PrivLLMSwarm: Privacy-Preserving LLM-Driven UAV Swarms for Secure IoT Surveillance

TL;DR

PrivLLMSwarm tackles privacy risks in LLM-driven UAV surveillance by enabling secure LLM inference via Secure MPC across a UAV swarm. It introduces MPC-optimized transformer components, GELU and SoftMax approximations, and a GPT-2 command generator trained with PPO in AirSim. The authors demonstrate high command semantic similarity, encrypted inference latency around 417 ms per image, and robust formation control in urban simulations, outperforming plaintext, DP, and FL baselines. They also publish a 30,000-sample synthetic dataset and open-source implementation to advance reproducibility and practical deployment in privacy-sensitive IoT applications.

Abstract

Large Language Models (LLMs) are emerging as powerful enablers for autonomous reasoning and natural-language coordination in unmanned aerial vehicle (UAV) swarms operating within Internet of Things (IoT) environments. However, existing LLM-driven UAV systems process sensitive operational data in plaintext, exposing them to privacy and security risks. This work introduces PrivLLMSwarm, a privacy-preserving framework that performs secure LLM inference for UAV swarm coordination through Secure Multi-Party Computation (MPC). The framework incorporates MPC-optimized transformer components with efficient approximations of nonlinear activations, enabling practical encrypted inference on resource-constrained aerial platforms. A fine-tuned GPT-based command generator, enhanced through reinforcement learning in simulation, provides reliable instructions while maintaining confidentiality. Experimental evaluation in urban-scale simulations demonstrates that PrivLLMSwarm achieves high semantic accuracy, low encrypted inference latency, and robust formation control under privacy constraints. Comparative analysis shows PrivLLMSwarm offers a superior privacy-utility balance compared to differential privacy, federated learning, and plaintext baselines. To support reproducibility, the full implementation including source code, MPC components, and a synthetic dataset is publicly available. PrivLLMSwarm establishes a practical foundation for secure, LLM-enabled UAV swarms in privacy-sensitive IoT applications including smart-city monitoring and emergency response.

Figures (7)

• Figure 1: Architectural overview of PrivLLMSwarm framework: Collaborative UAV System with LLM-Enhanced Edge Processing and Secure MPC for Joint Decision-Making in IoT environments.
• Figure 2: Four UAV (rotor) swarm drones configured for mission execution in the AirSim simulation environment, demonstrating the experimental setup for privacy-preserving swarm operations.
• Figure 3: Four UAV (rotor) swarm drones Return-to-Home Execution During Coordinated Flight Operations.
• Figure 4: Text-wise Semantic Similarity Comparison using cosine similarity method across different privacy approaches, demonstrating PrivLLMSwarm's superior balance of accuracy and privacy.
• Figure 5: Computation times breakdown for different components of the PrivLLMSwarm framework, highlighting the efficiency of MPC-friendly optimizations.