2026-10-15T00:38:47Z https://www.scidb.cn/oai

10.57760/sciencedb.43001 2026-07-16T09:41:32Z

2026-07-16 Bayesian Optimization Trial Log for PD Gain Tuning of a Quadruped Robot Subject to Lunar-Gravity Landing Impacts doi:10.57760/sciencedb.43001 en This dataset documents the complete trial history of a Bayesian optimization (TPE-based) process aimed at tuning the PD controller gains for a Unitree Go2 quadruped robot operating under lunar gravity (1.625 m/s²). A total of 200 parameter configurations were evaluated, each tested across 11 drop heights (0.45–0.55 m) in Isaac Sim simulations. For every successful trial, the settling time and steady-state oscillation amplitude were recorded; failed trials were assigned a penalty value of 10. The dataset includes the full parameter space, intermediate scoring metrics per height, and convergence traces, providing a comprehensive benchmark for studying controller robustness and optimization efficiency in low-gravity legged locomotion. Bayesian optimization; PD gain tuning; quadruped robot 韩凌杰 盛强 张聚乐 郭栋 李楠 PUBLIC https://creativecommons.org/licenses/by/4.0/ dataset Science Data Bank