Design and Development of a Multi-Control Gesture-Recognition based Robotic Arm

Abstract

With the advent of automation technology, Human-Machine Interface (HMI) has emerged as new trend in industry today. However, precision control in such systems has been the main hindrance in path for further improvements. This paper proposes a wireless communication HMI framework to operate a human-resembling robotic arm that tracks and follows the controller's arm movements. A 6 Degree-of-Freedom prosthetic arm is designed on a 3D CAD software `SolidWorks' to enact human arm activity. The end-effector of this setup is a mechanical human hand. This model is constructed and then simulated in Simscape. Arduino boards and IDE are utilized for instruction and data transfer. Simulink's support package for Arduino hardware is used for the arm's physical modelling. To resolve the aforementioned problem, various types of control structures are explored and implemented. These include gyro-accelerometer based position tracking, arm skeletal imaging and color tracking by using MATLAB. Performances of all the structures are analyzed and compared with the simulation. To obtain best results, the creation of a platform ROS (Robot Operating System) is proposed that corrects the input control for each set of time frame in accordance with the simultaneously run simulation. In future, this technique can be employed in all types of bilateral teleoperations and haptic feedback systems.