AERIS: Aerial-Edge Role-Driven Intelligence at Runtime via Orchestrated Language-Model Swarm

2026-06-29

AI Summary

This work addresses the challenge of efficiently deploying large language models on resource-constrained aerial robots to simultaneously achieve real-time responsiveness and sustained long-horizon task execution. To this end, the authors propose AERIS, a novel framework that introduces a role-based runtime orchestration mechanism for language model swarms. AERIS dynamically instantiates lightweight role-specific models, enables on-demand rebinding of perception and control modules, and aligns attention mechanisms with subgoal structures to jointly optimize instruction decomposition and closed-loop control. Integrated with a heartbeat-driven execution scheduler, AERIS establishes a robust end-to-end intelligent control loop on high-fidelity drone visual-language navigation benchmarks and demonstrates superior planning capabilities and rapid response performance in real-world scenarios.

Abstract

Integrating large language models into robotic systems holds promise for enhancing autonomy, yet practical deployment remains constrained by strict heartbeat-constrained scheduling and limited computational power. We propose AERIS: an edge deployment framework for aerial platforms. It organizes dedicated small language models combined with lightweight perception and control modules into roles that can be instantiated at runtime, and dynamically rebinds them across different executors as resources change, thereby pushing intelligent capabilities to the edge. AERIS achieves long-horizon instruction decomposition through an attention-subgoal alignment mechanism, which involves annotating the currently active instruction step in messages, thereby progressively approaching long-term objectives. We evaluate AERIS on a high-fidelity UAV Vision-and-Language Navigation benchmark. Under a heartbeat-timed execution mechanism, AERIS maintains a stable perception-decision-control loop between a low-frequency planner and a high-frequency controller, supporting real-time closed-loop operation. We further validate its deployability through two real-world experiments focused on planning and fast response. A demonstration video is provided in the supplementary materials.

Problem

• large language models
• aerial robotics
• edge deployment
• real-time control
• resource-constrained systems

Innovation

• edge intelligence
• language-model swarm
• runtime role orchestration
• attention-subgoal alignment
• UAV vision-and-language navigation