STC: Center for Human and Robot Co-Adaptation

For robots to transition from research laboratories to effective collaborators in homes, hospitals, workplaces, and public spaces, they must be able to understand, interact with, and adapt to people over time. Such advances have the potential to address national needs in independent living across the lifespan, post-acute and rehabilitative care, and a range of human-centered services. Realizing this potential will require a shift in research approaches to better understand how both human and robot behaviors, expectations, and capabilities can mutually benefit from ongoing interaction. The Center for Human and Robot Co-Adaptation addresses this gap by developing methods to understand, predict, and guide these reciprocal adaptations. These efforts will help ensure that assistive and service robots remain safe, trustworthy, and aligned with evolving human needs and values. The Center will also promote education and workforce development by providing training that integrates artificial intelligence, robotics, and human factors. Its Co-Adaptive Robotics for All Learners initiative will develop sustainable, reusable high-school robotics and AI curricula, public outreach activities, and pathways for students in this emerging STEM field. The Center’s research is structured around three interconnected threads. The first thread focuses on enabling robots to perform tasks that users need and value by developing lifelong co-learning algorithms. These algorithms will allow robots to acquire perception, manipulation, and communication skills from limited demonstrations and practice, while establishing shared mental models and communication conventions with users. The second thread investigates the foundations of human-robot co-adaptation in dyadic interactions, where a single human and a single robot adapt to each other through ongoing engagement. This research will develop frameworks for inferring individual values and preferences, estimating introspective competence, and supporting expectation-alignment methods to foster well-calibrated mutual trust. The third thread examines co-adaptation at the scale of groups and communities by studying group-aware task-planning algorithms, data-driven models of group dynamics, and methods for robots to reason about community norms. Across all threads, both cognitive and physical interactions will be studied and advanced. Physical interaction tasks necessitate advances in motor control, high-level planning, and models of human responses, with careful consideration of safety, accessibility, comfort, workload, and individual capabilities. Research activities will be grounded, tested, and refined through the Human Environment with Robots (HERO) Facility Network, an ecologically valid testing ground comprising always-on, participatory living laboratories in residential, commercial, and healthcare facilities across the partner institution network. These deployments will increase in complexity over the Center’s duration, supporting the development of generalizable principles, algorithms, and evaluation methods for long-term human-robot co-adaptation. 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.