Towards faster response in virtual reality applications for robotics control

Abstract

Virtual reality (VR) applications became increasingly popular recently. Their benefits now are not limited by the game and 3D animation software, as industrial VR applications are introduced to control various types of robots and manipulators or to assist in assembling 3D parts of a product either in CAD software or in the factory. However, the main problem with this technology is the high bandwidth requirement, which means some noticeable delay and latency in certain applications. This paper discusses the latency problem in VR-based imitation missions and compares its impact on different imitation systems. The first system uses a Leapmotion sensor as an input device for virtual control of the robot, the second uses HTC VIVE Pro controllers, and the third uses a deep learning hand recognition and tracking system based on VIVE eye cameras. The solutions to this problem according to the different described methods were proposed.