Engineers design autonomous robot that can open doors, find wall outlet to charge – sciencedaily
One of the flaws in the idea that robots will take over the world is that the world is full of doors.
And the doors are kryptonite for robots, said Ou Ma, professor of aerospace engineering at the University of Cincinnati.
âRobots can do a lot of things, but if you want one of them to open a door on their own and go through the door, that’s a huge challenge,â Ma said.
Students at the Intelligent Robotics and Autonomous Systems Lab at UC have solved this complex problem in three-dimensional numerical simulations. Now they’re building an autonomous robot that can not only open its own doors, but also find the nearest wall outlet to charge without human assistance.
This simple breakthrough in independence is a huge step forward for auxiliary robots that vacuum and disinfect office buildings, airports and hospitals. Auxiliary robots are part of a $ 27 billion robotics industry, which includes manufacturing and automation.
The study was published in the journal IEEE access.
Lead author of the study, Yufeng Sun, a doctoral student at UC College of Engineering and Applied Science, said some researchers solved the problem by scanning an entire room to create a 3D digital model so the robot could locate a door. . But it’s a time-consuming, custom solution that only works for the particular part that is being scanned.
Sun said developing an autonomous robot to open a door poses several challenges.
The doors are available in different colors and sizes with different handles which can be slightly higher or lower. Robots need to know what force to use to open doors in order to overcome resistance. Most public doors close automatically, which means if the robot loses its grip, it has to start over.
Since UC students use machine learning, the robot must “learn” itself to open a door, mostly through trial and error. It can take a long time at first, but the robot corrects its mistakes as it goes. The simulations help the robot prepare for the actual task, Sun said.
âThe robot needs enough data or ‘experiences’ to help it train,â Sun said. “This is a big challenge for other robotic applications using AI-based approaches to accomplish real-world tasks.”
Now, Sun and UC Masters student Sam King are converting Sun’s successful simulation study into a real robot.
âThe challenge is how to transfer this learned control policy from simulation to reality, often referred to as the ‘Sim2Real’ problem,â Sun said.
Numerical simulations typically only pass 60 to 70 percent in initial real-world applications, Sun said. He plans to spend a year or more bridging the gap to perfect his new autonomous robotics system.
So there is a lot of time to invest in robot-proof door locks.
Video of a robot that can open a door: https://www.youtube.com/watch?v=O_FV47hIRus
Source of the story: