Abstract: In teleoperated surgery, robotic telesurgery uses the concept of master-slave robotic configuration to provide remote surgical intervention; however, the safety of teleoperation is influenced by factors such as communication delay, jitter, packet loss, and varying network quality. The literature on telesurgery techniques based on digital twins has largely focused on virtual synchronization, visual feedback, buffering, replay, or fixed-delay operations. In this study, we propose a Network State-Aware Digital Twin Trajectory Prediction and Safety-Constrained Compensation for Robotic Telesurgery. Within the proposed framework, the characteristics of the surgical robot kinematics and network communication, such as end-to-end delay, jitter, packet loss ratio, and bandwidth state, are utilized to predict the end-effector trajectory of the future delayed teleoperation. This digital twin augments the state of the robot and performs command compensation based on workspace constraints, confidence estimation of predictions, and velocity limitation while ensuring safety. We applied our approach to a simulation of a teleoperated robotic surgery environment while considering surgical kinematics and network impairments. The performability of telesurgical robotics is enhanced through the joint improvement of trajectory accuracy, synchronization, command/service continuity, and safety in the presence of delays, jitters, and packet loss using the proposed approach.

Key words: surgical training, virtual twins, predictive trajectory estimation, teleoperation systems, video streaming, surgical phantoms