Space teleoperation with large time delay based on vision feedback and virtual reality

Abstract

Main difficulty encountered in space teleoperation is a large time delay in communication between earth and space [1]. Predictive control strategy [2] via virtual reality (VR) or augmented reality (AR) is an available approach in many literatures. As to a manipulator in 3D Cartesian coordinates, it should be predicted from both horizontal direction and vertical direction. If we only use VR in two directions, an operator cannot know the real environment's information, whereas we only use AR in two directions, the communication link can't afford it because of limited bandwidth. We propose a new shared control scheme incorporating the technologies of VR, AR and visual auto-aligning together, which utilizes VR and AR to predict the position and orientation in horizontal or vertical direction respectively, and aligns the target object automatically by visual auto-aligning in the close distance. A ground experimental system is set up, and many tasks of capturing a client satellite, adjusting the orientation of client satellite and replacing a component have been accomplished. The results prove that this method can reduce the influence of communication delay, and fulfill quickly space tasks.