C2RO is making robots smarter using the cloud
Robot-makers have a challenge, said Riccardo Badalone.
They want to take advantage of recent developments in artificial intelligence and make their robots smarter.
They could do that by re-engineering their robots and building more memory and more computing power into them. But that’s expensive and memory-intensive applications require more battery power, which could make mobile robots less useful, since they would have to charge more frequently.
Badalone is the CEO of C2RO, a company that says it has a solution: It wants to put robot brains in the cloud.
The company has developed a system to process sensor data from robots in the cloud in real time, said Soodeh Farokhi, the company’s CTO. It then sends control commands back to the robot.
That means the robot doesn’t need to have complicated computer systems on board, she said.
There are other advantages that come with the flexibility of cloud computing, Badalone said.
If robot operators want to add memory to their robots, they can do that remotely, without the need to physically upgrade their robots. If they’re not using as much memory, even temporarily, they can reduce it just as easily, he said.
“You can increase the capabilities of the robot instantly and with 100 per cent flexibility,” he said.
That extra processing power is important for robots that are used indoors. (Robots that do specific job indoors — like robots that clean or transport things — as well as those that interact with people are the company’s main focus.)
Those robots struggle to locate themselves, Farokhi said, because there’s no GPS for inside buildings.
“The way that we do it, we use visual SLAM, simultaneous localization and mapping,” she said.
It’s a technique where robots use on-board cameras to create “a 3D map of their environment indoors, and they can autonomously navigate themselves to understand the environment and go from A to B.”
But that requires significant computational power.
Connected robots could also act based on data from sensors and cameras that are off-board as well as to collaborate.
“When we have the map of one robot in the cloud, the other robots can use that map,” Farokhi said.
For robots that work together — to clean a large area, for example — that can be advantageous.
That’s not the only reason C2RO wants to put that capability in the cloud, though.
Connected robots could also be given new capabilities through artificial intelligence modules — pieces of software that give the robots specific artificial intelligencebased skills.
A module in the program that processes data gathered from a robot’s sensors could generate more usable information from that data, for example.
“That’s why it’s so powerful to have it in the cloud, because we can start combining different artificial intelligence modules to give the robot more capability,” Badalone said.
C2RO is currently developing several modules in-house and, in the future, it plans to let other companies and researchers develop their own modules for robots running on its platform.
C2RO founder and chief technology officer Soodeh Farokhi, with CEO Ricardo Badalone. The company seeks to take advantage of recent developments in AI to make their robots smarter.