Resource-Efficient Semantic Communication for Heterogeneous Agentic Teams

2026-09-30

AI Summary

This study addresses the cooperative communication bottlenecks among heterogeneous agents operating over constrained wireless links by proposing a goal-oriented semantic communication framework. The proposed method jointly optimizes message content, transmission timing, and reliability in a closed-loop manner, leveraging edge broadcasting for knowledge updates to ensure robust delivery of high-value messages and precise scheduling under unsaturated budgets while supporting large language model (LLM) agent integration. Experimental evaluations demonstrate that, in search-and-rescue missions, uplink channel occupancy is reduced by 1.2 to 8.5 times (up to 16.6 times when accounting for overhead). Furthermore, LLM-driven scenarios achieve bandwidth savings of 3.2 to 3.5 times, significantly enhancing overall communication resource efficiency.

Abstract

Teams of autonomous agents, including large language model (LLM) agents, must coordinate over scarce and unreliable wireless links. We propose goal-oriented semantic communication (GOSC), a closed-loop co-design that jointly decides what each agent sends, when it sends it, and how reliably it is transmitted, based on each message's value to the team task. An edge broadcast of the team's common knowledge closes the loop by updating these values. We prove that a message is sent only if its value exceeds the cost of delivering it and that more valuable messages receive more robust transmission rates, and we show that the scheduler solves each scheduling step exactly whenever the radio budget is not saturated, which held in 95% of scheduling decisions. In search-and-rescue missions validated on unseen scenarios, GOSC meets the same mission targets as carefully tuned periodic semantic schemes with 1.2--8.5 times fewer uplink channel uses. In most settings, this advantage persists with realistic packet overheads, reaching 16.6 times fewer uplink channel uses and 13.8 times lower cost when downlink costs are included; in the rescue task, it also persists when all agents share one uplink. Rough value estimates suffice, whereas values that ignore message content can fail. With three different LLMs, GOSC uses 3.2--3.5 times fewer channel uses, while completion-time gains depend on the model.

Problem

• Semantic Communication
• Multi-Agent Coordination
• Resource Efficiency
• Wireless Networks
• Large Language Models

Innovation

• Goal-Oriented Semantic Communication
• Heterogeneous Agentic Teams
• Closed-Loop Co-Design
• Resource-Efficient Scheduling
• Large Language Model Agents