Research Of Force Reflecting Teleoperation Based On Virtual Reality

The telepresence robot system with human-robot interaction is an effective way to perform tasks in unknown or hazard environment. The visual and force information are fed back to the operator though the human-robot interface based on multi-sensor. The operator teleoperates the remote robot to precisely accomplish various tasks as a part of the control loop. Force teleprensence is one of the most important contents of telepresence technology. The teleoperation system based on force telepresence is very important for contact task, such as grasp, assemblage etc.The key issue to carry out the delicate task for a force telepresence system is to improve the transparency of the system, when the system is stable. Through analyzing the teleoperation system based on position-force architecture, a gain-switching control scheme is developed. The preferable transparency has achieved through changing the position control gain of the slave. We give dynamics equation of the teleoperation system and analyze the system's stability at constrained motion with Lyapunov function.As the force-reflecting teleoperation system is often applied in long distance areas such as space and underwater exploration etc, the time delay in the communication between the master and the slave becomes the main problem in the teleoperation. Virtual reality is an effective method to eliminate the influence of time-delay. But it depends on the accuracy of the virtual model. For the sake of a more precise model, the modeling technology of virtual predictive model is analyzed. The information fusion based on multi-sensor and the transcendental information are used in the modeling. We utilize the visual, force and position information to build and modify the virtual model and improve its geometrical accuracy. Augmented Reality technology is applied in the teleoperation system. Graphic model is overlaid on the video images to correct the model when the discrepancy is beyond a certain threshold. We adopt the sliding average least square (SALS) method to identify the parameters of mass, damp, stiffness of the remote environment and use this information to modify the dynamic model of the environment. The environment dynamics is divided into three procedures: free motion, transition from free motion to contact and stable contact. Different models are used to describe the procedures respectively. We also set up an impedance model of the environment during the contact, and propose the calculating method of virtual contact force.We consider the predictive virtual model as a time forward observer and constructing the dynamics equation of the system. The control law of teleoperation system based on virtual predictive environment with error compensate observer is put forward. By using the Lyapunov function of the system, the system's stability is analyzed. The condition of the system's transparency is given in the end.A one-DOF force-reflecting teleoperation system based on virtual reality is built up, associated with the national 863 Hi-Tech space project. The proposed methods are verified in this system by experiments with time delay of 0~30 second. Experimental results show that the system is stable and has a satisfactory consequence in position and force traking.