A Vision-Based Assistive Robotic Arm for People with Severe Disabilities

Abstract

This paper presents a vision-based assistive robotic arm for people with severe disabilities. This system is composed of a robotic arm, a microcontroller, its controller, and a vision-based unit. The main body of the robotic arm that can be contained in a briefcase is about 5 kg, including two 12-V lead acid rechargeable batteries. This robotic arm is also capable of being mounted on a wheelchair. To obtain position coordinates of an object, image processing technique with a single Web camera was used. Position errors in the order of few millimeters were observed in the experiment. Experimental results of drinking water task with able-bodied subjects showed that they could smoothly carry out the tasks. The present results suggested that the resultant position errors were acceptable for drinking water command.