CAREER: Constraint-Aware Estimation, Learning and Control for Fluid Physical Human-Robot Collaboration

About This Grant

This Faculty Early Career Development (CAREER) award will support research that seeks to enable robots to work more safely and effectively alongside humans by developing a new constraint-aware framework that helps them understand and adapt to real-world challenges. While robots have made great strides in recent years, they still struggle with fluid physical collaboration in busy, unpredictable environments. The key issue is that today’s robots lack awareness of the physical, perceptual, and cognitive limitations of both themselves and the humans they interact with. Without this awareness, they cannot adapt effectively to changing tasks, human movements, or environmental conditions. This research project looks to bridge that gap by developing new algorithms and control methods that allow robots to better predict human intent, adjust their behavior in real time, and ensure safe and stable interactions. The goal is to make robots more useful in physically demanding and safety-critical tasks across industries such as healthcare, manufacturing, logistics, construction, and disaster response. If successful, these advances will help prevent workplace injuries by assisting with heavy lifting, reducing strain from repetitive tasks, and improving overall efficiency. Beyond research, this project also includes educational and outreach programs looking to train the next generation of engineers and expand STEM opportunities. By making robots smarter, safer, and more adaptable, the research seeks to bring us closer to a future where humans and robots can work together seamlessly in real-world settings. The goal of research supported by this CAREER Award is to enhance robots’ ability to predict human intent, adapt their behavior, and maintain stable physical interactions in dynamic, safety-critical tasks that neither robots nor humans can accomplish alone. To achieve these advanced perceptual and physical capabilities, the research focuses on three core thrusts. (1) Constraint-aware state and intent estimation will look to develop methods that account for perceptual and physical limitations in human-robot interaction. These methods should enable robust human state tracking despite imperfect perception, quantify confidence in predictions, and estimate feasible intent that respects shared constraints. (2) Constraint-aware adaptive and interactive learning seeks to move beyond simple imitation to create adaptation mechanisms inspired by human collaboration. By leveraging dynamical systems theory, differential geometry, and convex optimization, this thrust will seek to develop task representations that ensure stability, geometric compatibility, and adherence to physical constraints. (3) Fluid constraint-aware interaction control looks to combine confidence estimates from estimation and learning with both analytical and learning-based physical constraints to enable safe, efficient, and stable interactions. Achieving this task will involve investigating adaptive and energy-aware control techniques. Rooted in the principle that constraint-awareness enables fluid collaboration, this project looks to tackle the critical challenge of balancing safety, efficiency, and adaptability in robots designed to work seamlessly alongside humans. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.