Towards Low-Gravity Planetary Exploration using Reinforcement Learning for Walking, Jumping, and In-flight Attitude Control

2026-05-23

AI Summary

This study addresses the challenges of terrain traversal and landing faced by conventional ground robots in low-gravity planetary environments such as Mars. The authors propose a reinforcement learning–based control framework for a five-bar linkage quadrupedal robot, achieving, for the first time, coordinated control of walking, vertical jumping, forward leaping, and in-air posture adjustment under low-gravity conditions. By developing a Mars-gravity dynamics model and employing Sim2Real transfer techniques, the policy trained in simulation is successfully deployed on the physical robot. Experimental results demonstrate that the robot can reorient its body by 90° in midair within 2.6 seconds, achieve a vertical jump height of 3.1 meters and a forward leap distance of 3.9 meters in simulation, and effectively navigate complex terrains, thereby validating the efficacy and advancement of the proposed approach.

Abstract

This paper presents reinforcement learning (RL) policies for dynamic quadrupedal locomotion in planetary exploration scenarios. Building on a taskoptimized quadruped with a 5-bar leg design, we develop RL policies for walking, vertical jumping, forward jumping, and in-flight attitude control, explicitly tailored to the reduced gravity on Mars. These policies jointly enable such robots to overcome obstacles larger than themselves through coordinated jumping and precise in-flight reorientation for safe landings. We demonstrate Sim2Real transfer of the attitude control policy on the Olympus quadruped through single-axis reorientation tests, while all locomotion policies are validated in simulation. A complete Mars exploration mission scenario demonstrates coordinated policy deployment across challenging terrain. Experimental results show 90° attitude reorientation in 2.6 seconds, with simulations demonstrating 3.1 meter vertical jumps and 3.9 meter forward jumps under Martian gravity conditions. - Supplementary video: https://www.youtube.com/watch?v=qlSJ3P87A4A

Problem

• low-gravity locomotion
• planetary exploration
• quadrupedal robot
• in-flight attitude control
• obstacle traversal

Innovation

• reinforcement learning
• low-gravity locomotion
• quadrupedal robot
• in-flight attitude control
• Sim2Real transfer