RAMP: Reaction-Aware Motion Planning of Multi-Legged Robots for Locomotion in Microgravity

Authors: Warley F. R. Ribeiro, Kentaro Uno, Masazumi Imai, Koki Murase, Kazuya Yoshida · ICRA 2023

Licences

Open Access
Paper: CC BY 4.0

Everything below draws on the full paper under its CC BY 4.0 licence; its figure is credited to the authors.

TL;DR

A motion planner for gripping legged robots in microgravity that shapes swing-leg trajectories and moves the body to absorb swing momentum, reducing reactions that detach grippers.

Why it matters

Legged robots with gripping feet are proposed for asteroids and space stations, where hopping is uncontrolled and propellant-based flight is limited. RAMP gives a planning-level way to lower gripper pulling forces without needing friction estimates, and its tunable distribution factor offers a low-reaction alternative to reactionless control that can avoid singular configurations.

Problem

In microgravity the only forces that can pull a legged robot's grippers off the surface come from its own motion, so dynamic walking risks detachment and loss of the robot. Quasi-static walking avoids this but is too slow for mission time limits, friction-based compliant control needs friction data that is unavailable before a mission, and reactionless control was shown only for a single contact and can drive the robot into singular poses.

Proposed solution

RAMP combines a Low-Reaction Swing Trajectory, a 7th-order Bézier swing path whose coefficients are optimised to minimise peak momentum change while reaching the step height, with Whole-Body Swing Momentum Distribution. The latter computes a base velocity that cancels a fraction α of the swing momentum while holding the supporting grippers at zero velocity; α = 0.5 gives partial distribution (RAMP-PMD) and α = 1 full distribution (RAMP-FMD).

Key insights

Under microgravity, gripper pulling forces come only from the robot's own motion, so minimising its momentum change directly reduces detachment risk.
Optimising only the swing trajectory reduces reactions but is not enough; distributing swing momentum to the base and support legs is also needed.
Fully cancelling swing momentum (α = 1) can push a low-inertia robot into a singularity, so the distribution factor must be tuned to the robot's inertia, speed and manipulability.
The low-reaction swing trajectory reduces reactions partly by bringing the swinging limb closer to the body, lowering its inertia.

Main results

Simulation of a Hubrobo-like quadruped on a rough asteroid surface (10⁻⁶ G, 0.9 N gripper holding force): the baseline plan detached and the robot floated away after a few steps.
In the same simulation, RAMP-PMD (α = 0.5) advanced 40 cm without exceeding the gripper force limit, while RAMP-FMD (α = 1) moved the base too far and fell into a singular configuration.
On a dual-arm air-floating robot with one limb fixed to a plate, RAMP-FMD reduced maximum ground reaction force and moment to 40% of the baseline (a decrease of up to 60%).
Average force and moment fell to between 60% and 25% of the baseline average depending on the distribution factor (a decrease of up to 75%).
Body attitude changed by 11° with partial and 21° with full momentum distribution in the air-floating test, with no singularity problems on that higher-inertia robot.

Limitations

The distribution factor must be tuned for each robot to avoid distributing more momentum than the robot can absorb; full distribution caused a singularity in simulation.
The physical test is planar and covers a single step, because the air-floating robot cannot walk continuously; one limb is bolted to a plate, so actual detachment is not tested on hardware.
The quadruped walking results come from simulation only, using a compliant point-contact model with assumed surface stiffness and damping.
Trajectories and inverse kinematics for the experiment were computed offline.

Field impact

Finalist for the ICRA 2023 Outstanding Locomotion Paper award, with 7 citations. No code, data or hardware designs are released with the paper.

Abstract

Robotic mobility in microgravity is necessary to expand human utilization and exploration of outer space. Bio-inspired multi-legged robots are a possible solution for safe and precise locomotion. However, a dynamic motion of a robot in microgravity can lead to failures due to gripper detachment caused by excessive motion reactions. We propose a novel Reaction-Aware Motion Planning (RAMP) to improve locomotion safety in microgravity, decreasing the risk of losing contact with the terrain surface by reducing the robot's momentum change. RAMP minimizes the swing momentum with a Low-Reaction Swing Trajectory (LRST) while distributing this momentum to the whole body, ensuring zero velocity for the supporting grippers and minimizing motion reactions. We verify the proposed approach with dynamic simulations indicating the capability of RAMP to generate a safe motion without detachment of the supporting grippers, resulting in the robot reaching its specified location. We further validate RAMP in experiments with an air-floating system, demonstrating a significant reduction in reaction forces and improved mobility in microgravity.

Open Source Robotics