Context-Aware Industrial Robot Testing: Low-Cost Virtual Prototyping Environment

Abstract

This paper presents a case study regarding the simulation of a robotic workstation to virtually test objects detection and obstacle avoidance. Testing these features inside a virtual environment is useful, especially when human-robot cooperation and interaction are involved. Indeed, it allows users to avoid real dangerous conditions, lowering the possible risks of injuries by the users and cutting down the costs compared to a real testing environment. The work presented here exploits a framework where a virtual environment is connected to a Robot Operating System (ROS) able to simulate the kinematic of the robot and, on the other side, a physical ultrasonic sensor acts as the bridge with the real world. The latter, driven by an Arduino board, allows the virtual robot to recognize static obstacles in the real world, mapping the surrounding environment and computing a suitable trajectory to accomplish the given task. Thanks to the sensing capability, the virtual robot is also able to react to the presence of other obstacles (e.g. humans) entering the workspace at runtime. The seamless connection between the virtual and the physical world makes the framework suitable for the fast testing of new algorithms driving the behavior of the robot when interacting with dynamic environments.