|
The IRIM Seminar Series | March 31, 2021 | 12:15PM EDT
Towards Robust HRI: A Stochastic Optimization Approach
Stefanos Nikolaidis | Assistant Professor; Computer Science, University of Southern California
Access the Event Here : https://tinyurl.com/IRIMVSSspring4
Abstract
The growth of scale and complexity of interactions between humans and robots highlights the need for new computational methods to automatically evaluate novel algorithms and applications. Exploring the diverse scenarios of interaction between humans and robots in simulation can improve understanding of complex HRI systems and avoid potentially costly failures in real-world settings.
In this talk, I propose formulating the problem of automatic scenario generation in HRI as a quality diversity problem, where the goal is not to find a single global optimum, but a diverse range of failure scenarios that explore both environments and human actions. I show how standard quality diversity algorithms can discover interesting and diverse scenarios in the shared autonomy domain. I then propose a new quality diversity algorithm, CMA-ME, that achieves significantly better performance than the state-of-the-art in benchmark domains. Finally, I discuss applications in procedural content generation and human preference learning.
Speaker
Stefanos Nikolaidis is an Assistant Professor of Computer Science at the University of Southern California and leads the Interactive and Collaborative Autonomous Robotics Systems (ICAROS) lab. His research focuses on stochastic optimization approaches for learning and evaluation of human-robot interactions. His work leads to end-to-end solutions that enable deployed robotic systems to act optimally when interacting with people in practical, real-world applications.
Stefanos completed his PhD at Carnegie Mellon's Robotics Institute and received an MS from MIT, a MEng from the University of Tokyo and a BS from the National Technical University of Athens. His research has been recognized in the form of best paper awards and nominations from the IEEE/ACM International Conference on Human-Robot Interaction, the International Conference on Intelligent Robots and Systems and the International Symposium on Robotics
|
|
|
RESEARCH NEWS
Control System Helps Several Drones Team Up to Deliver Parcels
Many parcel delivery drones of the future are expected to handle packages weighing five pounds or less, a restriction that would allow small, standardized UAVs to handle a large percentage of the deliveries now done by ground vehicles. But will that relegate heavier packages to slower delivery by conventional trucks and vans?
A research team at the Georgia Institute of Technology has developed a modular solution for handling larger packages without the need for a complex fleet of drones of varying sizes. By allowing teams of small drones to collaboratively lift objects using an adaptive control algorithm, the strategy could allow a wide range of packages to be delivered using a combination of several standard-sized vehicles.
Beyond simplifying the drone fleet, the work could provide more robust drone operations and reduce the noise and safety concerns involved in operating large autonomous UAVs in populated areas. In addition to commercial package delivery, the system might also be used by the military to resupply small groups of soldiers in the field.
“A delivery truck could carry a dozen drones in the back, and depending on how heavy a particular package is, it might use as many as six drones to carry the package,” said Jonathan Rogers, the Lockheed Martin Associate Professor of Avionics Integration in Georgia Tech’s Daniel Guggenheim School of Aerospace Engineering. “That would allow flexibility in the weight of the packages that could be delivered and eliminate the need to build and maintain several different sizes of delivery drones.”
Read the Press Release Here
|
|
|
|
A research team at the Georgia Institute of Technology has developed a modular solution for drone delivery of larger packages without the need for a complex fleet of drones of varying sizes. By allowing teams of small drones to collaboratively lift objects using an adaptive control algorithm, the strategy could allow a wide range of packages to be delivered using a combination of several standard-sized vehicles.
|
|
Researcher & Student Accolades
Congratulations to the Robot Autonomy and Interactive Learning (RAIL) Lab Team of PI Sonia Chernova & students Devleena Das and Siddhartha Banerjee Best for their Best Paper award in the Technical Advances category at the IEEE/ACM International Conference on Human-Robot Interaction.
“Explainable AI for Robot Failures: Generating Explanations that Improve User Assistance in Fault Recovery” proposes a new type of explanation, εerr, that addresses the need to develop solutions for non-expert intervention and failure recovery that will be required if we are to adopt robotics systems broadly into domestic environments.
HRI '21: Proceedings of the 2021 ACM/IEEE International Conference on Human-Robot Interaction
March 2021, Pages 351–360 | https://doi.org/10.1145/3434073.3444657
|
|
|
Conference Presentation @ HRI 2021
Explainable AI for Robot Failures: Generating Explanations that Improve User Assistance in Fault Recovery
|
|
IRIM Faculty Lab Highlight
Georgia Tech Research Institute | Intelligent Sustainable Technologies Division
Gary McMurray Division Chief; Intelligent Sustainable Technologies Division @ GTRI, Associate Director; Institute for Robotics and Intelligent Machines (IRIM)
The Intelligent Sustainable Technologies Division (ISTD) develops innovative technology systems to enhance the productivity and competitiveness of Georgia’s food processing industry. A unique research unit of the Georgia Tech Research Institute, ISTD works collaboratively with university and industry partners on projects involving robotics, advanced sensors, environmental treatment, and food safety technologies. ISTD also conducts air quality research related to monitoring and reducing the environmental impact of vehicular emissions, designs and builds sustainable energy systems, and develops cutting edge systems to automate roadside maintenance for the Georgia Department of Transportation. Our goal is to transition technologies from concept to commercialization, as quickly and economically as possible.
Learn more about the lab here
|
|
|
|
|
Gary McMurray talks about his position as a Principal Research Engineer at GTRI.
|
|
GT Affiliate Event: Cognitive Augmentation
Pattie Maes | Professor of Arts & Sciences; MIT Media Laboratory
April 7, 2021 | 11AM EDT
Access the Lecture at: https://tinyurl.com/CHICEmaes
|
|
|
University Affiliate Event
Virtual Seminar Series on the Intersection of Control & Learning
Wednesdays 9 a.m. – 10 a.m. (Pacific Time)
|
|
|
EVPR COVID News & Guidance
Reminder: Lab Personnel Density Guidance
Georgia Tech is piloting a revised guideline to accommodate laboratories with lower personnel numbers and sufficient excess space in their laboratory. Read the revised guideline.
Weekly Testing Locations
If you live or work on campus, we strongly encourage you to get tested weekly, even if you aren’t experiencing Covid-19 symptoms. This is an essential part of protecting yourself and the Georgia Tech community. There are several options for getting tested, both on and off campus. See the current schedules and locations at this link.
Vaccine Roll-Out
The Institute has been working diligently with the Georgia Department of Public Health (DPH) to develop a vaccine rollout plan for the entire campus community. This plan consists of consecutive phases with corresponding groups. See the vaccine plane here.
|
|
|
|
|
|
