March 11, 2013
Robots Get Their Own Cloud, Inspiring Smarter Aritificial Intelligence
Michael Harper for redOrbit.com — Your Universe Online
For all the good robots can do, they´re built with one staggering disadvantage. No matter how sophisticated, no matter how powerfully built or how smart a robot may be, they´re only capable of doing what they´ve been programmed to do. The fact that many of the tasks we ask robots to complete are extremely complex only highlights this disadvantage.
Robots can be programmed to churn through calculations necessary to handle day-to-day change, of course, but the hardware and software which makes this necessary is often expensive and bulky. The RoboEarth Cloud Engine can help keep robot costs down while ensuring that they stay as small and light as possible.
The five European scientists responsible for this “Internet for robots” are calling it “Rapyuta” and have made it available to all robotic enthusiasts today.
"The RoboEarth Cloud Engine is particularly useful for mobile robots, such as drones or autonomous cars, which require lots of computation for navigation,” explains Mohanarajah Gajamohan, researcher at the Swiss Federal Institute of Technology (ETH Zurich) and Technical Lead of the project in a statement.
“It also offers significant benefits for robot co-workers, such as factory robots working alongside humans, who require large knowledge databases, and for the deployment of robot teams," added Gajamohan.
The RoboEarth team suggests the following scenario: A robot is tasked with making breakfast in the morning. It may be programmed to fry up some bacon and drizzle honey over biscuits, but what if the honey is suddenly in a different sized bottle? The robot could become confused, not knowing what ingredient is contained in the bottle or even how to pick it up. Furthermore, if the chairs are placed in different spots each day, it could be time for lunch before the robot gets the table set for breakfast.
With Rapyuta, (a clever reference to a film where Rapyuta is a castle in the sky inhabited by robots) the morning breakfast robot could figure out what ingredient the mysterious new bottle contains by relying on computational processes stored in the cloud.
“Each robot connected to Rapyuta will have a secured computing environment giving them the ability to move their heavy computation into the cloud,” reads an explanation of the RoboEarth Cloud Engine. “In addition, the computing environments are tightly interconnected with each other and have a high bandwidth connection to the RoboEarth knowledge repository.”
With this engine built in, robots will be able to not only offload some of their work to the cloud, but they´ll also be able to pull down other solutions previously completed by other robots. For instance, a second breakfast robot who encounters the new honey bottle shape could place a call to the cloud and get the answer quickly.
The introductory video also mentions a second robot in the hypothetical household – a vacuuming robot. With Rapyuta, the breakfast robot and the vacuuming robot can share information about their house with one another, rather than sussing out an ever changing layout on their own. If the breakfast robot encounters a new chair configuration, it can share its final solution with the vacuuming robot. This means the cleaning robot doesn´t need to rely purely on an infrared scanner to prevent itself from bumping into the legs of the chair. It can instead use a map created by its robotic coworker.
The team claims more computational tasks can be moved into the cloud as time moves on. The European team is also sensitive to concerns that such a cloud could negatively impact jobs. The makers of the RoboEarth Cloud Engine say this development could in fact create more jobs than it replaces.