Digital Twin Technology for Robot Grasping in Complex Environments

Robotic manipulation tasks involving complex and cluttered circumstances remain a challenge for existing autonomous robots. The digital twin (DT) platforms open new avenues for improving robustness of grasp by real-time model learning and prediction-guided correction, and virtual-to-physical adaptation. This paper reviews the DT technology for robotic grasping. First, the fundamental concepts and architectures of DTs for robotic grasping tasks was provided. Then, a comprehensive review of the current research status was conducted, including the static objects in simulated environment, adaptive grasping under varying conditions, and algorithms. Finally, a summary of the current key issues was made based on the current situation, and the direction of future development was discussed. This paper provides a systematic overview of this evolving field and offers development insights for enhancing the autonomy of robots operating under uncertain conditions.