Co-located with the IEEE International Conference on Ubiquitous Robotics 2024 (UR2024).

The workshop aims to serve as an engaging platform for leading researchers and practitioners to discuss the latest advancements in cloud and fog robotics. It will delve into the cutting-edge methodologies of distributed training and inferencing, which are critical for enhancing robot intelligence. Additionally, the workshop will explore innovative strategies for the smart offloading of cognitive processes in robotics, addressing the complexities and computational demands of current robotic systems. The objectives include creating a vibrant community by inviting active researchers to share their insights, fostering collaborative learning and idea exchange. The program will feature talks from invited speakers who are at the forefront of these fields, along with a poster session to showcase emerging research, providing attendees with a comprehensive view of the current trends and future directions in cloud and fog robotics.

Topics of interest

  • Cloud and fog robotics platforms
  • Cloud robot intelligence
  • Distributed learning and inferencing
  • Smart offloading of robotic tasks
  • Multi-robot collaboration via cloud and fog computing

Date, Time and Location

  • Monday 24 June, 2 pm – 5 pm
  • KC 907, Kimmel Center, New York University, USA

Invited Speakers

Joohyung Kim, Associate Professor, UIUC, USA

Talk Summary

PAPRAS: Plug And Play Robotic Arm System

The demand for robots capable of working in close proximity to humans and engaging in physical interaction has been steadily increasing. Currently, robots are employed in various settings such as airports, restaurants, and amusement parks to guide, assist, and entertain people. However, despite technological advancements, there remains a scarcity of robotic applications that fully meet public expectations. Enhancing the usefulness of robots in our daily life necessitates a deeper understanding of human environments and tasks, improved methods for task solution by robots, and the development of designs that enable robots to interact with humans in a natural and safe manner. In this talk, I will introduce KIMLAB’s recent endeavors in designing and implementing PAPRAS (Plug And Play Robotic Arm System) for general-purpose applications.

Short Bio

Joohyung Kim is currently an Associate Professor of ECE (Electrical and Computer Engineering) and MechSE (Mechanical Science & Engineering), and the director of KIMLAB (Kinetic Intelligent Machine LAB) at University of Illinois, Urbana-Champaign. His research focuses on design and control for humanoid robots, system for motion learning in robot hardware, and safe human-robot interaction. He received BSE and Ph.D. degrees in Electrical Engineering and Computer Science (EECS) from Seoul National University, Korea, in 2001 and 2012. He was with Disney Research as a Research Scientist from 2013 to 2019. Prior to joining Disney, he was a postdoctoral fellow in the Robotics Institute at Carnegie Mellon University for DARPA Robotics Challenge in 2013. From 2009 to 2012, he was a Research Staff Member in Samsung Advanced Institute of Technology and Samsung Electronics, Korea, developing biped walking controllers for humanoid robots.

Sandeep Chinchali, Assistant Professor, University of Texas at Austin, USA

Talk Summary

Distributed Inference, Learning, and Control Between Fleets of Robots and the Cloud

Today, we are starting to place high hopes and demands on fleets of networked robots. Ideally, they should be cheap, compute/memory/power efficient, and highly resilient and autonomous in the physical world. However, this is at odds with the way machine learning, and especially generative AI is trending today. We are building larger and larger models that are fairly accurate, but compute, power, and data hungry. In this talk, I’ll describe our work on augmenting robotic inference and learning with cloud resources using algorithms that gracefully trade-off accuracy with systems costs of inference, storage, bandwidth, labeling, and compute costs. I will highlight our papers in top robotics and ML venues like RSS, CoRL, ICML, and MLSys.

Short Bio

Sandeep Chinchali is an assistant professor in UT Austin’s ECE department. He completed his PhD in computer science at Stanford and undergrad at Caltech, where he researched at NASA JPL. Previously, he was the first principal data scientist at Uhana, a Stanford startup working on data-driven optimization of cellular networks, now acquired by VMWare. Sandeep’s research on cloud robotics, edge computing, and 5G was recognized with the Outstanding Paper Award at MLSys 2022 and was a finalist for Best Systems Paper at Robotics: Science and Systems 2019. His group is funded by companies such as Lockheed Martin, Honda, Viavi, Cisco, and Intel and actively collaborates with local Austin startups.

Youngjae Kim, Assistant Manager, LG Electronics, South Korea

Talk Summary

Cloud robotics and on-device A.I. in the generative A.I. era

As generative A.I. technology advances to the level of commercialization, on-device A.I. technology is getting high attentions from the industry. It is inherently based on low-power enabling technology and requires no network connection and, hence, no subscription fee for any A.I. service. At the same time, however, its limitations are clear. For example, it won’t be able to generate the correct information when it was asked a live sport game score. We believe that the ultimate future will be cloud-native, multi-modal, and deeply networked. On the road to that future, two technologies may be complementary in a sense that cloud robotics technology help the development of future on-device applications and low power technology developed for on-device A.I. can be used to save power in the cloud servers. LG electronics is also working on both. In this talk, I will introduce what LG electronics is doing under this subject.

Short Bio

Youngjae Kim is a communication theory backed robotics engineer. He currently works for the Advanced Robotics Lab. at LG Electronics Inc. as a Senior Research Fellow (VP). Previously, he worked for Velodyne LiDAR Inc. and Apple Inc. on 3D point cloud processing and cellular modem design, respectively. His research interests include cloud robotics applications and household work automation using mobile manipulators. He holds M.S. and Ph.D. degrees both from Stanford University and a B.S. degree from Seoul National University, all in Electrical Engineering.

Program (tentative)

  • 2:00 pm – 2:05 pm: Opening Remarks
  • 2:05 pm – 2:50 pm: Invited Talk by Joohyung Kim
  • 2:50 pm – 3:00 pm: Break
  • 3:00 pm – 3:50 pm: Invited Talk by Sandeep Chinchali
  • 3:50 pm – 4:40 pm: Invited Talk by Youngjae Kim
  • 4:40 pm – 4:50 pm: Short paper session: “LLM-based Nursing Diagnostic AI with Human-System Interaction For Effective Training and Nursing Home Support” by Myungeun Lee, The George Washington University, USA
  • 4:50 pm – 5:00 pm: Closing Remarks

Organizers

  • Minsu Jang, Electronics and Telecommunications Research Institute, South Korea
  • Youngjae Kim, LG Electronics, South Korea