Robot avatars have been a sci-fi staple for decades, letting fictional humans remotely carry out delicate and dextrous tasks, from hospital care to mining on the Moon. But, as usual, our imagination outstrips reality. The best commercially available “avatars” are telepresence bots, which are little more than iPads on wheels, while cutting-edge humanoid robots look flashy but struggle to keep their balance. The X Prize Foundation wants to change this, and today, it announced its latest challenge: building “real-world avatars” by 2021.
As per the foundation’s modus operandi, the goal here is to spur development by offering teams cash prizes. Researchers will sign up to enter the prize and develop avatars (the deadline for entry is October 31st) and will have to meet certain criteria. There will be two milestone prizes in 2020 and 2021 worth $1 million each, and one grand prize worth $8 million.
The foundation has not detailed the exact challenges and specifications these avatars will have to meet, but the broad strokes are ambitious. Teams will have to build avatars that will enable users to “remotely see, hear, touch and interact with physical environments and other people” from 100 kilometers away. These robot avatars must also be able to “execute tasks across a variety of real-world scenarios,” with the foundation citing future applications like helping provide “critical care and deploy immediate emergency response in natural disaster.” The prize is sponsored by Japanese airline ANA.
Speaking to The Verge, X Prize Foundation CEO Marcus Shingles said the goal was to create a truly multipurpose robot. We’ve already developed robots for specialized tasks like deep-sea mining and moving patients around hospitals, but the hard part is creating a machine that can take on a variety of tasks — just like a human.
“It’s intended to be multi-purpose,” says Shingles. “Imagine a remote village with a single avatar system that provides utility to everyone in the village. The robot could be controlled by different operators to repair solar panels, help with carpentry, and lots of other jobs.”
Shingles says he imagines the winning team could combine a number of existing technologies, including virtual reality and artificial intelligence. But, he says, the foundation doesn’t want to be too specific about what it thinks the final design would look like. Going the humanoid route is certainly a possibility (especially if you want the robot to be able to use tools built for humans), but it’s not a certainty.
“We were hesitant to even give any graphical images of what the avatar looks like because we don’t want to lead the witness too much,” says Shingles. “It’s also why we’re not going to get real specific about what you have to do in the final [tests] because if we say you have to dig a hole, we’re going to have a bunch of teams making avatars with shovels.”
Looking at the current cutting-edge developments in robotics, it’s also hard to imagine a legged robot getting very far in such a competition. As we’ve seen from events like the DARPA Robotics Challenge, bipedalism is still very much a human specialty. It just takes too much work to maintain balance using electric or pneumatic motors, when four legs are more stable and wheels are faster. But, if the prize money can attract research teams, the X Prize Foundation could end up fostering new paradigms, which the institution hopes will lead to new products that benefit humanity.
“Any time we do an X Prize, we’re always looking to generate a commercial appeal, because if we can create that breakthrough technology, we need the market to help adopt it, to get it to scale, and get the price down,” says Shingles.
Of course, just because there’s money on the line doesn’t mean the problem will get solved. The Lunar X Prize, which aimed to get rovers on the Moon using private funding, came to nothing this January when all the teams involved missed the final deadline. But even a failure can push the field forward, and this latest X Prize will hopefully achieve more than that.