Scenario Selection and Prioritization for ADAS Hardware-in-the-Loop Validation: A Structured Critical Review

Hardware-in-the-Loop (HIL) validation is an important component of Advanced Driver Assistance Systems (ADAS) verification, enabling production-intent hardware and embedded software to be evaluated under controlled and repeatable conditions. However, the high dimensionality of driving scenarios, combined with limited HIL execution time and real-time test-bench constraints, makes exhaustive scenario execution impractical. This review critically examines scenario selection and prioritization approaches relevant to ADAS HIL validation, based primarily on literature published between 2021 and 2026 while retaining foundational earlier studies where necessary. The reviewed approaches are organized into four complementary families: risk-based, coverage-based, history/data-driven, and optimization-based methods. They are compared in terms of their underlying objectives, data requirements, computational characteristics, advantages, limitations, and applicability to HIL environments. The literature indicates that risk-based approaches provide safety traceability, coverage-based methods help prevent systematic gaps in the scenario space, history/data-driven techniques enable adaptive regression testing, and optimization-based methods provide mechanisms for balancing multiple validation objectives under resource constraints. However, an important evidence gap remains: recent scenario-selection and prioritization studies are predominantly evaluated using simulation or recorded datasets, whereas direct HIL studies primarily address test infrastructure, sensor stimulation, real-time execution, and fault injection. Consequently, the effectiveness of advanced selection strategies under physical HIL constraints remains insufficiently demonstrated. To provide a common basis for discussing these approaches, recurring elements identified in the literature are consolidated into an illustrative conceptual reference formulation combining safety relevance, coverage, historical evidence, and execution cost. This formulation is presented as a synthesis of existing concepts rather than as a novel or empirically validated optimization method. The review highlights the need for future studies that experimentally evaluate adaptive scenario-selection strategies under realistic HIL timing, sensor-fidelity, initialization, and resource constraints.