CARAH Aims to Transform Aging at Home with Caregiver Robotics
Imagine a robot that can learn to fetch a bottle of water, retrieve medication from
a cabinet, or pick up an object from the floor—not by being programmed by robotics
experts, but by learning directly from a family caregiver. That is the vision behind
Caregiver-Guided Assistive Robot for Aging at Home (CARAH), a new artificial intelligence
research project led by Associate Professor, Nilanjan Chakraborty, (Mechanical Engineering), Dr. C. Ramakrishnan and Dr. I.V. Ramakrishnan (Computer
Science)
Supported by the Johns Hopkins Artificial Intelligence and Technology Collaboratory for Aging Research (JH AITC), a National Institute on Aging-funded initiative that advances AI-driven technologies to improve the health, independence, and quality of life of older adults, CARAH is designed to address one of the most significant challenges facing an aging population: helping older adults remain safely and independently in their own homes while easing the growing demands placed on caregivers.
As people age, routine daily activities—retrieving household items, managing medications, or navigating the home safely—can become increasingly difficult. While assistive robots have long promised to support independent living, their adoption has been limited because today's robots are difficult to train and often fail to adapt reliably to the unique layouts and routines of individual homes.
CARAH takes a different approach.
Instead of needing to be programmed by professional roboticists, CARAH enables robots to learn incrementally from just a handful of caregiver demonstrations. Even more importantly, the robot actively communicates what it has learned, allowing caregivers to see how it understands a task and provide additional demonstrations only where they are needed. This bi-directional learning process makes the robot's behavior more transparent, improves reliability, and dramatically reduces the amount of training required.
Professor Chakraborty, the lead investigator on the project, describes how this caregiver-guided approach could make assistive robots more practical, adaptable, and trustworthy in the home:
``For assistive robots to become useful in people’s homes, families should not need a robotics expert every time the robot encounters a new task or a new environment. With CARAH, we want caregivers to be able to teach the robot directly, see what it has learned, and help it improve when necessary. Our hope is to make robots adaptable and trustworthy enough to support older adults in everyday activities while reducing some of the routine physical demands placed on caregivers and allowing them to spend more of their time on the human aspects of care. “
The project's AI software is designed to be hardware-agnostic, making it compatible with a wide range of robotic platforms. For this project, the research team will integrate the software with a mobile robotic platform equipped with a robotic arm, cameras, and environmental sensors capable of performing everyday manipulation tasks. Initial demonstrations will focus on activities such as retrieving water or medication, picking up dropped objects, and accessing items stored in cabinets or drawers—tasks that can reduce fall risks and free caregivers to spend more time interacting with loved ones rather than performing routine chores.
Researchers will evaluate CARAH through laboratory studies and in Stony Brook University's Home of the Future Lab, a fully furnished, sensor-equipped smart apartment located within the Center of Excellence in Wireless and Information Technology (CEWIT). Designed to simulate real residential environments, the facility provides an ideal setting for testing assistive technologies under realistic conditions and understanding how they can be integrated into everyday life.
The interdisciplinary project brings together expertise in robotics, artificial intelligence, machine learning, and human-computer interaction to develop assistive technologies that are not only technically capable but also practical for families and caregivers. By allowing robots to learn efficiently from non-experts while adapting to changing environments, CARAH represents a significant step toward making intelligent home assistants a reality.