Research On Predictive Display Technology Based On Real-time Reconstruction Model In Space Teleoperation

In recent years,the technology of artificial intelligence has developed rapidly,but the fully autonomous intelligent robot,which can sense,make decisions,plan and perform tasks independently,is still the direction of human science and technology development.Facing with the dangerous environment that many people are not easy to reach,such as space,deep water environment,nuclear radiation environment,it is a good choice to use robots instead of human to perform tasks.While avoiding the threat of dangerous environment,it can also improve the efficiency of operations.At present,teleoperation robot technology is the main means to solve the task in such dangerous environment.The dangerous environment that is difficult to reach is often unknown.The key of teleoperation technology under unknown space environment is generally the problem of large time delay and environmental unknown.Teleoperation technology based on predictive model has good robustness and transparency for large time delay,and has attracted much attention.In this paper,the technology of predictive display Teleoperation Based on real-time reconstruction model in space teleoperation is studied,and the problems of large time delay and environmental unknown are solved at the same time.The main work is as follows:Firstly,the key technologies of real-time 3D reconstruction from end-to-end environment are studied.Kinect Fusion is used to reconstruct the unknown environment.Through three-dimensional reconstruction,the slave environment is structured,and the spatial position information of the base of the manipulator is obtained.In the master,the virtual manipulator after registered is superimposed on the video screen to form the augmented reality visual prediction display together with the three-dimensional reconstruction model.Real-time three-dimensional reconstruction results is also the foundation for the subsequent force prediction.Secondly,aiming at the difficulty of force prediction and display in teleoperation technology based on Augmented Reality under unknown environment,KF-VPS force rendering algorithm is proposed,which combines real-time three-dimensional reconstruction technology and force rendering technology in virtual scene to solve the force prediction and display problem between virtual manipulator and real slave environment.The key of the algorithm is to adopt a data representation of three-dimensional scene,which can not only express the results of three-dimensional reconstruction conveniently,but also perform collision detection and contact force calculation conveniently.Thirdly,the human-computer interaction technology in the master is studied,and a graphic interaction method based on 3D interactive path planning is proposed.Aiming at the problems of difficulty and inefficiency of direct operation using equipment in the master such as hand controller,isomorphic manipulator and exoskeleton in some multi-obstacle or narrow space,this graphic interactive mode allows the operator to complete the path planning of manipulator by drawing the desired trajectory of manipulator in 3D interactive interface,and instructs the manipulator to complete teleoperation.In the mission planning stage,human decision-making and judgment ability is fully developed without the involvement of the manipulator,it is safe and efficient.In the task execution stage,the manipulator can complete the task independently without the operator's participation,it reduces the operator's pressure.Finally,an teleoperating system experimental platform based on reconstruction model is built,and the research contents of this paper are analyzed and verified.The experimental results show that the force prediction algorithm and the graphical interaction method based on 3D interactive path planning are effective.