Really good science fiction movies help us imagine a more exciting future, where giant robots engage in combat for our viewing pleasure or there’s a cure for baldness. These films also remind us just how far those fantastic visions are from our more mundane reality. By now we should all own flying cars and be able to control machines with our minds. At the very least, we hoped by now to have artificially intelligent robots catering to our every whim. It turns out we’re not far off from that particular future, at least when it comes to buying stuff.
Retail automation through robotics and AI is changing the way our consumption society consumes. On one end, machines can predict the sort of useless stuff we like to buy online, while at the other end, robots are delivering said useless stuff to our front doors. Even when forced to buy something from an actual store, autonomous check-out solutions ensure the possibility of interacting with a human is relatively low. Behemoth retailer Amazon (NASDAQ:AMZN) put retail automation technology in the spotlight with the 2012 purchase of Kiva Systems, a company which developed warehouse robots and related technologies, and which was acquired for $775 million.
Now known as Amazon Robotics, based in Boston, the wholly owned subsidiary combines “autonomous mobile robots, sophisticated control software, language perception, power management, computer vision, depth sensing, machine learning, object recognition, and semantic understanding of commands.” All of that so you can get stuff made in China delivered to you in days, if not hours. A number of startups have attempted to fill the hole created when Amazon snatched up Kiva, some of which we profiled in our feature on industrial warehouse robots. Companies we’ve profiled like inVia Robotics, which has raised $29 million in funding, are developing warehouse robots that can work right alongside humans, giving rise to a term called cobots. (Most of that funding came just days ago in the form of a $20 million Series B round.) Let’s look at some other startups working on warehouse robots.
Warehouse Robots for Fulfillment Centers
If you’re going to have to collaborate and work alongside a robot, it might as well be one named Chuck. Chuck is a warehouse robot from a Boston-based startup called “6 River Systems”, which has raised $46 million, including a $25 million Series B at the start of the year. The co-founders had both previously worked for Kiva Systems, now Amazon Robotics, so they had something of a track record before starting the company in 2015. The startup claims that Chuck, which looks a bit like a treadmill with the readout awkwardly stuck in the middle, is technologically similar to autonomous vehicles, with sensors like LiDARs and machine learning algorithms for cloud-based navigation and control.
The company says 70 percent of warehouse operations are still mostly manual, so there is plenty of work for Chuck. The warehouse robot actually comes in a couple of different sizes and configurations to handle various tasks. Chuck seems like the most bro-ish cobot of the crowd, designed to work closely with its human slave associate while telling off-color jokes and hitting on chicks.
Warehouse Robots for On-Demand Grocery Delivery
A three-year-old company called CommonSense Robotics out of Tel Aviv, Israel, raised $20 million earlier this year in a bid to take a piece of the on-demand grocery market. The Series A in February brought total funding to $26 million. CommonSense appears to be engaged in a version of urban guerrilla warfare, with plans to operate out of existing retail spaces no bigger than 10,000 square feet in the middle of cities, rather than in cavernous warehouses on the outskirts. So, it’s not selling its low-profile warehouse robots, but instead contracting with retailers to put together orders for them directly in what CommonSense calls micro-fulfillment centers. The Israeli company has designed robots that maneuver through tightly packed stores, picking up totes filled with groceries that a human then packs for delivery. More robots then take those packaged items to a dispatch area.
While the robots are outfitted with sensors, a central computer actually coordinates the complex choreography for the entire system, learning where to make improvements, such as relocating popular items to make them more easily accessible for the tiny robot army. The company claims that its warehouse robots can fill the average order in less than five minutes, about 10 times faster than a human, according to TechCrunch. And probably without crushing the Charmin.
Of course, it’s not just startups like CommonSense thinking about filling grocery orders using robots. Walmart is spending heavily on robotics technology, and just 13 hours ago, announced that they’re going to use a robotics solution for online ordering called Alphabot. It’s being built by a company we haven’t come across before called Alert Innovation, and apparently, they designed the solution just for Walmart with an undisclosed amount of funding. From the horse’s mouth:
When completed, automated mobile carts will retrieve ordered items – stored warehouse-style in this new space – then deliver them to our associates at one of four pick stations.
That’s great, because when we used Walmart’s online ordering service a few months back, the tools they had working in the Vancouver Washington store said they didn’t get the +$400 order we posted the day before, and then refused to rectify the problem by filling the order while we stood in the store twiddling our thumbs like a bunch of morons. Told you that would come back to haunt you lads.
Warehouse Robots that Work in Three Dimensions
Also founded in 2015, A French startup called Exotec, led by former roboticists and software engineers with GE Healthcare, has scooped up $21.2 million in investments, including a $17.7 million Series B in June. Its robot army can move like the Wonkavator—it can go up and down, sideways and slantways, and frontways and squareways. Well, more specifically, it can navigate horizontally and vertically, using a laser scanner to help it maneuver through three-dimensional space. The fast-moving Skypod is reputedly four times faster and can climb five times higher compared to traditional “shelf-mover” robots. It can also handle twice the number of picks per hour and save warehouse workers up to 10 miles of walking per day, which they can now spend looking for new jobs.
The young startup has several customers, including Cdiscount, A French e-commerce company. The new Euros will go toward deploying another 1,000 Skypods by 2019; we just hope the AI brain behind this system isn’t called Skynet.
Cloud-based AI for Warehouse Robots
Munich, Germany-based Magazino raised $24.8 million in disclosed funding back in February for its wonkavator-sized (and looking) warehouse robots. The company claims that funding round represents one of the “largest funding rounds for robotics startups in Europe.” If true, then it sounds like $20-million-plus Series A is a pretty good round. Magazino’s robots are more of the old-school industrial variety but with a new age software called ACROS (Advanced Cooperative Robot Operation System) that forms the “brain” of the robot. This software creates what the company calls perception-guided robots that use cloud-based artificial intelligence, a technology we’ve come across before.
The company has developed several kinds of warehouse robot systems, depending on the job. If you had to pick a favorite, however, it’s probably TORU, which can grab individual objects such as shoe boxes by using “3D-camera technology and numerous sensors, the robots can localise, identify and securely grasp objects on shelves.” That’s good because some of our MBAs have big feet. And you know what they say about big feet. Big feet, big shoes.
Arming Warehouse Robots
Plus One Robotics came out of stealth mode this year with a $2.4 million Seed round in May. The San Antonio, Texas-based startup, founded in 2016, is developing a robotic arm that can move dynamically, almost organically, to pick up packages and parcels along a conveyor belt. That’s different from your normal warehouse robot arm that makes a single repetitive action from a fixed position.
The PickOne’s accuracy is thanks to 2D and 3D cameras that image the input stream of made in China e-commerce booty in fractions of a second in order to choose pick positions for the arm. It can do 25 picks per minute, the same as a human, but like forever.
Update 11/17/2018: Plus One Robotics just raised $8.3 million in Series A funding which brings the company’s total funding to $10.6 million to date.
Maybe Warehouse Robots?
There is a bit of mystery about San Francisco-based Darkstore, which has picked up $1.7 million since it was founded two years ago. For instance, all the news stories are pretty much the same: Darkstore works by enabling e-commerce retail brands to store their inventory in major cities and provides on-demand and same day home delivery at lower prices. That’s because it squats in leftover space in storage facilities, like abandoned malls, and charges a modest percentage on each item that leaves its pop-up warehouse. It’s unclear where robots fit into this picture, as the most high-tech part of the operations sounds like a smartphone app. However, the company’s website says it leverages the “latest in robotic automation and AI.” It’s almost as mysterious as the Dark Knight.
Swift Warehouse Robots
Founded in 2012, Pittsburgh-based IAM Robotics has grabbed about $1 million for a warehouse robot that features a robotic arm like Plus One Robotics but on a mobile platform closer-looking to Magazino. Its “mobile manipulation robot,” Swift, can travel autonomously through the warehouse, finding and identifying products, and then picking them up with one of a variety of grasping ends that can lift up to 15 pounds.
Their system also includes SwiftLink, software that manages and tracks the robots, as well as Flash, a high-tech inventory instrument that records a product’s barcode, 3D dimensions, weight and a high-resolution image of any product placed inside it. Flash is sold separately.
Once upon a time, warehouse robots were large, dangerous mechanical monsters that could easily crush a soft-fleshed humanoid. Today, they have been uncaged but subdued, thanks to powerful but increasingly inexpensive sensor systems and the magic of artificial intelligence. In fact, it’s interesting that while most of these robot systems for e-commerce are using some form of AI, computer vision or machine learning, the companies aren’t exactly shouting it out from the rafters. That might be because the algorithms aren’t any better than fancy software. Or, as we have proposed, AI is becoming like electricity, where of course you’re going to build autonomous robots with artificial intelligence. Doing it any other way would be stupid.