Abstract
<title>Abstract</title> <p>Beyond-visual-line-of-sight (BVLOS) unmanned aircraft operations increasingly depend on heterogeneous space-air-ground networks, on-board autonomy, remote sensing, edge computing and external geospatial information. Existing safety approaches typically assess these dependencies separately: command-and-control (C2) performance is treated as a communications problem, perception confidence as an autonomy problem, and operational authorisation as a pre-flight compliance problem. This separation creates an assurance gap because degradation can propagate across layers before any single subsystem crosses its conventional failure threshold. This method article introduces ASSURE-BVLOS, an assurance-aware architecture that combines five continuously updated risk dimensions: C2 integrity, sensing reliability, data freshness and provenance, environmental exposure, and vehicle health. A weighted multiplicative risk index, an explicit uncertainty penalty, hysteretic authority transitions and a digital safety-case evidence loop are defined. The method is evaluated using 100,000 reproducible Monte Carlo mission states spanning disaster response, linear-infrastructure inspection and maritime surveillance, with mixed terrestrial, low-Earth-orbit and geostationary links. Against a static, teleoperation-centred baseline, the proposed policy reduced simulated hazardous outcomes from 12.76% to 9.62%, corresponding to a 24.6% relative risk reduction, while eliminating simulated continuation under the predefined high-C2-risk condition. Sensitivity analysis indicates that benefits are retained across mission classes but depend strongly on correct threshold calibration and trustworthy input confidence. The contribution is not a certified control law; it is a transparent systems-engineering method for linking communications, sensing, data pipelines, autonomy and traceable runtime evidence. The framework is designed to support subsequent hardware-in-the-loop and flight-test validation and to provide a structured bridge between runtime assurance research and operation-centric risk assessment.</p>