Robotic arm being made with neuromorphic engineering

Intel and Accenture announced support on Wednesday for neuromorphic research into the development of a robotic arm to assist patients with spinal injuries in performing basic tasks like drinking from a cup.  

Intel’s neuromorphic technology along with funding and technology from Accenture, as well as algorithmic support from Applied Brain Research, will be used by Israeli research teams at the Open University in the robotic arm development. 

The devices will be tested with children at ALYN Hospital, an adolescent rehab center.

Neuromorphic computing has low power requirements and offers the ability to learn and adapt to novel situations in real time, according to Intel scientists. Mike Davies, director of the Intel lab, said the technology will be a “natural fit” for an assistive technology like the robotic arm.

More than 75 million people worldwide require a wheelchair to get around and those with spinal cord or neuromuscular injuries causing arm impairments could take advantage of a robotic arm to reduce the need for a personal caregiver by offering assistance in feeding and other tasks.

Existing assistive devices are expensive, but Intel argued that its Loihi neuromorphic research chip could help reduce the costs.  Loihi allows real-time learning for adaptive control of gestures.  It also allows for greater energy efficiency, meaning less frequent charging for users.  Using affordable parts and Loihi, the cost of the robotic arm could be 10% of existing devices, Intel said in a blog.

The researchers will use ABR’s recurrent error-driven adaptive control hierarchy to help move the arm with fewer errors. Study participants will eventually control the arm mounted on a wheelchair using a small joystick.

“The ability of robotic arms to benefit people today is largely limited due to high cost and excessive power consumption,” said Elishai Ezra Tsur, lead project researcher at Open University.  The support from ABR, Intel and Accenture will help the university explore adaptive controls in neuromorphic hardware in hopes of lowering the cost for a user-friend robotic arm, he added.

RELATED: Intel shares Pohoiki Springs chassis with AI researchers