Communication-Efficient Digital-Twin Coordination for Heterogeneous LLM Embodied Agents over Computing Power Networks
2026-07-10 • Artificial Intelligence
Artificial IntelligenceMultiagent Systems
AI summaryⓘ
The authors address challenges in managing teams of robots or agents that use different language-based AIs to work together in places like factories. They point out problems with current systems that rely on many conversations between agents, which cause heavy communication, delays, and quality issues. Their solution, called LDT-Coord, uses a lightweight digital twin that collects each agent's chosen actions and timing info to coordinate without lots of talking. This method reduces communication drastically while keeping task success rates high, even when agents have different AI capabilities.
Embodied agentsLarge language models (LLMs)Digital twinCoordination frameworkCommunication overheadMulti-agent systemsConstrained POMDPPPO-Lagrangian algorithmTask success rate
Authors
Nuocheng Yang, Sihua Wang, Zihan Chen, Tony Q. S. Quek, Changchuan Yin
Abstract
Embodied agent teams powered by heterogeneous large language models (LLMs) are being widely deployed in physical artificial intelligence such as smart factories, warehouses, and service robotics. To enable collaboration among such an agent team, efficient coordination mechanisms that operate reliably under limited network resources are required. However, existing heterogeneous LLM-agent coordination frameworks that rely on multi-round natural-language-based conversations introduce three coupled challenges. First, inter-agent dialogue incurs communication overhead that grows rapidly with team size. Second, the quality of coordination is constrained by the heterogeneous capabilities of the agent team's LLMs. Third, agents may suffer from action delays due to iterative negotiation. To address these challenges, we propose LDT-Coord, a networked coordination framework built upon a lightweight digital twin (DT). Specifically, each agent independently selects its intended action and reports both the action decision and a structured temporal constraint over shared resources to the DT server, thereby decoupling coordination performance from natural-language reasoning ability. Then, DT executes a training-free, rule-based orchestrator algorithm to resolve cross-agent conflicts and returns coordination instructions to prevent such conflicts. To further reduce communication overhead, we formulate agent reporting control as a constrained partially observable Markov decision process (C-POMDP) and solve it with the PPO-Lagrangian algorithm. Simulation results show that LDT-Coord achieves a task success rate comparable to conventional coordination methods while reducing communication overhead by more than 70x and maintaining robustness under LLM heterogeneity.