Connectivity-Resilient Autonomous Navigation for Beyond-Visual-Line-of-Sight UAV Systems

Abstract

Purpose: This study aims to develop and evaluate a connectivity-resilient autonomous navigation framework that enables UAVs to maintain safe, efficient, and mission-compliant operation under varying connectivity conditions.

Methodology: This research adopts a system-driven design and analytical evaluation approach, integrating communication-aware modeling with artificial intelligence–based navigation strategies. The framework combines connectivity prediction models, reinforcement learning–based decision-making, and adaptive path planning algorithms to dynamically adjust navigation behavior. Simulation-based experiments are conducted across diverse operational scenarios, including urban, rural, and disaster environments, using realistic communication degradation profiles. Performance is evaluated using key metrics such as mission success rate, path efficiency, connectivity uptime, and energy consumption, with comparative analysis against traditional navigation methods.

Findings: The results demonstrate that the proposed framework significantly improves navigation robustness and mission success under intermittent connectivity conditions. Connectivity-aware path planning reduces exposure to communication dead zones, while the reinforcement learning engine enables adaptive decision-making in uncertain environments. Compared to conventional approaches, the system achieves higher mission completion rates, improved path efficiency, and optimized energy utilization. Nonetheless, performance trade-offs are observed in computational overhead and model training complexity, particularly in highly dynamic environments.

Unique Contribution to Theory, Policy and Practice: This study advances the field of autonomous UAV navigation by introducing a unified framework that explicitly integrates communication awareness into navigation intelligence. It contributes to theory by bridging the gap between UAV autonomy and network resilience, and provides a scalable architecture for real-world BVLOS deployment. From a policy and practice perspective, the findings support the development of safer BVLOS regulatory frameworks and offer actionable insights for UAV system designers, aviation authorities, and industry stakeholders seeking to enable reliable long-range autonomous operations.