Cloud Robotics

Cloud robotics is an emerging field of robotics rooted in cloud computing, cloud storage, and other Internet technologies centered on the benefits of converged infrastructure and shared services. It allows robots to benefit from the powerful computational, storage, and communications resources of modern data centers. Other definitions stress the links between robotics and related emerging fields such as the Internet of Things, Web of Things, robot app stores, sensor networks, Big data, and others.

Erwin Mochtar WijayaErwin Mochtar Wijaya a student of Monash University explains the challenges of cloud robotics. According to Erwin “The concept of cloud robotics is similar to cloud computing where shared computing resources are provided by the cloud to private devices such as personal computers/laptops based on demands.

Similarly, in cloud robotics, computing resources can be dynamically allocated to multiple robots with a single CPU acting as the “cloud”. The concept of cloud robotics allows some limitations which exist in network robotics to be eliminated.

Picture: Erwin Mochtar Wijaya

To illustrate, in network robotics, where each robots communicate with each other by adopting a certain protocol such as ad-hoc routing, the system performance depends on stability of the network topology. This is not practical as a multi-robot system is commonly designed to perform tasks such search and rescue, exploration and surveillance where the network topology is always dynamic.

Another limitation in network robotics is the computational tasks must be performed by the on-board computers attached to every single robot. The on-board CPU has clearly limited processing power and memory spaces due to its size and power constraints. Cloud robotics offers the solution to both of these problems.Cloud Robotics

                                                                                                                                                Picture: The concept of cloud robotics

The first problem mentioned above can be overcome by adopting a clone-based model where every robot has a “clone” in the cloud which can communicate to one another. In this case, as long as the communication link between the robot and its clone is reliable, robots can cooperate without any problem as they communicate through their “clones” which exist in the cloud. The solution to the later problem is much more obvious where robots can offload high computational tasks to the cloud given the tasks do not have hard real-time requirements.



Wireless Sensor and Robot Networks Laboratory (WSRNLab)
Monash University