LEO satellite agents improve onboard task offloading reliability and efficiency
ORCA: Onboard Reflective Collaborative Agents for Reliable Task Offloading in LEO Satellite Networks
Networking and Internet Architecture
Summary
LEO satellites need to work together to process important tasks quickly, but their different abilities and changing connections make this hard. The authors created ORCA, a system where small onboard agents perceive, think, act, and learn to decide how to share tasks. Their tests show ORCA helps satellites handle tasks efficiently and stay reliable, even when some satellites act in unexpected or harmful ways. This helps satellites work better without needing a central controller.
What this means in practice
- •For satellite network operators: Manage onboard task sharing in satellite constellations to maintain reliable processing despite failures or attacks on some satellites.
- •For distributed computing engineers: Design decentralized collaboration protocols for networks with changing links and variable node trustworthiness, inspired by ORCA’s agent framework.
Authors
Yunxin Geng, Shengxiang Hao, Yichen Wu, Haotian Liu, Weiwei Jiang, Lu Cao, Han Meng, Yuxuan Pan, Zhiqing Wei
Abstract
Low Earth orbit (LEO) satellites increasingly perform time-critical onboard processing, yet heterogeneous computing capabilities, rapidly changing inter-satellite links (ISLs) and uncertain neighbor behavior make decentralized task offloading difficult under partial observations. This paper presents ORCA, an architecture of Onboard Reflective Collaborative Agents for autonomous onboard task offloading. ORCA organizes offloading decisions into a four-Agent \emph{Perceive}$\rightarrow$\emph{Reason}$\rightarrow$ \emph{Act}$\rightarrow$\emph{Learn} loop. It aims to improve efficiency and reliability in dynamic, heterogeneous, and unreliable LEO satellite networks. The simulation results of a heterogeneous 60-satellite Walker-delta constellation demonstrates that ORCA delivers competitive system efficiency in cooperative settings while achieving superior task-success robustness and offloading reliability under previously unseen malicious node behavior, indicating that outcome-grounded reflection can improve the reliability of decentralized onboard offloading without centralized runtime coordination.