SLM, LLM or Agentic AI? Toward Intelligent UAV-Enabled WPT Systems in Low-Altitude Economy Networks

Jul 16, 2026

Upload by Logic

Unmanned Aerial Vehicles (UAVs) enable low-altitude economic networks, but efficient and adaptive optimization remains difficult under resource-constrained, dynamic conditions. The work studies language-model approaches for UAV-enabled wireless power transfer (WPT). A lightweight Small Language Model (SLM) is built from a pre-trained BERT backbone with UAV embeddings, contextual features, a geometry-aware path decoder, and ensemble inference. An Agentic AI framework uses four collaborative agents in a closed loop of generation, optimization, evaluation, and reflection, then compares SLM/LLM/agentic results via simulations.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages15

Reading Time38 min

Introduction

Background

Motivation

SLM, LLM or Agentic AI? Toward Intelligent UAV-Enabled WPT Systems in Low-Altitude Economy Networks Feibo Jiang, Senior Member, IEEE, Li Dong, Lei Mao, Kezhi Wang, Senior Member, IEEE, Xianbin Wang, Fellow, IEEE, Abbas Jamalipour, Fellow, IEEE. arXiv :2607 .00255v 1 [ cs .IT] 30 Jun 2026 Abstract—Unmanned Aerial Vehicles (UAVs) have become key enabling platforms for low-altitude economic networks, yet achieving efficient and adaptive optimization under resourceconstrained and dynamic environments remains challenging. This paper investigates language models for UAV-enabled Wireless Power Transfer (WPT) systems. First, a lightweight Small Language Model (SLM)-based solution is developed using a pre-trained BERT backbone, enhanced UAV embeddings and contextual features, a geometry-aware path decoder, and ensemble inference to achieve low complexity, low latency, and high energy efficiency. Second, an Agentic AI-based framework is designed to exploit the reasoning and interactive capabilities of Large Language Models (LLMs). It integrates four collaborative agents—Initializer, Actor, Critic, and Reflector—to form a closed loop of generation, optimization, evaluation, and reflection for iterative UAV path and energy optimization. Finally, simulations compare the SLM-, LLM-, and Agentic AI-based approaches. Index Terms—Small Language Model, Large Language Model, Agentic AI, Unmanned Aerial Vehicle, Wireless Power Transfer, Path Planning I. INTRODUCTION A. Background In recent years, the Low-Altitude Economy Network (LAENet) has emerged as a new frontier of future information infrastructure and economic growth. It is gradually evolving into a multi-layered network that integrates transportation, energy supply, air–ground collaboration, and intelligent services. Within LAENet, Unmanned Aerial Vehicle (UAV) has become This work was supported in part by the National Natural Science Foundation of China under Grants 62572184, and 41604117; in part by the Natural Science Foundation of Hunan Province under Grants 2024JJ5270 and 2025JJ50365; in part by the Changsha Natural Science Foundation under Grants kq2402098 and kq2402162 (Corresponding author: Li Dong.) Feibo Jiang ([jiangfb@hunnu.edu.cn](jiangfb@hunnu.edu.cn)) is with the Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing and the Yuelushan Digital Intelligence Laboratory (Artificial Intelligence and International Communication), Hunan Normal University, Changsha 410081, China. Li Dong ([Dlj2017@hunnu.edu.cn](Dlj2017@hunnu.edu.cn)) is with the School of Computer Science, Hunan University of Technology and Business, Changsha 410205, China, and also with the Xiangjiang Laboratory, Changsha 410205, China. Lei Mao ([202520294367@hunnu.edu.cn](202520294367@hunnu.edu.cn)) is with School of Information Science and Engineering, Hunan Normal University, Changsha 410081, China. Kezhi Wang ([Kezhi.Wang@brunel.ac.uk](Kezhi.Wang@brunel.ac.uk)) is with the Department of Computer Science, Brunel University London, London UB8 3PH, UK. Xianbin Wang ([xianbin.wang@uwo.ca](xianbin.wang@uwo.ca)) is with the Department of Electrical and Computer Engineering, Western University, London N6A 5B9, Canada. Abbas Jamalipour ([a.jamalipour@ieee.org](a.jamalipour@ieee.org)) is with the School of Electrical and Computer Engineering, University of Sydney, Australia, and with the Graduate School of Information Sciences, Tohoku University, Japan. a core node supporting network operation and value realization due to its high mobility, rapid deployment, and wide-area coverage capabilities. However, UAV systems also face multidimensional challenges related to limited computing, communication, and energy resources, which place higher demandson their ability to operate over extended periods. Energy consumption is a critical bottleneck that constrains both the flight endurance and service capabilities of UAVs. Due to payload and size limitations,

FAQs

What problem does the paper address in UAV-enabled WPT systems?

It addresses the challenge of achieving efficient and adaptive optimization for UAV-enabled wireless power transfer under limited computing, communication, and energy resources in dynamic environments.

How does the proposed SLM-based approach work?

It builds a lightweight Small Language Model using a pre-trained BERT backbone, enhanced UAV embeddings and contextual features, a geometry-aware path decoder, and ensemble inference to reduce complexity and latency while improving energy efficiency.

How does the Agentic AI framework optimize UAV paths and energy?

It employs four collaborative agents—Initializer, Actor, Critic, and Reflector—running in a closed loop that iteratively performs generation, optimization, evaluation, and reflection to refine UAV path and energy decisions.

Jul 16, 2026

Upload by Logic