Quantum-Resistant Homomorphic Encryption for Secure Federated Learning in CubeSat Constellations for Real-Time Exoplanet Detection

• Author(s): Govind Mulchandani
• Paper ID: 1719027
• Page: 2033-2038
• Published Date: 19-06-2026
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 9 Issue 12 June-2026

Abstract

The growing adoption of CubeSat constellations has introduced a new era in exoplanetary science by enabling low-cost, distributed, and continuous astronomical observation. Simultaneously, advances in onboard artificial intelligence have made real-time processing of photometric data increasingly feasible within spaceborn systems. However, the integration of federated learning into satellite networks introduces significant challenges related to data privacy, communication security, and long-term cryptographic resilience. Existing satellite communication protocols rely heavily on classical public-key cryptography, which is vulnerable to future quantum computing attacks. Given the extended operational lifespan of space missions, this creates a critical security concern for scientific data and collaborative machine learning processes. This paper proposes a quantum-resistant federated learning framework for CubeSat constellations dedicated to real-time exoplanet detection. The proposed architecture integrates lightweight homomorphic encryption based on the learning with errors (LWE) problem with a resource-aware federated learning pipeline optimized for low Earth orbit (LEO) environments. By enabling secure aggregation of encrypted model gradients without exposing raw observational data, the framework preserves both privacy and scientific integrity while maintaining resilience against quantum adversaries. The system architecture further incorporates hierarchical inter-satellite communication, onboard TinyML processing, and FPGA-accelerated cryptographic operations tailored to CubeSat hardware limitations. Experimental evaluations conducted using simulated Kepler and TESS light curve datasets demonstrate that the proposed framework achieves strong detection performance while operating within realistic satellite power, memory, and bandwidth constraints. The results indicate that quantum-resistant secure federated learning is both technically feasible and practically deployable for future distributed space missions. This work establishes a foundation for secure autonomous scientific collaboration in next-generation satellite constellations.

Keywords

Federated Learning, CubeSats, Post-Quantum Cryptography, Homomorphic Encryption, Exoplanet Detection, Learning with Errors, TinyML, Space Security

Citations

IRE Journals:
Govind Mulchandani "Quantum-Resistant Homomorphic Encryption for Secure Federated Learning in CubeSat Constellations for Real-Time Exoplanet Detection" Iconic Research And Engineering Journals Volume 9 Issue 12 2026 Page 2033-2038 https://doi.org/10.64388/IREV9I12-1719027

IEEE:
Govind Mulchandani "Quantum-Resistant Homomorphic Encryption for Secure Federated Learning in CubeSat Constellations for Real-Time Exoplanet Detection" Iconic Research And Engineering Journals, 9(12) https://doi.org/10.64388/IREV9I12-1719027