Pose based bilateral teleoperation using wave variables and wave prediction.
详细信息
文摘
Teleoperation over large distances can result in substantial communication delays. Such delays may result in poor control performance and eventually instability, especially in the case of Internet-based teleoperation with variable and unpredictable communication delays.;In this thesis, a stable, fast, and effective remote robot control method via the Internet is investigated, using a pose-based visual servoing (PBVS) teleoperation system. First a general wave variable approach is studied to guarantee teleoperation system stability despite time delays in the communication links. Then, motivated by the wave reflection problem in the conventional wave variable method, a wave variable based predictor is explored to improve teleoperation performance and to achieve faster and more accurate robot control. The predictor model is obtained through system identification of the controlled remote robot using the modified least squares method. The predictor is designed using a multi-step Kalman prediction, which uses the delayed wave variable measurements to predict the corresponding current system states and provide accurate predicted feedback information to the operator for control use.;The proposed approaches investigated in the thesis are applied to a simulated PBVS teleoperation system model and an experimental PBVS CRS PLUS robot system, which is manipulated by an operator at the local site. Both simulation results and experimental results verify the effectiveness of the wave variable method and illustrate that wave prediction can greatly enhance teleoperation performance and system robustness.