Abstract
Socially assistive robots (SAR) have shown great potential to augment the social and educational development of children with autism spectrum disorders (ASD). As SAR continues to substantiate itself as an effective enhancement to human intervention, researchers have sought to study its longitudinal impacts in real-world environments, including the home. Computational personalization stands out as a central computational challenge as it is necessary to enable SAR systems to adapt to each child's unique and changing needs. Toward that end, we formalized personalization as a hierarchical human robot learning framework (hHRL) consisting of five controllers (disclosure, promise, instruction, feedback, and inquiry) mediated by a meta-controller that utilized reinforcement learning to personalize instruction challenge levels and robot feedback based on each user's unique learning patterns. We instantiated and evaluated the approach in a study with 17 children with ASD, aged 3–7 years old, over month-long interventions in their homes. Our findings demonstrate that the fully autonomous SAR system was able to personalize its instruction and feedback over time to each child's proficiency. As a result, every child participant showed improvements in targeted skills and long-term retention of intervention content. Moreover, all child users were engaged for a majority of the intervention, and their families reported the SAR system to be useful and adaptable. In summary, our results show that autonomous, personalized SAR interventions are both feasible and effective in providing long-term in-home developmental support for children with diverse learning needs.
Highlights
Human development follows non-linear trajectories unique to each individual (Vygotsky, 1978)
We introduce a hierarchical framework for Human Robot Learning that decomposes Socially Assistive Robotics (SAR) interventions into computationally tractable state-action subspaces contained with a meta-controller
To address the challenge of long-term personalization in SAR in a principled way, we present a solution to the problem as a controller-based environment which we define as hierarchical human-robot learning
Summary
Human development follows non-linear trajectories unique to each individual (Vygotsky, 1978). In a long-term setting, this means interventions must continuously and rapidly adapt toward the user’s unique personality. While personalized services are paramount, they are neither universally nor equitably affordable This provides motivation for humanmachine interaction research that seeks to develop personalized assistance via socially assistive agents, whether disembodied, virtually embodied (DeVault et al, 2014), or physically embodied (Mataric, 2017). A significant body of SAR research has focused on user learning, with a specific focus on developing personalized robot tutors for young children (Clabaugh and Mataric, 2019). Several studies on intelligent tutoring systems (ITS) have involved computational models of student learning patterns; in contrast to SAR, these works have predominately focused on university students in highly controlled environments (Anderson, 1985; Murray, 1999). From that body of past work, key principles about SAR for learning have been grounded in theories of embodied cognition, situated learning, and user engagement
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
