Every year, heatstroke-related child fatalities in cars highlight the urgent need for action. Automakers are developing sensor-based systems to detect children and initiate rescue measures. However, validating these systems through simulation faces a critical challenge: a lack of real-world child data to accurately validate simulation models. Our study addressed this gap by testing market-standard sensors with real children in various scenarios and comparing the results to simulation outputs. Using controlled environments and standardized methodologies, we provide valuable insights into improving detection accuracy and safety. This presentation will share technical approaches, challenges, and results from this research.