LLM-DaaS: LLM-driven Drone-as-a-Service Operations from Text User Requests
Lillian Wassim, Kamal Mohamed, Ali Hamdi
TL;DR
The paper addresses the challenge of converting free-text drone service requests into structured DaaS tasks under uncertain weather. It introduces LLM-DaaS, which fine-tunes multiple LLMs (e.g., Phi-3.5, LLaMA-3.2 7b, Gemma 2b) on paired free-text and structured DaaS requests and couples this with a weather-aware routing and drone-selection pipeline. The work formalizes DaaS as $DaaS = \langle W, P, I \rangle$ with $W = \langle T, WS, WD, H, P \rangle$, $P = \langle N, E \rangle$, and $PA = \langle H, A^*, D, W_{adj} \rangle$, and demonstrates a dataset of 5,000 structured requests along with extensive LLM fine-tuning that yields near-perfect G-Eval scores, alongside pathfinding comparisons such as Route 46 and Route 47. The approach shows robust automation of free-text to executable drone tasks, real-time weather adaptation, and multi-drone composition, indicating strong potential for scalable, efficient, and reliable DaaS operations in dynamic environments.
Abstract
We propose LLM-DaaS, a novel Drone-as-a-Service (DaaS) framework that leverages Large Language Models (LLMs) to transform free-text user requests into structured, actionable DaaS operation tasks. Our approach addresses the key challenge of interpreting and structuring natural language input to automate drone service operations under uncertain conditions. The system is composed of three main components: free-text request processing, structured request generation, and dynamic DaaS selection and composition. First, we fine-tune different LLM models such as Phi-3.5, LLaMA-3.2 7b and Gemma 2b on a dataset of text user requests mapped to structured DaaS requests. Users interact with our model in a free conversational style, discussing package delivery requests, while the fine-tuned LLM extracts DaaS metadata such as delivery time, source and destination locations, and package weight. The DaaS service selection model is designed to select the best available drone capable of delivering the requested package from the delivery point to the nearest optimal destination. Additionally, the DaaS composition model composes a service from a set of the best available drones to deliver the package from the source to the final destination. Second, the system integrates real-time weather data to optimize drone route planning and scheduling, ensuring safe and efficient operations. Simulations demonstrate the system's ability to significantly improve task accuracy, operational efficiency, and establish LLM-DaaS as a robust solution for DaaS operations in uncertain environments.