Innovative technologies take aim at functional disabilities
A number of exciting research projects aimed at increasing independence for persons with disabilities have resulted from a collaboration between Electrical and Computer Engineering Professors Harry Tyrer and Gui DeSouza, Darren Gabbert with MU’s Adaptive Computing Technology (ACT) Center, and students in DeSouza’s Vision-Guided and Intelligent Robotics (ViGIR) lab.
In his position at ACT, Gabbert – who was born with a progressive neuromuscular disease – works to find innovative solutions for those with mobility, visual and learning disabilities, a job he finds “challenging, fun and sometimes baffling.”
Acutely aware of the gap that exists in areas of assistive technology for those with severe disabilities, Gabbert was introduced by Tyrer to DeSouza as someone with human-computer interface work that might help with the “baffling” part of Gabbert’s job. The pair agreed to research assistive technology and Gabbert donated a Permobil power wheelchair, which DeSouza’s students equipped with a computer and sensors.
The “smart” wheelchair, dubbed Permobot by Gabbert – a name adopted by DeSouza’s graduate students Daniel Conrad, Ruizhi Hong, and Randy Melloy – is able to interpret images detected by cameras for path planning and obstacle avoidance.
“Darren is the one who sparked this idea,” said DeSouza. “He wanted to be more independent and we hope we can give this to him.”
DeSouza explained that an electromyography (EMG) switch, used by Gabbert to control his wheelchair and smartphone through a sensor on a single functioning pectoralis muscle, is a workable human interface for severely disabled persons, but progress in assistive technology has been slow.
“It takes months for the user to switch, for example, from a joystick to EMG interface,” said DeSouza. “So now, we are interested in research that will allow users to switch seamlessly from one mode to another – pointing, nodding, facial expression, eye gaze control and other modes. And we are interested in switching not only how a human interfaces with the system, but also what is being controlled by that person, for example, a computer, a cell phone, wheelchair navigation, etc.”
Gabbert, DeSouza and Tyrer created the Robotic Assistive Technology (RAT) group a couple of years ago and since then, DeSouza has been recruiting students to join the RATs. In 2009, Gabbert met with DeSouza’s team of students to take a look at progress on the Permabot. “I immediately recognized in these young engineers that same zeal of using their technical knowledge and skills to make a difference in someone’s life,” said Gabbert. “Other smaller projects have been the outcome of conversations amongst our RAT team.”
Each semester DeSouza assigns projects to students, some as junior-year ECE projects, some as capstones projects and others as graduate level class projects or undergraduate research. Inspired by Gabbert, the students have responded favorably to the work.
“Knowing the product of my work will impact someone’s life, is a reward in itself,” said RAT team member Joe Scaduto.
RAT projects included:
A power wheelchair control interface, investigated as a capstone project by Mahmood Sobahy, Truc Bui, Sarah Danner and Joe Ayala, allows a wheelchair to be operated by humming noises. Students programmed a microcontroller to analyze and recognize four frequencies associated with forward, backward, right and left movements of the power wheelchair.
Ahmad Alshbly, Mohammed Abu Haleegah, Tony Hoff and Jeff Piersol equipped a wired EMG switch with a wireless Bluetooth interface, allowing greater freedom for the operator to switch back and forth between functions: wheelchair, cell phone and computer operation.
A graphical interface for a smart phone was developed by Scaduto, which allows the user to control the operation of the phone via the EMG switch. The switch, connected to the COM port via the Bluetooth interface, triggers horizontal and vertical bars to pan the display creating cross hairs for target selection.
As a project for his Machine Learning and Pattern Recognition class with Dr. DeSouza, Luis Alberto Rivera worked on pattern recognition algorithms to detect muscle activity, classifying EMG signals and allowing for multiple controls to be selected – as opposed to a simple “on/off” EMG switch.
Gabbert and his colleagues in the Division of IT put together a poster of the RAT team’s various research projects and took it to an EDUCAUSE conference – a nonprofit association whose mission is to advance higher education by promoting the intelligent use of information technology – where Gabbert said it was “well received.”