Networked robotics continues to be a key component of various industrial systems. 6G will empower networked robotics with the envisaged capabilities like hyper-reliable and low-latency connectivity and in-network intelligence. While communication and control techniques have been widely investigated; co-design and AI-based solutions for autonomy and reasoning are still emerging. In this article, we propose a generative AI-in-the-Loop (GITL) framework for the control and coordination of networked robotic systems. The generative AI (GAI) agent oversees three interrelated loops: the communication loop, the robot control loop, and the AI model control loop, making holistic decisions by interpreting task requirements, network status, and robotic operations. Acting as a high-level cross-domain coordinator, the GAI agent leverages background knowledge of AI models, robot behaviors, and network policies, ensuring seamless interplay across different operational scenarios. This approach ultimately enables a highly autonomous networked robotic system capable of handling complex tasks with minimal human intervention. We anticipate that the GITL paradigm will play a pivotal role in unlocking the full potential of networked robotic systems, enabling enhanced autonomy, adaptability, and coordination in the emerging 6G and Industry 5.0 era.
Li et al. (Mon,) studied this question.