Adaptive Machine Learning Framework for UAV Trajectory Optimization in O-RAN

2026-06-23

AI Summary

This work addresses the inefficiency of frequent retraining required for trajectory optimization models when deploying unmanned aerial vehicles (UAVs) as O-RUs in dynamic and unknown environments. To overcome this limitation, the authors propose an adaptive trajectory planning framework that integrates enhanced continual transfer learning. The approach leverages a library of pretrained models coupled with a model selection mechanism to rapidly transfer knowledge from similar environments, while also incorporating a continually refined fallback model to guarantee baseline performance in the absence of relevant prior scenarios. Evaluated on real urban maps with ray-tracing-based channel modeling, the method reduces convergence time by 44%–56% compared to training from scratch and achieves up to 40% faster convergence than conventional transfer learning, substantially improving deployment efficiency and environmental adaptability.

Abstract

The deployment of unmanned aerial vehicles (UAV) as open radio units (O-RUs) in 6G cellular systems presents a promising opportunity to achieve scalable and adaptive network coverage. However, optimizing UAV trajectories in dynamic and unfamiliar environments remains a critical challenge, particularly due to the need for extensive retraining in each new scenario. In this paper, we introduce a novel UAV trajectory optimization framework that integrates enhanced continual transfer learning within the O-RAN architecture. The proposed system maintains a library of pre-trained models and employs a model selection mechanism to identify and transfer knowledge from the most relevant environments, minimizing adaptation time and improving efficiency. When no sufficiently similar model is available, a fallback model empowered by continuous refinements ensures baseline performance. The framework leverages real-world city maps and ray tracing techniques to enhance learning reliability and improve trajectory planning. Simulation results demonstrate that the proposed model selection-based transfer learning approach reduces convergence time by 44% to 56% compared to retraining from scratch, and up to 40% compared to traditional transfer learning without model selection.

Problem

• UAV trajectory optimization
• O-RAN
• continual transfer learning
• 6G cellular systems
• dynamic environments

Innovation

• continual transfer learning
• UAV trajectory optimization
• O-RAN
• model selection
• ray tracing