot and place it on a tiny treadmill that is actually a robot. I click on a button to my left, then close the robot so that it meets my expectations, in the direction of one of the 300 rectangular holes of the ground corresponding to the postal codes. Once on site, the robot engages its conveyor belt and the packing slides from its back down a chute to the ground, where it can be loaded into a truck for delivery.
This is not an experimental system in a robotics laboratory. These are real parcels for real people using real robots, in tomorrow's Amazon sorting center, not far from the Denver airport. With a little luck, my robot friend and I have just successfully transported a package to someone from Colorado. Otherwise, blame the technology, not the user.
Seen from the top, the scale of the system is dizzying. My robot, a small orange plate known as "disk" (or more formally and mythically Pegasus), is only one of hundreds of its kind swarming on a "field" of 125,000 square feet pierced with chutes . It's a symphony of electric humming, with robots stopping at each other at intersections and delivering their packages to the slides. After each "mission", they form a neat queue in the stations of the periphery, waiting for humans to scan a new package, load the robots again and send them for another mission.
Matt Simon covers cannabis, robots and climate science for WIRED.
You do not have to look far to see how the logistics hidden behind your Amazon deliveries are overwhelming. On the other side of the building are four humans doing things at the old, standing at the bottom of a slide filled with packages. Frantically, they pick up the parcels, carry the label on each of them and guide them to the appropriate chute. At the bottom of the chutes, even more humans grab the parcels and stack them on pallets for delivery. It is extremely laborious and, in a word, chaotic.
Amazon needs this robotic system to overload its order processing process and make same-day delivery a widespread reality. But the implications affect the very nature of modern work: humans and robots merge to form a coherent workforce that promises to leverage the unique skills of both parties. Added to this is a familiar anxiety – even an existential enigma – that, as robots grow, push more and more people out of work. But in reality, it's not that simple.
If only the Luddites could see us fill the orders online now.
This Colorado warehouse is in a way a monument to robots. This is not Amazon accomplishment centers you've probably heard about, in which humans grab all items of your order and pack them in a box. It's a sorting installation, which receives all these boxes and puts them on trucks to your neighborhood. The distinction is important: these stocky discs are not responsible for finely handling your shampoos, books and t-shirts. They are mules.
Mules very very finely tuned. A system in the cloud, much like air traffic control, coordinates the route of each robot through the ground, taking into account potential interference from other drives on other routes. This coordination system also decides when a robot should take off and anchor in a loader and when to return to work. Sometimes the route selection can become even more complicated because particularly congested postal codes include more than one zipper. The system must therefore take into account the traffic patterns by determining the portal that a robot should visit.
"It's basically a very big sudoku puzzle," says Ryan Clarke, senior director of special technology applications Amazon Robotics. "You want every column and every line to have an equal number of drops. How can we make sure that each line and column are similar? The ultimate goal is to minimize clutter through a uniform distribution of field traffic. Thus, in addition to changing the routes of the robots, the system can actually change the channel assignments according to the demand, so that neither the robots nor the human sorters they work with encounter bottlenecks.
To map all this madness, Amazon runs simulations. These in turn explained how the discs themselves should perform. What is the optimal speed? What is the optimal acceleration and deceleration, as long as you want deliveries to be as efficient as possible while preventing robots from clashing? After all, a shock could throw a package on the ground, which other robots would notice with their vision sensors and route around, adding another layer of complexity to the terrain. (By the way, the robots have sensors at each end of their conveyor belt, so if a package starts to slide to the side, the tape automatically engages to tuck it in.)
Amazon runs complex simulations to coordinate robots in the field.
Amazon
The temptation could be to operate these machines as quickly as possible. "But it's like having a Ferrari in downtown San Francisco: all you do is make stops and starts," says Clarke. "We tried to adjust many different parameters and found that more speed and more acceleration actually had the opposite effect. They bumped against each other and caused more heaps. "
Ready for more complexity? Amazon had to modify the built space itself to keep the machines happy. Humans doing things old on the other side of the building, for example, like to bask in the photons flowing into the skylights. Above the field of the robot, however, the skylights are covered, because the reflections could shake the sensors of the machine: to navigate, they use a camera on the belly that reads the QR codes on the ground. Even the air conditioners hanging on the roof are modified. On the human side, they blow the air directly down, but above the robots, they blow on the side, because bursts of air could project light packets on the treadmills of the machine.
LEARN MORE
The WIRED guide of robots
Even worse, precarious packets such as liquids could send the system into chaos. Thus, although the system is automated, humans still monitor robots on flat screens below the field, where packages fall in the corridors, and respond to seizures. "Think if I had a package containing a gallon of paint and that gallon of paint had been damaged and had sunk in one of those chutes," says Steve McDonnell, general manager of the sorting center. "In a few minutes, I'm able to close that chute, redirect the drives to another chute, and I'm done."
The key here is flexibility – not a word that comes to mind when thinking about robots. Flexibility in robot journeys, in their destinations, in the number of robots on the field at a time. For example, you might think that the more machines there are, the better. Amazon could deploy up to 800 drives at the same time, but it could block the ground like city traffic. Instead, they usually operate 400 or 500, the others are parked on the side and waiting to be distributed.
Beyond the coordination of robots themselves, there is the question of how to make them good collaborators for human employees. Human work involves placing packs in cartons 6 feet high under the field, taking care not to discard large packages first. For this work to be manageable, the robots must distribute parcels between the multiple chutes for a particular postal code, so that a given chute does not overflow. At the same time, the system is considering the best way to group packages at the bottom according to their departure time, so that workers do not have to chase their hunting.
"The interaction between the associate and the readers is almost like a game of 3D chess," says McDonnell, "because you can optimize the field of readers, but then you can make the work of the Partner more difficult. below field."
Across the field From the human workers who distribute the parcels to the controls, a prototype robotic arm, named Robin, is at the end of a treadmill. His "hand" is a vacuum manipulator designed to hang boxes and flat packs.
This robotic arm is a test of what it might look like to further automate package reorganization work. The idea is that the conveyor will deliver packages to the arm, which will then load the drives. "We're going to feed things a little differently than we do with humans," said Rob Whitten, senior technical program manager. "We will not be content with a lot of things, we will instead throw soft balls. We will give it a little more structure so that it can handle it. "For parcels that they can not handle, as they were too heavy or of a weird shape, humans would step in to help."
While walking through the line of robot-loaders, I come across a worker who set aside a broken box that spilled bottles and other bowels. This human with unique abilities could do two things here: use his problem-solving skills to say, Something wrong, I have to put them aside, then manipulate these objects with extremely fine motor skills.
This robot arm has no prowess to solve problems, nor fine motor skills. Imagine if clear laundry liquid had broken inside a package and soaked the bottom of the box. A human being could smell the detergent or feel its viscosity before seeing it. A robot arm dependent solely on sight would miss the problem, loading the package on a control robot that would then swindle the ground.
same if they had a semblance of judgment, robots are always awful to handle complex objects like bottles. That's why Amazon simplifies things, with a suction arm designed to stick to flat surfaces, as opposed to an analog of the human hand. For a while, humans will (almost) literally hold the hand of these robots.
Matt Simon
The bottom line is that we humans have to adapt to machines as much as machines have to adapt to us. Our careers depend on it.
Amazon is running simulations to find out how to keep its human workers comfortable loading robots with packages. This includes their range of motion from the point of view of ergonomics and their safety. Or questions such as the best way for a human to grab a parcel, scan it, place it and reach out to press the button that sends the robot on the road. "There is an art to make perfectly harmonious between what the robot does and what humans do," said Brad Porter, vice president of robotics at Amazon.
This is the kind of dynamic environment that is ideally suited to the development of Amazon's next iteration of its system. The company is working on a new modular robot called Xanthus with different fasteners, for example to hold containers instead of using a treadmill. In a sense, this machine will bridge the gap between distribution centers, where humans load products into boxes by hand, and sorting centers, where they work mostly with assembled boxes.
The new modular Xanthus robot from Amazon can be equipped with accessories allowing it to carry different types of goods.
Amazon
"You can see how, combined with the addition of a sensor platform, perhaps, you can create a stand-alone drive that carries the gear," says Porter. But you can also take this same sled end and replace the bag transport unit with a carrier tray and deploy it in the sorting center.
This is the huge advantage of Amazon: it has the funds and talent to develop robots internally, each of them to solve problems specific to Amazon. Other warehouses are beginning to be robotic, but they work with machines from other companies. For example, Boston Dynamics, maker of the impressive SpotMini and Atlas hypnotics, will soon be introducing a box-lifting robot called Handle. But it is a general machine, not developed exclusively for a customer.
Amazon, on the other hand, can iterate through a robot until it is perfectly suited to a specific task. "They are building it for themselves and for their environment and circumstances," said John Santagate, Director of Service Robotics Research at IDC, who conducts market research. "It is difficult to create a product that suits everyone."
And every employee hired in a role in front of the machine does something that no human has ever done: lower level employees in this facility have been encouraged to help monitor the gigantic system that surrounds them. , as well as humans intimately integrated with it. "The fully automated or highly automated distribution center is not a North Star that we are trying to touch," Porter says. "Are we seeing additional levels of automation, higher and higher? Yes, I think it will increase as the capabilities of our systems increase. "
Here is the big question, though. Is this type of automation required to completely replace human jobs or to replace rooms of these jobs? "Most research seems to suggest that automation is moving towards shifting skills, not jobs," says David David Edelman, former special assistant to President Obama on the issue. digital economy, and now director of the MIT project on Technology, Economy and National Security. "This suggests that these people can, through Amazon, be re-empowered or exploit other skills already acquired in the same job."
These days, industries that are short human labor need automation to survive. Consumers still want fresh produce, but California farms face a labor shortage of 20% and are turning more and more to agricultural robotics. Amazon's business is booming, but America is experiencing a historically low unemployment rate and so workers have plenty of work opportunities. "The demand from this company is increasing, but the availability of resources to meet this demand does not necessarily increase," says Santagate. "In fact, it's probably in the process of contracting." Robots fill the void.
Here, in tomorrow's sorting center, I walk along the field and hear the morning break for humans, called with the help of loudspeakers. The training robots continue to move for a few minutes, with their incessant electric white noise, until the place suddenly becomes almost silent. After delivering their packages to the chutes, the robots are out of work. They park on the side of the field, some in charging stations. It is only when the speakers sound the end of the break that the machines restart, ready for their users to feed them more.
If only the Luddites could see our codependence now.
More great cable stories
[ad_2]
Source link