Some may remember Johnny 5, the loveable robot from the 1980s film Short Circuit, whose never-ending desire for “input” led him on a series of comical hijinks. While onscreen this is mere science fiction, researchers at Drexel University are working towards making Johnny 5 science fact through cloud robotics. To further research on cloud robotics, Drexel hosted a workshop from February 27 to February 28 with the goal of bringing experts from industry, government and university together to explore new opportunities in this emerging area.
Cloud robotics refers to approaches to robotics that exploit advances in cloud computing and big data, and have the potential for developing a new generation of robotics with many applications. As computing technology increases in its ability to process big data, robotics researchers are looking for novel ways to harness the power of cloud computing and pervasive internet connectivity to improve robotic autonomy.
Currently, robots are limited in their abilities due to processor constraints like speed and memory and the availability of data. By harnessing the power of cloud computing, future robots may have the ability to offload computer-intensive tasks like image processing and voice recognition and even download new skills instantly. It’s a technology that would allow researchers to develop, lighter, cheaper, more efficient “smart robots” that can adapt to environments through instant information via the internet.
The workshop was hosted by M. Ani Hsieh, an assistant professor with the Mechanical Engineering and Mechanics and head of the Drexel robotics program, along with Volkan Isler, an associate professor at the University of Minnesota. It was funded by the National Science Foundation and the Army Research Laboratory.
Hsieh hopes that the take away from this workshop will be new ideas on how to advance cloud robotics. The two-day workshop featured several technical presentations and breakout sessions to explore the opportunities in Cloud Robotics, and the science and technology gaps that need to be addressed in this field.
Among the topics discussed in the workshop were:
- Cloud-based methods for solving fundamental robotics problems such as object manipulation, navigation and learning by demonstration
- Standards for facilitating synergies between robots and the cloud, and interoperability across platforms
- Novel methods and techniques to manipulate real-time big data
- Cloud robotics data-sets and testbeds; Benchmarks and real-world robotics testbeds which can provide context for basic research and enable technology transition
- Novel cloud services based on robotics and embedded sensing such as semantic labeling of spatial maps
- Time-scale and communication issues for cloud-based solutions to robotics applications
The term “Cloud Robotics” was first coined by James Kuffner, faculty member at the Robotics Institute at Carnegie Mellon University and researcher at Google, in 2010. However, the idea for a cloud based robot isn’t necessarily new. The idea was first developed by Masayuki Inaba at the University of Tokyo who envisioned the concept of a “remote brain”. But as the internet’s ability to handle massive loads has increased, researchers are now able to exploit this boom in connectivity for robots.
Hsieh and the other conference attendees believe that a new era of robotics is just around the corner, however, there are still numerous challenges that researchers must overcome, particularly the sometimes slow nature of cloud-based applications. Either way, the time may come when Johnny 5 will find all of the input he will ever need right in the cloud.
Brian Nicholas
Public Relations and Recruitment Coordinator
Department of Mechanical Engineering and Mechanics