Intelligent ADL Assistance Through Multimodal Shared Control of a Wheelchair-Mounted Assistive Robot to Support Daily Independence and Community Living - This three-year DRRP will develop and evaluate an AI-enabled Multimodal Shared-Control System (AIMSCS) for wheelchair-mounted robotic arms (WMRAs) to increase independent, timely completion of user-prioritized daily activities by powered-wheelchair users (PWUs) with severe upper-limb limitations, reduce delayed assistance, and support household and community participation. Prior NIDILRR-supported work produced the multifunctional Robotic Assistive Arm (mR2A), a WMRA platform, with integrated control and safety features (90DPGE0018). User testing, stakeholder engagement, knowledge translation (90DPKT0015), and customer discovery identified joystick mode switching, repeated commands, fine positioning, and waiting for caregiver help as barriers. AIMSCS will enable users to request, confirm, supervise, pause, cancel, or override approved activities through speech, touchscreen, wheelchair controls, switches, scanning, eye gaze, or a support person. Requiring only two or three selections per request, it is designed to remain accessible across users and changing abilities and resolve recurring “ask-points” that delay participation. Aim 1 will finalize user, caregiver, clinical, assistive-technology, safety, home-use, and service-delivery requirements through co-design. Aim 2A will develop, integrate, and verify AIMSCS on the mR2A across unfamiliar objects, lighting, clutter, occlusion, speech variability, sensor failures, and task interruptions. Aim 2B will conduct a randomized, counterbalanced simulated-home comparison of AIMSCS shared control and manual joystick control in 28 PWUs, targeting 24 completers. Aim 3 will conduct a four-week home evaluation with up to 24 PWUs, targeting 20 completers and at least 15 caregiver/support-person dyads. Prespecified safety and performance criteria at Months 9, 15, 22, and 25 will govern advancement and refinement; a Month 36 readiness review will establish the approved configuration and future Proof-of-Adoption requirements. Primary outcomes are independent task completion in Aim 2B and sustained home use in Aim 3, with most attempts completed without hands-on help. Secondary outcomes include task time, workload, usability, satisfaction, occupational performance, psychosocial impact, assistance and waiting, resolved ask-points, caregiver intervention, reliability, and participation. Device logs and blinded assessments will be integrated with participant and caregiver interviews to identify functions that perform well technically but poorly in daily use. Final products will include a version-controlled AIMSCS-enabled mR2A; accessible user and caregiver interfaces; independent safety software; verification, home-deployment, and training protocols; cost, service-delivery, dissemination, utilization, and technology-transfer materials; and a framework linking ask-points, system capabilities, and participation outcomes. The project will determine whether AIMSCS is safe, accessible, reliable, and usable and for which users, access profiles, activities, and environments, generating evidence for future Proof-of-Adoption evaluation. Its configurable design can extend to other WMRAs and users not served by current controls.
