Taking open source to the nextlevel.
3D Printed Robots Will Save The World
Taking open source to the next level.
Across the world, from Greece to Guatemala; Illinois to Iran, people are assembling life-size humanoid robots. They’re creating a legion of AI-powered clones that anyone with access to a 3D printer can build.
It began by accident. By day, French sculptor and designer, Gael Langevin, creates objects for big brands to use in commercials. In 2011, a client asked for a futuristic prosthetic. Langevin decided to 3D print one. “I searched online for a prototype, but there was nothing,” he says. In the end, the idea was dropped, but by then, Langevin was hooked.
We spoke on a bright Monday morning in spring. He tells me last night he was up “til 4 am working on a fiddly piece of code.” It’s clear this side project is peripheral in name only.
“I’ve loved hands since I was a kid,” Langevin explains enthusiastically. He decided to sculpt a hand and 3D print it. There was nothing like it available online, so he shared the templates he used to build the hand and the codes necessary to control it on the popular 3D printing site Thingiverse. According to Langevin, the reaction was immediate:
“It went wild. Within a few weeks people from all over the place contacted me to ask if it could really be used as a prosthetic. So I decided to go on, making the arm, the shoulder… and every time I designed something I shared it online so everyone could reproduce it.”
And thus, InMoov — the world’s first open source 3D printed human-size robot — was born.
Anyone can use and adapt the software and hardware — so the robot is constantly evolving.
The project quickly reached critical mass and Langevin migrated the files to a dedicated site, a one-stop shop for its many makers. Anyone can use and adapt the software and hardware — so the robot is constantly evolving. Thanks to this open source approach and 3D printing technology, Langevin’s talent and creativity can be used as a stepping stone for anyone interested in robotics.
Initially, Langevin was disconcerted by the way other makers built upon his designs. “When I saw pictures of other robots, they were different colors, for example, and I was like, ‘Whoa! Why?’ Now when I see one, I think wow, this is interesting.”
Open Bionics creates affordable 3D printed bionic hands for amputees.
Langevin’s project provides the foundation for pioneering initiatives like E-nable, a US-based global community of over 6,000 individuals who use their 3D printers to create prosthetic hands and arms for those in need. Using Langevin’s early design, E-nable continues to leverage its network to tackle other problems through open source technology.
Across the pond, the Bristol-based company, Open Bionics, creates affordable 3D printed bionic hands for amputees. Controlled via EMG sensors, the hands allow users to grip and manipulate objects and perform everyday tasks. These prostheses usually cost hundreds of thousands of dollars, but 3D printing lowers costs and makes the technology accessible to those who would otherwise be unable to benefit. The organization has won multiple awards and in 2016, Open Bionics made Fast Company’s list of the top ten most innovative robotics companies.
InMoov is also used in universities and schools across the world. The robot serves as an experimental portal for all kinds of tech, from sensors and vision tracking to human relation and artificial intelligence. The teamwork, problem-solving and research skills involved are, according to Langevin, essential aspects of education today.
The beauty of open source is that there are no predetermined goals aside from those you set yourself.
The extent to which people change Langevin’s original designs varies: Some follow all the rules or make slight tweaks, while others, like E-nable and Open Bionics, take an idea and run with it. The beauty of open source is that there are no predetermined goals aside from those you set yourself.
After 3D printing the robot, InMoovers must painstakingly put it together and sync it with the AI software, a challenge that requires skill, patience and perseverance. It’s no surprise the puzzle appeals to hobbyists. “You have to be a kind of MacGyver,” says Langevin. “Some 16-year-olds follow the tutorial and come up with a fully functioning robot without asking a single question. And then you’ll see a group of 10 adults working together over six months, and they don’t even manage the arm.”
InMoov’s software is as important as its 3D printing possibilities. Powered by My Robot Lab, an open source framework developed by American programmer Greg Perry, the robot can answer questions by recognizing your voice and associating it with your face. Using Microsoft Kinect, the robot can move around an area, detecting and avoiding objects. You can teach it gestures by performing them and asking the robot to mimic your movements.
What’s more, a virtual version of the robot can connect to one in the real world. You can control a robot on your computer screen and a robot in China will replicate those movements in real life. It’s this collaborative approach that inspires Langevin. “Everyday you load your robot and you don’t know what it’s going to do.”
Langevin thrives off collaboration, but he is also driven by darker realities. He says:
“We need AI that’s not made by companies. In the near future, there’s going to be powerful AI — more powerful than us. It’s quite scary to imagine it made by private companies. This project has led me to the conclusion that we need an AI that is totally open source — that everyone can modify. It has to be totally transparent, so if anything happens, we have access to everything.”
Like any parent, Langevin’s work is never done.
The good news is that plenty of people are already onboard. Thanks to the power of new technologies, people are working together in unprecedented ways, using the networks as a blueprint for creating a better world. Connected, motivated and empowered, there’s no telling what they might achieve.
As for InMoov’s founding father, like any parent, Langevin’s work is never done. More sleepless nights await. In addition to diving into machine learning, he still has to finish and release the final parts of the robot’s body. “Everyone’s waiting for the lower legs,” he tells me.