Architecture of Human-controlled Arm Manipulator Operator Training System

Abstract

Industrial robots of different kinematic structures and functionality are widely used nowadays in different areas of industry and most of them function autonomously with a pre-programmed set of actions or typical trajectories. Some special purpose mobile and stationary robots are human-controlled and in most cases require qualified operator skills. This article describes general approaches and practical aspects of the operator training and qualification system for arm-based special purpose manipulators. The set of training system functions include: general manipulator control (basic motion) and grasping actions, manipulator operations in restricted working areas, operator actions using virtual reality helms and multi-monitor visualization, etc. The hardware and software architecture of training system should allow for an effective training process and teamwork of trainee and instructor to reach good results. The paper proposes a ROS-based software part of the training system with a description of inter-process communication messages, data formats, integration with a database and the following analysis of trainee action.