Robust Optimal Design with Latin Hypercube Sampling Method for Remote Sensing Satellite in LEO Orbit

The design of remote sensing satellites in Low Earth Orbit (LEO) presents significant challenges due to inherent orbital uncertainties such as altitude, inclination, right ascension of the ascending node (RAAN), and elevation angle. To address these issues during early-stage design, this study proposes an integrated Robust Design Optimization (RDO) framework that combines the Teaching-Learning-Based Optimization (TLBO) algorithm with Latin Hypercube Sampling (LHS). The primary objective is to minimize the satellite's total mass while ensuring stable performance under environmental perturbations. Unlike conventional approaches that treat uncertainty quantification and optimization separately, the proposed method embeds stochastic behavior directly into the design loop, capturing the interaction between uncertain parameters and design decisions. The framework is validated using real-world case studies, including Aqua, VRSS1, and CloudSat satellites, showing a significant reduction in design error compared to classical optimization. Results confirm that the TLBO-LHS integration enhances design robustness, reduces system mass, and improves mission reliability. This methodology offers a scalable and practical solution for the robust design of other complex space systems affected by uncertainty.