Project Grant 2535891
- The National Science Foundation Division of Electrical, Communications and Cyber Systems awarded the University of Texas at Austin $320,000 on October 1, 2026, under the Engineering program (CFDA 47.041) to develop robotic systems that adapt to unexpected failures without requiring retraining. The project, titled "Collaborative Research: Learning In-Context Improvement for Robotic Systems via Training-Time Failure Simulation," addresses the brittleness of current robot controllers by...
- The National Science Foundation Division of Civil, Mechanical, and Manufacturing Innovation awarded the University of Washington $187,587 on January 1, 2027, under the Engineering program (CFDA 47.041) to support fundamental research on human-inspired cyber-physical security in human-robot interactions. The research develops models of human decision-making in adversarial contexts, methods for identifying robot vulnerabilities to physical attack, and control strategies that allow robots to resist...
- The National Science Foundation Division of Civil, Mechanical, and Manufacturing Innovation awarded the University of Washington $450,064 on June 15, 2026, under the Engineering program (CFDA 47.041) to develop decentralized control strategies enabling teams of robots to explore and inspect constrained spaces such as aircraft wings and ship compartments. The research will design time-discounted ergodic controllers on spatially discretized region graphs to coordinate multiple robots in...
- The National Science Foundation Division of Computing and Communication Foundations awarded the University of Washington $600,000 on July 15, 2026, under the Computer and Information Science and Engineering program (CFDA 47.070) to develop artificial intelligence systems that adapt to new experiences and users by learning from implicit human feedback during deployment. The research addresses the problem that AI models trained on fixed datasets often fail on new data outside their training...
- This Project Grant award of $489,214 from the National Science Foundation's (NSF) Engineering program (CFDA 47.041) will fund research at Oregon State University to develop techniques to help robots recover from grasping failures in various settings, including homes and underwater environments. The goal is to improve the grasping success rate and reduce the number of attempts required, benefiting multiple sectors of the U.S. economy, from fisheries to household consumer goods. Key focus areas...
- The National Science Foundation Division of Electrical, Communications and Cyber Systems awarded Washington State University $225,000 on July 1, 2026, under the Engineering program (CFDA 47.041) to develop reinforcement learning methods with high-probability safety constraints for cyber-physical systems. The project will design learning algorithms that quantify uncertainty, constrain exploration, and adapt policies cautiously in response to environmental shifts, enabling safety-critical...
- This $349,000 National Science Foundation project grant supports research at Rice University to develop new robotics capabilities under uncertainty. The NSF Division of Civil, Mechanical, and Manufacturing Innovation is funding this three-year award through its Engineering program (CFDA 47.041). Specifically, the researchers will establish generic frameworks for robot manipulation systems to self-identify using exploratory motions while maintaining stability. By changing the traditional...
- The National Science Foundation Division of Electrical, Communications and Cyber Systems awarded the University of Texas at Austin $360,000 on October 1, 2026, under the Engineering program (CFDA 47.041) to develop robot safety methods using deep generative modeling. The project replaces hand-crafted safety constraints with learned representations of safe behavior by aggregating expert demonstrations across a wide variety of tasks and operating conditions. The research produces a reusable safety...
- The National Science Foundation (NSF) Engineering Directorate awarded a $199,683 Project Grant to Middle Tennessee State University to investigate effective engineering educational methods for preparing future robotics engineers. The grant, funded under the NSF Engineering (CFDA 47.041) program, aims to leverage industry partner needs to formulate requirements for robotics engineers and develop immersive learning environments emphasizing human-robot interactions. The research will use embodied...
- The National Science Foundation Division of Electrical, Communications and Cyber Systems awarded the University of California Irvine $605,580 on October 1, 2026, under the Engineering assistance program (CFDA 47.041) to develop modular shared robot learning systems that enable non-expert users to teach robots new tasks with reduced demonstration burden. The project combines shared learning—reusing knowledge gained from teaching one robot across different environments, tasks, hardware, and user...
The National Science Foundation Division of Electrical, Communications and Cyber Systems awarded the University of Washington $360,000 on October 1, 2026, under the Engineering program (CFDA 47.041) to develop robotic systems that adapt rapidly to failures and environmental changes without full retraining. The research addresses the brittleness of current learned robot controllers—a robot arm tuned to assemble one product cannot adjust when parts change shape, and a legged robot reliable on pavement stumbles on gravel or after motor wear. The project develops methods that expose robots deliberately to failures during training, enabling them to learn a general skill of error recovery and apply it to novel problems never encountered before. The approach mirrors human learning: when a person fumbles with an unfamiliar tool, they do not relearn hand movements but notice what went wrong and adjust on the next attempt. The resulting controllers promise more reliable robots for manufacturing, delivery, surgical assistance, and other applications where machines must adapt to new tools, changing terrain, and gradual wear without constant human re-engineering. The project runs through September 30, 2029, with work performed in Seattle, Washington. Educational components include training undergraduates and summer research students directly in the research, developing new course modules, and conducting outreach to high school students in robotics and machine learning.
Mod # | Description | ReasonForModification | Federal Obligation | Date |
|---|---|---|---|---|
| Not listed | $360.0k | 8/11/26 |