201 Brain-Machine Interface Control of a Robotic Arm for Object Grasping is Improved With Computer-Vision Based Shared Control. (August 2015)

• Record Type: Journal Article
• Authors: Tyler-Kabara, Elizabeth C.; Downey, John; Weiss, Jeffrey; Muelling, Katharina; Venkataraman, Arun; Valois, Jean-Sebastien; Javdani, Shervin; Herbert, Martial; Bagnell, J. Andrew; Schwartz, Andrew; Collinger, Jennifer
• Is Part Of: Clinical neurosurgery. Volume 62(2015)Supplement 1
• Publication Date: 2015-08
• Subjects: Nervous system -- Surgery -- Congresses; Neurosurgery; Nervous system -- Surgery; Neurologie; Congresses; Conference papers and proceedings; 617.48
• DOI: 10.1227/01.neu.0000467165.82278.28
• Languages: English
• ISSNs: 0069-4827

Abstract

Abstract : INTRODUCTION: Brain-machine interface neuroprosthetic arms for people with upper limb impairment are developing quickly, but could be improved through intelligent computer-vision-based assistance. Grasping and manipulating objects requires very accurate control of a prosthetic arm and hand, and is required for these limbs to eventually be used clinically. With the computer helping to stabilize the hand during grasping, the user's control would not need to be as accurate, and they would be free to concentrate on the larger goals of the arm movements. METHODS: A brain-machine interface was used to control a robotic arm to complete a subset of tasks from the Action Research Arm Test to determine 2 subjects' functional control of the arm. The task was done with and without computer-vision-based assistance. The computer-vision system identified objects and how the objects could be stably grasped. Once the user approached the object, the system helped the movements to ensure a stable grasp. RESULTS: Both subjects successfully completed the tasks more often with the grasp assistance than without. The assistance lowered the speed with which the arm moved while near the objects, but did not increase the amount of time required to complete the task. This shows that the assistance made the movements both more accurate and more efficient. Both subjects reported that the arm was easier to use with assistance. CONCLUSION: By integrating brain-machine interface-based high-level control with computer-vision-based low-level control of a robotic arm, people with tetraplegia showed improved functional use of the arm. This result highlights the importance of combining neuroscience- and robotic-based assistive technologies to create a highly flexible and effective neuroprosthetic arm for people with upper limb impairment.