Beyond Black-Box Testing: Active Binary Frequency-Division Modulation Inversion and Physical Tomography for Internal State Decoupling in True Random Number Generators

A Model-Based Inversion Framework via Active Frequency-Division Modulation This work proposes a shift from black-box statistical testing toward active white-box physical diagnosis for true random number generators (TRNGs). We present a tomographic methodology using active binary frequency-division modulation and algebraic inversion to decouple internal metastable and reference components, providing a reusable three-step framework (modeling–modulation–inversion) suitable for high-reliability design and process feedback. This manuscript has been submitted to the IEEE Transactions on Computers; the final version of record may differ. © Zhou Shicong et al., 2026. Preprint licensed under Creative Commons Attribution 4.0 International.