Convergent Frameworks for Evaluation and Ethical Deployment of Autonomous Aerial and Ground Vehicles: Simulation-informed Edge-AI, Scenario Generation, and Explainable Decision Architectures

Abstract

This article synthesizes contemporary advances in autonomous vehicle systems—spanning unmanned aerial vehicles (UAVs), connected ground vehicles, and the convergent infrastructure that enables them—into a unified research narrative oriented toward rigorous evaluation, ethical decision-making, and deployable system design. Building strictly from the provided scholarship, we construct a conceptual and methodological framework that integrates cut-out scenario generation for safety assessment, simulation-centered verification, edge-compute-enabled artificial intelligence, sensor-fusion-driven perception and localization, and explainable decision architectures. The work articulates a layered methodology in which reasonability-bounded scenario generation (Muslim et al., 2023) seeds simulation-driven testbeds (Khatiri et al., 2024) executed in edge-enabled platforms (McEnroe et al., 2022; Zhang et al., 2023). We examine how deep sensor fusion approaches (Fayyad et al., 2020; Koch, 2023) and multipolicy behavior prediction (Galceran et al., 2017) underpin safe motion planning and lateral control (Biswas et al., 2022), and how explainable AI mechanisms (Atakishiyev et al., 2021) and ethical decision-making models (Patil et al., 2025) shape acceptance and regulatory readiness. Our descriptive results articulate the expected behaviors, failure modes, and evaluation metrics that arise when these components are integrated—providing a richly detailed narrative of findings from a theoretical and systems engineering standpoint. The discussion unpacks theoretical implications for robustness, generalizability, and socio-technical governance; highlights methodological limitations inherent in simulation-only validation and distributional shift; and prescribes a staged research agenda that emphasizes cross-validation between lab-scale, simulation, and real-world testbeds such as those envisioned in the SAKURA project (SAKURA Project, 2023). We conclude by offering concrete recommendations for researchers, test engineers, and policymakers seeking practical pathways to reliable, explainable, and ethically defensible autonomous systems in both aerial and ground domains.