Las Vegas/ Science, Tech & Medicine

Pitt-Led RAMMP Program Adds Robotic Arms to Power Wheelchairs

AI Assisted Icon
Published on November 18, 2026
Pitt-Led RAMMP Program Adds Robotic Arms to Power WheelchairsSource: U.S. Navy Petty Officer 1st Class Chad J. McNeeley / Wikimedia Commons

A chair that opens a door is useful. A chair that can also carry a cup to someone’s mouth, and stop when the rider wants the last inch, is a different kind of engineering bet. RAMMP, a five-year program that began in September 2025, pairs a powered wheelchair with a robotic arm, artificial intelligence, an assistive-technology operating system and a digital-twin simulation environment.

The University of Pittsburgh’s Human Engineering Research Laboratories leads the work with funding from the Advanced Research Projects Agency for Health, an award Purdue University’s Institute for Accessible Science describes as up to $41.5 million. Assistive-technology company ATDev is the industry partner. News 3 reported Tuesday that co-founder Owen Kent, a wheelchair user, is developing the system so riders can handle more everyday tasks from a seated position, sometimes with less hands-on caregiver help.

More Than a Ride Across the Room

Purdue’s account of RAMMP is a stack, not a slogan: the chair, the arm, the software that interprets a scene, an operating system built for assistive tech, and a digital-twin simulation environment. Curb Cuts describes the result as an open-source platform and identifies ATDev as the company building wheelchair and autonomy capabilities inside the Pittsburgh-led program.

The scale of the audience is easy to blur, so the counts are worth keeping separate. A 2025 report from the Disability Rights Education and Defense Fund estimates that about 5.5 million people in the United States use a mobility device. Kent has put the power-wheelchair portion at about 3.5 million, The Robot Report noted. DREDF’s larger point is not a headcount. Beyond transport, these devices matter because they are how many users hold onto independence and autonomy.

That is the gap ATDev says it is building into, not around. In a September interview with the Center for Data Innovation, the company said its goal is to extend the capabilities of existing assistive devices rather than replace them. News 3 reported that Kent co-founded ATDev from his own experience with the limits that remain after a wheelchair has already solved getting down the hall: the reach, the door, the counter, the wait for another person’s hands.

Vision, Intent and a Dial the Rider Sets

ATDev says RAMMP uses computer vision to identify objects and activities, then an AI system to interpret what the user intends. Curb Cuts reported that the team built that perception layer with Meta’s open-source DINOv3 and SAM vision models. News 3 separately said ATDev is working with Meta’s AI for Good program, which aims to apply artificial intelligence to projects with social impact, and that the collaboration is meant to advance RAMMP’s assistive capabilities.

The control claim is more specific than “the chair drives itself.” ATDev’s materials, as relayed by Curb Cuts, put the choice on the rider: “You decide, task by task, how much the chair does.” Full help on one errand and none on the next are supposed to be settings, not a personality the software assigns. News 3 said the design is meant to expand what someone can do while seated and, for certain activities, may reduce the need for caregiver assistance.

That last part is an intention, not a published clinical result. Purdue lists a digital-twin simulation environment alongside the hardware. The five-year clock started in September 2025, but that schedule is not a ship date.

Two Stories About Who Pays

Money is where the public accounts pull apart. In that September interview, Kent said ATDev had established reimbursement pathways through Medicare and other healthcare programs for its AI-enabled robotic wheelchairs. A separate MassRobotics account says the company’s founders expect to achieve an FDA-regulated, insurance-reimbursable wheelchair within five years. Those statements do not describe the same moment, and neither one is an insurer’s coverage decision or an FDA record reproduced in the reporting.

A robotic arm and an AI stack sitting on a power chair could give a benefits review more pieces to classify, not fewer. Kent and ATDev have said they hope technical gains widen access to tools that make daily tasks easier. A hope is not a billing code.

A Coffee Run, and Arms That Came First

Pittwire described a University of Pittsburgh demonstration in which a RAMMP prototype opened doors, retrieved and carried a coffee from Starbucks, and brought the cup to a wheelchair user’s mouth. The April 23 account reads as a lab sequence, not a product launch: one outing in which the chair and the arm had to share a task instead of leaving the handoff to a person standing nearby.

RAMMP’s addition is the attempt to wrap manipulation into a powered chair with vision, intent-reading software and a shared operating system, then let the rider meter how much of each task the machine finishes.

Early in the Five-Year Clock

A September 2025 start on a five-year ARPA-H award puts RAMMP about 14 months into its funded life, with Purdue’s figure capped at $41.5 million and the commercial path still described differently depending on who is talking. The prototype’s coffee run is the image that travels. The quieter test is whether a rider can refuse help on one task, accept it on the next, and find a payer that treats that choice as medical equipment rather than a demonstration.