Sampling Strategy Design for Model Predictive Path Integral Control on Legged Robot Locomotion

1 st Chuyuan Tao, 2 nd Fanxin Wang*, 3 rd Haolong Jiang, 4 th Jia He, 5 th Yiyang Chen, 6 nd Qinglei Bu · 2026

Read Sampling Strategy Design for Model Predictive Path Integral Control on Legged Robot Locomotion by 1 st Chuyuan Tao, 2 nd Fanxin Wang*, 3 rd Haolong Jiang, 4 th Jia He, 5 th Yiyang Chen, 6 nd Qinglei Bu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-01409.txt.

Description

Six authors led by Chuyuan Tao study how a sampling-based controller should draw its trial moves when steering a legged robot. The controller, MPPI, tests many random control sequences, scores them, and blends the best into the next command. For a four-legged robot, naive random sampling wastes trials and produces jerky motion.

The team compares several sampling designs inside the same framework. Some draw each step independently; others build smooth spline curves through a few control points, which ties the steps together. Simulations on a quadruped platform measure smoothness, task success, robustness, and how many samples each design needs. The comparison shows sampling design is not a detail but a main lever on performance. The findings give practical guidance for deploying this controller class on complex legged machines, where every wasted trial costs stability.

Catalog record

CREATOR1 st Chuyuan Tao, 2 nd Fanxin Wang*, 3 rd Haolong Jiang, 4 th Jia He, 5 th Yiyang Chen, 6 nd Qinglei Bu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE31 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-01409

Browse the full library · About the Tower of Babel