It is difficult for all of us to imagine that in a relatively short time we can acquire a robot or a very competent artificial intelligence for our home or to accompany us in our lives, basically because something like this must have a very high price or complex maintenance and technical characteristics .
However, all this is largely solved thanks to the so-called "robotics in the cloud", that is, the migration to powerful centralized computer systems of much of the intelligence provided to our robots. Something that is finally easy to achieve thanks to the spectacular growth in data transmission speed .
As Martin Ford explains in his book The Water of Robots :
Today it is possible to leave much of the processing required by advanced robotics in the hands of large data centers and thus give robots access to an entire network of resources. This allows less expensive robots to be built as they are endowed with less power and memory, and it also allows updating the software of many machines at the same time. If a robot uses this centralized computer intelligence to learn and adapt to its environment, that learning will be instantly available to any other machine accessing the system, and it will be easy to extend it to a large number of robots.
The Google Cloud
Google has been one of the companies that is investing the most in robotics in the cloud. Already in 2011, it announced its support for this procedure by offering a platform that allows robots to take advantage of all the services designed for Android devices.
Google has also introduced the Googles service, which allows us to take a picture of an object, and the system will automatically recognize the object, sending us information about it. This makes it possible to expand the databases and offer greater visual recognition capabilities to the robots that are in the cloud. It will be a gigantic library that any robot can use .
Watson
Another great initiative for the cloud has been that of IBM and its Watson supercomputer, which can now reside on huge collections of servers connected to the Internet. In this way, developers can link directly to the system and incorporate Watson’s revolutionary cognitive technology into software programs and mobile applications.
Amazon
Amazon is also providing cloud computing services, as Ford explains:
Cycle Computing is a small company specialized in large-scale computing that was able to solve through Amazon’s cloud service, and in just 18 hours, a complex problem that would have taken a personal computer more than 260 years. The company estimates that before the advent of cloud computing it would have cost approximately $ 68 million to build a supercomputer capable of accomplishing its goal. In contrast, it is possible to rent 10,000 servers in the Amazon cloud for around $ 90 an hour.