I-Corps: Translation Potential of Mobile Manipulation for Object Retrieval in Rehabilitation Hospitals

This I-Corps project focuses on the development of an assistive robotic system that enables object retrieval and manipulation through voice commands. The technology addresses challenges in healthcare, assisted living, and service environments where individuals may have limited mobility or require assistance with daily tasks. The solution combines accessibility with automation to enhance operational efficiency while promoting user independence.

When deployed in medical facilities, the system allows staff to focus on specialized care tasks rather than routine object retrieval. In residential settings, it gives users greater autonomy in their daily activities. The technology also has potential applications in industrial and commercial environments where human-robot collaboration can streamline workflows and reduce physical strain on workers.

Beyond its immediate applications in healthcare and assistance, the system's ability to reliably grasp and manipulate objects opens possibilities in manufacturing, logistics, and service industries where object handling automation can increase productivity and workplace safety.

This I-Corps project utilizes experiential learning coupled with a first-hand investigation of the industry ecosystem to assess the translation potential of the technology. This solution is based on the development of vision-guided robotic grasping systems that can pick up objects using input from a single camera. The underlying technology leverages physics simulation for data collection and learning, enabling successful transfer of grasping capabilities from simulated to real-world environments.

The system demonstrates robust performance in grasping previously unseen objects, with even higher success rates achieved using simplified robotic hands. Research results show that the technology performs particularly well when handling familiar objects, making it suitable for deployment in structured environments where the set of target objects remains relatively constant.

The technical approach combines computer vision, machine learning, and robotic control to create a reliable and adaptable object manipulation system.

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.