The validation of perception sensor models is critical for ensuring the credibility of virtual testing frameworks employed in Advanced Driver Assistance Systems (ADAS) and Autonomous Driving (AD) technologies. This study presents a comprehensive overview of the challenges and methodologies associated with sensor model validation. Key contributions include a detailed classification of sensor models by their modeling techniques (geometrical, stochastic, and physical) and output types (object lists and detection lists). Emphasis is placed on the necessity of precise environmental simulations, high-quality reference data, and the selection of appropriate validation metrics. The innovative Double Validation Metric (DVM) is introduced, offering robust evaluation by separately quantifying bias and scattering errors. The study also highlights the importance of modular co-simulation architectures, enabling flexible integration of various sensor models and adherence to emerging standards such as ASAM OSI and OpenMATERIAL. Future directions underscore the need for a standardized material database and validated co-simulation frameworks, addressing critical gaps in the virtual validation of automated driving technologies.