Carnegie Mellon University's School of Computer Science aims to change the world — the planet and its people — for the better by the year 2040. A team of Human-Computer Interaction Institute researchers has joined with colleagues at CMU and other organizations to do just that by improving how assistive technologies like prosthetic hands are fabricated and distributed.
A professionally made prosthetic hand can cost between $6,000 and $10,000. But a global network of volunteer makers called e-NABLE uses relatively affordable 3D printing technology to produce simple mechanical prosthetic hands for underserved populations around the world. For free. Since its founding in 2013, the grassroots organization has responded to more than 1,500 requests for prosthetic hands and connected those recipients, mostly children, with volunteers in the maker community who fabricate the hand on a 3D printer. Yet despite the vast number of makers around the world who contribute their time and effort to the project, e-NABLE cannot meet the demand for its prosthetics.
Enter Carnegie Mellon.
In their project, "Revolutionizing Assistive Devices Via Distributed Innovation, Fabrication and Refinement," HCII professors and lead project investigators Jennifer Mankoff and Scott Hudson have teamed up with e-NABLE founder and Rochester Institute of Technology Research Scientist Jon Schull to find ways to make customization, production and distribution of these types of do-it-yourself assistive technologies (DIY-AT) more streamlined and effective. They hope that one day, all people who need such devices will be smoothly matched with the most effective solution for their needs.
The research will focus on three specific aspects of DIY-AT: developing algorithms that improve how makers create computationally customizable physical devices; creating virtual service teams that allow multiple stakeholders to work together smoothly across ability levels and geographic distance to solve bigger, more complex problems; and studying the specific needs of clinical expert participation in assistive technology creation and provision, with the hope of revolutionizing the process.
Making Modeling Better
While the team's goals might seem lofty, they're already well on their way to changing how the world creates and distributes assistive devices. "We are trying to change the way 3D modeling happens so we can make it easier for people who are not highly trained while still ensuring high quality results," Mankoff said. "We are trying to understand how to make 3D modeling a process that can more easily be driven by requests, that supports testability and quality control, and that is modular."
In work that will be featured at this year’s ACM Conference on Human Factors in Computing Systems (CHI2016), the researchers explore the needs that current prosthetics fail to meet and how DIY technology can fill them.
"One of the things that's starting to show up in e-NABLE and in our case studies is that people don't just want hands," Mankoff said. "They want end effectors — tools that attach to the hand and can do all sorts of things. We have people who want to hold a cello bow, and someone who wants an attachment that allows him to move the handles on his hand-controlled bike.” Because these end effectors have to connect to real world objects, shaping and printing them to meet the attachment needs for those objects becomes incredibly important.
One way they're improving how devices are printed is Project Encore, which was featured at the 2015 ACM User Interface Software and Technology Symposium (UIST) this past November. Historically, 3D printing has been limited to printing items from scratch — not enhancing existing objects. Encore changes that by giving 3D fabricators the ability to print over an object, attach an item to an object or print through the holes in an object. Its software allows users to import a 3D item and click on where they want to add an attachment, such as a handle. The system then performs a series of geometric calculations to provide the information for fabricating the attachment and the printer makes it happen.
Building Effective Teams
But the DIY-AT community faces challenges beyond device fabrication. e-NABLE's process relies on a diverse group of makers spread across the globe who have a range of skills and time to devote to the project. To make their work more scalable — to help connect more applicants with the devices they need or to expand the devices beyond hands — the Carnegie Mellon researchers will create virtual service teams that apply research in machine intelligence, crowdsourcing and collaboration to solve complex problems like those inherent in DIY-AT. Herbert A. Simon Professor of HCII Robert Kraut will lead this activity, applying his extensive research on how open-source communities use information to improve processes to the maker community.
"The goal is to improve the efficiency of building a customer prosthetic device for a particular recipient by allowing multiple people to contribute different skills in measuring, building and fitting the device, while still presenting a single point of contact for recipients and their caregivers," Kraut said.
HCII Associate Professor Jeff Bigham's work in organizing crowdworkers and uniting them to accomplish shared goals will also be applied to the DIY-AT community. The researchers hope that the result of this effort will be consistent results and experiences for users, regardless of who or what the DIY-AT team comprises.
"Community members have a variety of prior experiences and varying amounts of time to help. That makes providing always-available personalized support difficult for clients," Bigham said. "We are adapting crowdsourcing methods to allow the dynamic community that happens to be available to appear more like an always-available contact point."
Revolutionizing the Process
The team's work will have benefits beyond the realm of assistive technology. "We believe that the proposed research can dramatically reshape how custom devices are produced — that it has the potential to produce a new, computationally enhanced way of working that can be widely applied," the team wrote in their proposal.
But they're still determined to improve the assistive technology process by adding better follow-up and learning more about how people use their devices. "We're trying to design a method for tracking what happens to these hands," said Mankoff. "There's very little data of that sort for any assistive technologies.”
With collaborators Jon Pearlman and Mary Goldberg from the University of Pittsburgh’s Department of Rehabilitation Science and Technology, the researchers are developing a survey to send to users about how they use their devices. They'll augment that survey with sensors on the hands that will allow them to see how they're actually being used, and whether they can predict device abandonment —which happens an unfortunate 30 percent or more of the time, even with professionally produced assistive technology. The CMU team and their University of Pittsburgh colleagues also plan to introduce new mechanisms for patient follow-up, and increase collaboration between clinicians and the virtual service teams.
In the end, the team hopes that more people will receive the prostheses they need more quickly and efficiently, with increased follow-up. More broadly, they want to revolutionize how prostheses in general — not just hands — are created and distributed.
"This project is trying to create a capability to create real-world objects that serve real-world needs without requiring the level of expertise that's currently required — making it easier to make things that matter and that particularly help people with disabilities," Mankoff said. "For now it's prosthetics, but I hope that eventually we can move beyond that."
For more on their research, visit the Make Abilities website.