December 20th, 2018
By Lynnette Reese, Editor-in-Chief, Embedded Intel Solutions
Instead of invasive body sensors, HOOBOX Robotics’ kit, powered by Intel® AI technologies, uses a 3D Intel RealSense camera mounted on the wheelchair to stream data that AI algorithms process in real time to control the chair.
Everyday, most people use artificial intelligence (AI) for tasks like photo tagging, talk-to-text, online searches, handwriting recognition for mobile check deposits, and ride-sharing apps. Other examples of AI in common use include email spam filters, credit card protection, and autopilot in commercial airlines. However, we continually see AI and machine learning contribute to cutting-edge innovations such as autonomous cars, precision medicine, and preliminary diagnoses.
AI is also helping quadriplegics control their motorized wheelchairs using facial expressions. The HOOBOX Robotics’ Wheelie 7 kit, powered by Intel® Intel artificial intelligence technologies, is an aftermarket add-on that allows the user to pick from ten simple facial expressions to control their motorized wheelchair. Facial expressions such as kissing, wrinkling the nose, smiling, and raising the eyebrows are translated into commands to go forward, turn right, turn left, and halt. The HOOBOX system accurately detects expressions and manipulates the wheelchair in real-time without the need for body sensors or machine training ahead of time.
HOOBOX web site claims that there are in excess of 60 people in the United States who are presently testing the kit, the majority of whom are people with amyotrophic lateral sclerosis, quadriplegics, or the elderly. The HOOBOX Wheelie 7 kit takes about seven minutes to install and employs a 3D Intel® RealSense™ Depth Camera SR300 to stream data to on-board AI processors. The system can process, in real-time, the AI algorithms that control the wheelchair.
Since immediate responsiveness to the user’s facial expressions is critical to the success of the system, the Wheelie 7 incorporates Intel® Core™ processors using the Intel® OpenVINO™ Toolkit to accelerate the facial recognition software’s inference activity. The fact that AI training of each individual is not necessary speaks to a universal-ness of human facial expression.
A HOOBOX video demonstrates the Wheelie translating facial expressions into commands to control a motorized wheelchair. “The mission scenario is to drive from the room A, passing through the corridor, overcoming its obstacles (that require precise turns and displacements) and getting back to the room A. In this round, the user took 18 expressions to complete the course in 2 minutes.”[i]
Dr. Paulo Pinheiro, co-founder and CEO of HOOBOX Robotics, states, “The Wheelie 7 is the first product to use facial expressions to control a wheelchair. This requires incredible precision and accuracy, and it would not be possible without Intel technology. We are helping people regain their autonomy.”[ii]
According to the National Spinal Cord Injury Statistical Center, there are over 280,000 people in the United States who have spinal cord injuries. The center also claims that there are approximately 17,700 new cases each year. Physical mobility has the most impact on the quality of life for those with spinal cord injuries.[iii]
As of this writing, The Wheelie 7 kit includes a 30-day trial and is available in a subscription plan at $300/month on a 12-month subscription plan.
Lynnette Reese is Editor-in-Chief, Embedded Intel Solutions and Embedded Systems Engineering, and has been working in various roles as an electrical engineer for over two decades. She is interested in open source software and hardware, the maker movement, and in increasing the number of women working in STEM so she has a greater chance of talking about something other than football at the water cooler.