Cloud Robotics in the Real World
Cloud Robotics is primarily used for tasks that do not require real-time execution. This helps to maintain local resources for services that are more demanding and allows cloud resources to be used for those tasks that fewer constraints on time.
When it comes to learning new information, Cloud Robotics dramatically improves the speed at which machines learn. By connecting to the Cloud, robots do not need to learn new detail individually. If one robot is aware of the information, it can be shared with others through a simple download.
This type of information is useful in autonomous vehicles where Cloud Robots update road and traffic conditions on GPS maps. For cloud robotics to continue growing in use and popularity, three different hurdles need to be overcome.
With regards to speed, the current speed of information flow that we have come to enjoy online with emails and apps is simply not enough for cloud robotics. For these machines to function at their maximum capacity, they need access to speeds that mimic brain synapses, and while we’re getting closer, we still have a ways to go.
In terms of isolation, there are instances where a robot cannot be connected to a network. While for some, this might evoke notions of the Cylon and BSG, there are other real-world situations where limited access is appropriate. For example, simple security reasons or even location can impact access. In these circumstances, while the Cloud might not be viable, edge devices might suffice to provide the same capabilities.
Finally, connectivity itself. The internet as a whole has become dramatically more stable; however, in situations where a surgical robot is dependent on a connection to the Cloud, any interruption could be life-impacting. The advent of 5G should help alleviate some of these issues with its improved latency, but another solution might also be the creation of mini-clouds that function as a fallback.
Cloud Robotics in the World of AI
Machine learning is one area where cloud robotics has an impact. If a single robot takes 100 hours to learn a task by itself, a robot that is part of a much larger team can learn quite a bit faster.
Access to Cloud services lets robots quickly learn and identify items within their environment, which can assist with tasks like cleaning and sorting. In addition, by using the Cloud for computational power, robots themselves can be designed and created with less internal computing power. This saves on energy, which makes robots lighter and a lot less expensive.