Robots learn safe whole-body movement for moon surface tasks
Learning Safety-Constrained Whole-Body Loco-Manipulation for Lunar Surface Operations
Robotics
Summary
Operating robots on the moon is risky because failures can't be fixed easily and people can't watch them all the time. The authors designed a way for a four-legged robot with arms to learn how to move and use its tools safely while working on the moon's surface. The robot figures out how to shift between walking and manipulating objects by itself, using special rules that keep it balanced in low gravity. Their approach was tested in moon-like simulations with different ground types and showed good precision and safety.
What this means in practice
- •For space robotics engineers: Develop autonomous lunar robots that safely perform walking and tool use without constant human control in variable moon terrain.
- •For industrial robot programmers: Program robots to coordinate whole-body movement and tool handling in complex, constraint-heavy environments with safety guarantees.
Authors
Alvaro Belmonte-Baeza, Miguel Cazorla, Gabriel J. García, Carlos J. Pérez-Del-Pulgar, Jorge Pomares
Abstract
Robotic systems for lunar surface operations must act autonomously and preserve their own integrity in an environment where failures are unrecoverable and continuous supervision is impossible. This paper presents a Constrained Reinforcement Learning framework for quadrupedal mobile manipulators that learns a single whole-body policy for tracking a 6D end-effector pose specified in task space, with operational limits enforced during learning through a constrained learning mechanism. We introduce a reward formulation in which the transition between locomotion and manipulation emerges from the reward geometry rather than from an explicit gating mechanism, a stability term derived from the Gravito-Inertial Inclination Margin and motivated by the reduced-gravity dynamics of the lunar surface, and a design rule for reward regularization that follows from the multiplicative structure of the termination-based objective. The framework is evaluated in a lunar-analog simulation spanning three terrain morphologies at five difficulty levels under lunar gravity, with regolith contact properties calibrated to the JSC-1a simulant and varied from loose to dense. The learned policy attains mean position and orientation errors of 3.22 cm and 6.83 deg. under nominal conditions, retains this precision across regolith densities and over the operationally representative range of terrain difficulty, and reduces soft-constraint violation time by up to an order of magnitude relative to an unconstrained baseline while incurring no hard-constraint violations.