Reproduce “Confidential FRIT via Homomorphic Encryption” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Haruki Hoshino, Jungjin Park, Osamu Kaneko, Kiminao Kogiso
A confidential data-driven gain-tuning framework for Cyber-Physical Systems (CPS) is introduced, enabling the exact computation of Fictitious Reference Iterative Tuning (FRIT) on encrypted data. It shows ElGamal encryption allows for approximately 200 times faster server-side computation compared to CKKS, both achieving comparable control performance with 128-bit security.
Key insights
- Reformulating matrix inversion into a summation of cofactor-derived products allows for exact homomorphic computation, distinguishing this method from prior approximation-based approaches.
- ElGamal encryption, despite being partially homomorphic, is highly effective for the transformed FRIT algorithm, offering substantially better computational efficiency than CKKS for this specific application.
- The proposed methods successfully manage quantization errors, producing control system performance and closed-loop characteristics nearly identical to conventional, unencrypted FRIT.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.