Get technical briefs, research updates, and collaboration opportunities.
MEET RYAN
A research and education platform for Physical AI and Healthcare Human-Robot Interaction
Unlike traditional service robots designed for mechanical or repetitive physical tasks, Wellness Robots represent a convergence of Artificial Intelligence (AI), speech and language processing, and facial recognition. These humanoid systems are engineered for Empathetic Engagement, using natural voice intonations and facial expressions to interact with humans in a manner that simulates personality and expressiveness.
The convergence of speech, language, and facial recognition technologies along with artificial intelligence has made it possible for robots to interact naturally with people — giving them something they never had before: an engaging personality.
RYAN employs multi-modal AI to dynamically adapt conversation strategy, activity selection, and emotional tone based on real-time resident response. Context-aware dialogue management enables sustained, coherent interactions over extended sessions — a requirement for therapeutic and cognitive monitoring applications.
RYAN serves as a robust platform for investigating emotion recognition and affective feedback loops. Its humanoid face and vocal intonations allow researchers to benchmark multimodal sentiment models against live human interaction data, providing a controllable physical stimulus for studying machine-generated affect.
Advanced speech recognition and language generation enables RYAN to conduct natural, fluent conversations across multiple languages. This multilingual capability is of particular research interest in diverse care populations and cross-cultural HRI studies, removing linguistic barriers common in prior social robot deployments.
As a physical deployment testbed, RYAN moves Large Language Model (LLM) research beyond text-only benchmarks into real-world, embodied contexts. Researchers can examine variables like turn-taking latency and grounded reference during extended, multi-turn dialogues in high-stakes environments like senior care.
The following table maps RYAN's core capabilities to established evidence streams in social robotics, gerontechnology, and clinical care research — situating this technology within the broader scientific literature.
Initial feasibility study demonstrating the robot's ability to engage with elderly individuals, resulting in lower stress levels among respondents.
Initial feasibility study demonstrating the robot's ability to engage with elderly individuals, resulting in lower stress levels among respondents.
Investigating the effectiveness of the robot's artificial emotional intelligence comparing empathetic vs. non-empathetic versions. Found significant improvement in the mood state of users with both versions.
Investigating the effectiveness of the robot's artificial emotional intelligence comparing empathetic vs. non-empathetic versions. Found significant improvement in the mood state of users with both versions.
Details of a clinical trial designed to measure the impact of the Ryan robot on quality of life, AD symptoms, and caregiver burden over an 8- to 10-week interaction period.
Details of a clinical trial designed to measure the impact of the Ryan robot on quality of life, AD symptoms, and caregiver burden over an 8- to 10-week interaction period.
20 participants spent 887 hours with Ryan, playing 12 different cognitive games. Average SLUMS scores improved by 3.3 points and PHQ-9 depression scores improved by 2.43 points.
20 participants spent 887 hours with Ryan, playing 12 different cognitive games. Average SLUMS scores improved by 3.3 points and PHQ-9 depression scores improved by 2.43 points.
Six children with ASD and six typically developing children tested facial expression recognition using Ryan. Increasing expression intensity significantly improved recognition accuracy.
Six children with ASD and six typically developing children tested facial expression recognition using Ryan. Increasing expression intensity significantly improved recognition accuracy.
