A Portable Vision-based and Force Controlled Collaborative Robot System for Entertainment Purposes

Abstract:
The goal of this work is to realize a smart game with a lightweight robot. The manipulator, equipped with a visual sensor through the camera mounted on the arm and a passive gripper, plays against a human. As an example of smart game, we have chosen the game connect four for implementation. The robot used in this work is a Universal Robot UR3e equipped with an internal Force/Torque sensor on the flange. A low-cost USB camera is employed as the visual sensor. The gripper mounted on the robot has adhesive characteristics and therefore is independent from external power supply. All these properties make the proposed system able to be easily transported to different locations as an educational and entertainment service. To make the system intelligent enough to perform rational decisions, the Minimax algorithm from game theory is applied. This ensures a proper decision-making process on the part of the robot. To implement the game autonomously, the robot must feature the following functions and abilities: detection of board status, waiting and recognizing the opponent’s move, the ability to recognize and distinguish own’s disks, reaching and grasping the disks, calculation of the next move, dropping the piece in the specified place. The communication between the system’s units is achieved using ROS middleware and ROS drivers for the camera and the robot. Finally, we have validated our developed robot through actual playing against human.