Combination of Flex Sensor and Electromyography for Hybrid Control Robot

Abstract

The alternative control methods of robot are very important to solved problems for people with special needs. In this research, a robot arm from the elbow to hand is designed based on human right arm. This robot robot is controlled by human left arm. The positions of flex sensors are studied to recognize the flexion-extension elbow, supination-pronation forearm, flexion-extension wrist and radial-ulnar wrist.The hand of robot has two function grasping and realeasing object. This robot has four joints and six flex sensors are attached to human left arm. Electromyography signals from face muscle contraction are used to classify grasping and releasing hand. The results show that the flex sensor accuracy is 3.54° with standard error is approximately 0.040 V. Seven operators completely tasks to take and release objects at three different locations: perpendicular to the robot, left-front and right-front of the robot. The average times to finish each task are 15.7 ssecond, 17.6 second and 17.1 second. This robot control system works in a real time function. This control method can substitute the right hand function to do taking and releasing object tasks.