Omnidirectional Platform for Autonomous Mobile Industrial Robot

Abstract

Omnidirectional mobile platforms are holonomic robots that can independently and simultaneously perform translational and rotational motions. In order to develop an autonomous omnidirectional mobile manipulator, this paper presents a platform based on four mecanum wheels. It has a higher carrying capacity and mobility than a standard four-wheel platform. The used manipulator is a Fanuc LR Mate 200 iD/7l robot with an R-30iB Mate Plus Controller. The heavy weight of the industrial arm and the controller makes collision-free navigation a challenge. To navigate with this robot in an unknown semi-structured indoor environment, a Hokuyo 2D Lidar and a Realsense D435i camera have been used. The Central Processing Unit is an Nvidia Jetson TX2 running Ubuntu Linux on which ROS (robot operating system) was installed. The robot is capable of autonomously performing Simultaneous Localization and Mapping (SLAM), navigation, obstacle detection, and object recognition, vision-guided robot motions. A map of our workplace was generated. Most mobile robot motion control approaches rely on dynamic or kinematic models. The study also covers mathematical modeling of the four-wheeled omnidirectional platform that leads to the robot's kinematics. The simulations were carried out using MATLAB to establish and verify the kinematic model of the omnidirectional platform. The robot was controlled to follow curves with a constant translation velocity of 1m/s.