Secure uplink transmission strategy for deep learning-based surveillance UAV

Abstract

This paper addresses the challenges in unmanned aerial vehicles (UAVs) surveillance, particularly under constraints of limited computational capability and energy resources. We introduce the UAV Surveillance Optimization Framework (USOF), an innovative algorithm that integrates with edge computing (EC) servers to optimize surveillance performance. Employing a Lyapunov optimization framework, the USOF algorithm is designed to jointly optimize four critical performance metrics: end-to-end (E2E) latency, frames per second (FPS), energy consumption, and data confidentiality. Our numerical results demonstrate that the USOF algorithm significantly enhances surveillance UAVs’ performance by maintaining a balanced optimization across these metrics. This approach presents a substantial advancement in UAV surveillance, offering a more efficient and secure method of operation that could be pivotal in modern surveillance applications.