Blog

ICRA 2020 moved to virtual conference; presentation posted below

Posted 05/23/2020

Given the ongoing social restrictions due to COVID-19, ICRA 2020 has been moved to a virtual format. I have created a presentation for my work which will be available on-demand through the ICRA proceedings and I have also posted it here on my website (available below).

Two papers accepted to ICRA 2020

Posted 02/03/2020

I am happy to announce that the research I worked on while at Honda Research Institute last summer has been accepted to ICRA 2020! This was a very cool project on teaching robots to hug people based on full-body haptic cues, conducted under the guidance of Katsu Yamane. In addition, my colleague Geoff Clark's paper was also accepted to ICRA (congratulations on the awesome work Geoff!).

For more information, please see the accompanying videos and/or consult the full papers (when available).

J. Campbell and K. Yamane. Learning Whole-Body Human-Robot Haptic Interaction in Social Contexts. International Conference on Robotics and Automation (ICRA), Paris, France, May 2020.

G. Clark, J. Campbell, S.M.R. Sorkhabadi, W. Zhang, and H. Ben Amor. Predictive Modeling of Periodic Behavior for Human-Robot Symbiotic Walking. International Conference on Robotics and Automation (ICRA), Paris, France, May 2020.

I will be presenting my work at IROS 2019

Posted 10/28/2019

I will be presenting the results of my collaboration with Prof. Koh Hosoda (Osaka University) and Prof. Shuhei Ikemoto (Kyushu Institute of Technology) at IROS 2019 in Macau, China. This work was performed during an NSF EAPSI/JSPS Summer Program fellowship at Osaka University and examines how we can leverage imitation learning to reproduce complex interactions with humans in non-tractable pneumatically-controlled robots.

For more information, please see the accompanying video and/or consult the full paper.

J. Campbell, A. Hitzmann, S. Stepputtis, S. Ikemoto, K. Hosoda, and H. Ben Amor. Learning Interactive Behaviors for Musculoskeletal Robots Using Bayesian Interaction Primitives. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Macau, China, November 2019.

I will also be attending CoRL from October 29 to November 2 in Osaka, Japan and would love to meet up with anyone else attending!

Invited talk at IROS 2019

Posted 10/25/2019

I will be giving an invited talk at the IROS 2019 Workshop on Progress in Ergonomic Physical Human-Robot Collaboration on behalf of my advisor, Prof. Heni Ben Amor.

Machine Learning and Predictive Biomechanics for Human-Robot Collaboration

Collaborative robots and other forms of modern assistive technology, e.g., smart prosthetic devices and exoskeletons, have the potential to change millions of lives for the better. However, for this vision to become reality, a theoretical foundation is needed that allows for the specification of safe and meaningful physical interactions between humans and robots. In this talk, I will discuss Bayesian Interaction Primitives (BIP) -- unified statistical framework for modeling dynamics among multiple agents using a compact, probabilistic, and data-driven methodology. BIPs can be used to derive algorithms for learning and adaptation which incorporate the future biomechanical state of a human user into the decision-making process. In turn, predicted biomechanical variables can be used to steer physical human-robot interaction towards biomechanically safe movement regimes. Finally, I will discuss applications of our approach to collaborative robotics and the control of intelligent prosthetics.

IntPrim v2.0 Released on Github!

Posted 10/20/2019

I have just released IntPrim v2.0 on our laboratory's Github page! This version is a significant upgrade from v1.0, and adds support for eBIP, automatic basis space selection, new analysis tools, and more. It can be found at the following repository:

https://github.com/ir-lab/intprim

Paper accepted to NeurIPS 2019 Workshop on Robot Learning!

Posted 10/5/2019

Our workshop paper on using imitation learning to combine language and vision has been accepted to the NeurIPS workshop on robot learning. Congratulations to the lead author, Simon Stepputtis, for his awesome, new work!

S. Stepputtis, J. Campbell, M. Phielipp, C. Baral, and H. Ben Amor. Imitation Learning of Robot Policies by Combining Language, Vision and Demonstration. NeurIPS Workshop on Robot Learning (NeurIPS-WRL), Vancouver, Canada, December 2019.

Paper accepted to AAAI Fall Symposium on Artificial Intelligence for Human-Robot Interaction!

Posted 9/1/2019

Our paper detailing a new data set collected in a user trial of hugging interactions with robots has been accepted to the AAAI Symposium on HRI. Congratulations to Kunal Bagewadi for his great work!

K. Bagewadi, J. Campbell, and H. Ben Amor. Multimodal Dataset of Human-Robot Hugging Interaction. AAAI Fall Symposium on Artificial Intelligence for Human-Robot Interaction (AI-HRI), Arlington, Virginia, November 2019.

Invited talk at RSS 2019

Posted 6/1/2019

I will be giving an invited talk at the RSS 2019 Workshop on AI and Its Alternatives for Shared Autonomy in Assistive and Collaborative Robotics.

Learning Interaction Primitives for Human-Robot Collaboration and Symbiosis

In this talk, I will present a methodology for learning physical human-robot interaction from demonstrations. The result of this learning process is a compact representation, called "Interaction Primitive", which models the spatio-temporal relationship between multiple agents. Interaction Primitives can be used in human-robot collaboration and shared control tasks for both action recognition, as well as action generation. Most importantly, they generate probabilistic beliefs over key information that is needed for safe and fast-paced physical interaction. I will present extensions of this approach that address multimodal datasets and complex, non-linear inference schemes. Finally, I will also discuss a number of real-world applications, including intelligent prosthetics, collaborative robot manipulation, as well as throwing-and-catching games.

I will be presenting my work at RSS 2019

Posted 6/1/2019

At this year's RSS in Freiburg, Germany, I will be presenting my work on ensemble Bayesian Interaction Primitives (eBIP). This is a Monte Carlo approximation of Bayesian Interaction Primitives which yield several important gains in inference accuracy, computational complexity, and ease of training.

For more information, please see the accompanying video and/or consult the full paper.

J. Campbell, S. Stepputtis, and H. Ben Amor. Probabilistic Multimodal Modeling for Human-Robot Interaction Tasks. Robotics: Science and Systems (RSS), Freiburg, Germany, June 2019.