Driven by the rapid evolution of embodied foundation models and advanced robotics, embodied agents continuously interact with the physical world through a "perception-decision-action" closed-loop, becoming critical carriers for AI deployment. Confronted with open and dynamic environments, individual agents inherently face limitations in perception breadth, computational boundaries, and action robustness. This inevitably drives the emergence of Embodied Agent Networks, which leverage networked multi-agent collaboration to transcend the performance bottlenecks of single agents.
Distinct from the singular "information transmission" attribute of traditional networks, Embodied Agent Networks, oriented toward complex physical interactions and collaborative operations, exhibit comprehensive characteristics of high intelligence, autonomy, and task-driven functionality. Their operation relies not only on underlying connectivity but also demands the deep integration of communication, sensing, computing, and control capabilities, thereby supporting diverse application scenarios such as multi-agent collaboration, high-reliability low-latency interaction, and integrated sensing-communication-computing.
However, existing communication network theories, rooted in the Shannon paradigm, primarily adhere to the design principles of "bit-fidelity, open-loop transmission, and functional decoupling." There remains room for adaptation to align with the intrinsic demands of Embodied Agent Networks during dynamic interactions—such as "closed-loop feedback, cross-layer synergy, and joint optimization of multidimensional resources." This poses new research challenges for network architecture evolution, topology adaptability, and intelligent resource orchestration. This Special Issue aims to establish a cutting-edge academic exchange platform to systematically explore the foundational theories, system architectures, and key enabling technologies of Embodied Agent Networks. By aggregating original research achievements in this field, it seeks to foster joint discussions and explorations of new theories, new technologies, and new methodologies for Embodied Agent Networks operating in complex environments.
This Special Issue solicits high-quality and original research papers focusing on critical theoretical and architectural frameworks, collaboration mechanisms, resilient networking, and integrated sensing-communication-computing technologies for embodied agent networks. The topics of interest include, but are not limited to:
Topics include (but not limited to):
1. Theory and Architecture of Embodied Agent Networks
Network architecture, protocol design, and standardization pathways for embodied agent tasks
Communication mechanisms and information-theoretic limits of embodied agent networks
Multidimensional evaluation framework and benchmark testing for the communication-sensing-computing performance of embodied agent networks
2. Collaborative Networking and Communication-Sensing-Computing Convergence Technologies for Embodied Agents
Physical-layer transmission and air interface link construction for embodied agent tasks
Dynamic and Resilient Networking and Self-Healing Mechanisms for Embodied Agents
Cloud-Edge-Device Collaboration for Task Offloading and Acceleration of Embodied Model Inference
Integrated Resource Scheduling and Waveform Design for Communication, Sensing, and Computing in Embodied Agent Networks
3. Network Security and Energy Efficiency Optimization for Embodied Agents
Inherent Security Mechanisms and Multimodal Privacy Protection Adapted to Embodied Agent Interactions
Low-Power Design and Green Communication Technologies for Long-Duration Operation of Embodied Agent Networks
4. Typical Scenarios and Cross-Domain Collaborative Applications
Embodied Agent Networks and Applications for Smart Industry, Precision Agriculture, Emergency Rescue, and Other Scenarios
Networked Collaboration Technologies for Multi-Domain Heterogeneous Embodied Agent Networks (Low-Altitude, Ground, Surface, and Underwater)
Task Simulation and Performance Evaluation of Embodied Agents Driven by Virtual-Physical Interaction and Digital Twins
Schedule
Submission Deadline: Sept 15, 2025
Acceptance Notification (1st round): Oct. 25, 2025
Minor Revision Due: Nov. 25, 2025
Final Decision Due: Oct. 15, 2025
Final Manuscript Due: Oct 30, 2025
Guest editors
Tao Jiang, Huazhong University of Science and Technology, China
Lin Cai, University of Victoria, Canada
Yansha Deng, King’s College London, United Kingdom
Fuhui Zhou, Nanjing University of Aeronautics and Astronautics, China
Submission guidelines
This special issue focuses on satellite Internet and aims to publish cutting-edge research on embodied agent networks theory and architecture as well as key technologies.
Papers should be submitted in two separate .doc files (preferred) or .pdf files: 1) Main Document (including paper title, abstract, key words, and full text); 2) Title page (including paper title, author affiliation, acknowledgement and any other information related with the authors’ identification) through the Manuscript Central. Please register or login at http://mc03.manuscriptcentral.com/chinacomm, then go to the author center and follow the instructions there. Remember to select “Embodied Agent Networks: Theory, Architecture, and Key Technologies—2027” as your manuscript type when submitting; otherwise, it might be considered as a regular paper.