Pattern recognition based movement control and gripping forces control system on arm robot model using LabVIEW

Abstract

Most arm robot has an inefficient operating time because it requires operator to input destination coordinates. Besides, main problem of arm robot is object’s vulnerability when it is manipulated by the robot. This research goals is to develop an arm robot control system which has ability to automatically detect object using image processing in order to reduce operating time. It is also able to control gripping force for eliminating damage to objects caused by robot gripper. This research is implemented in LabVIEW 2011 software to control arm robot model which can represent industrial scale robot. The software is designed with informative visualization to help user learn and understand robotic control concept deeply. The system can automatically detect object position based on pattern recognition method which has four steps: pre-processing process to initialize picture taken by camera, segmentation process for separating object from the background, classification process to determine characteristics of object, and position estimation process to estimate object position in the picture. The object’s position data are then calculated by using kinematic equation to control the robot’s motion. The results show that the system is able to detect object and move the robot automatically with accuracy rate in x-axis is 95.578 % and in y-axis is 92.878 %. The system also implements modified PI control method with FSR as input to control gripping force with maximum overshoot value 10 %. Arm robot model control system developed is successfully meet the expectation. The system control can be implemented to industrial scale arm robot with several modification because of kinematic similarity between model and industrial scale robot.