Load-bearing Assessment for Safe Locomotion of Quadruped Robots on Collapsing Terrain

TL;DR

The paper tackles safe locomotion for quadruped robots on collapsing terrains, where exteroceptive sensing alone is unreliable. It proposes a unified framework that combines terrain probing via joint measurements, load-bearing assessment, trajectory optimization (TO), and perceptive model predictive control (MPC) to compute a GRF envelope for each leg and adapt footholds. The contributions include a TO that derives the GRF envelope, an MPC that enforces stability while allowing probing, and a state-machine for probe coordination, all validated on collapsing planks and rocky terrain. The results demonstrate safe traversal without specialized hardware, improving stability and safety for potential use in search-and-rescue and planetary exploration scenarios.

Abstract

Collapsing terrains, often present in search and rescue missions or planetary exploration, pose significant challenges for quadruped robots. This paper introduces a robust locomotion framework for safe navigation over unstable surfaces by integrating terrain probing, load-bearing analysis, motion planning, and control strategies. Unlike traditional methods that rely on specialized sensors or external terrain mapping alone, our approach leverages joint measurements to assess terrain stability without hardware modifications. A Model Predictive Control (MPC) system optimizes robot motion, balancing stability and probing constraints, while a state machine coordinates terrain probing actions, enabling the robot to detect collapsible regions and dynamically adjust its footholds. Experimental results on custom-made collapsing platforms and rocky terrains demonstrate the framework's ability to traverse collapsing terrain while maintaining stability and prioritizing safety.

Figures (12)

• Figure 1: On the top left, the Aliengo robot traverses a customized collapsing terrain with moving planks; on the top right, a terrain with collapsing rocks. On the bottom left, the selection of probing positions is aided by the terrain map (orange square); on the bottom right, the contact force envelope optimized for the probing leg.
• Figure 2: A simplified overview of the proposed framework.
• Figure 3: Safe regions for foot placement computed by the VFA algorithm for a scenario with a step. The white spheres indicate the regions of the foothold heightmap safe for foot placement, and the black spheres indicate the unsafe regions. They are enclosed by 2D polygons (orange, blue, yellow, and purple), which are used as constraints in the TO.
• Figure 4: Trajectory optimization: (a) the optimized feet trajectories are represented with colored lines: yellow, blue, purple, and orange. The probing leg is the right-front, and the GRF envelope is indicated by the orange force vectors on the target foot position; (b) the maximum and minimum forces are marked in red, green, and blue, for the x, y, and z-direction, respectively. These forces together compose the GRF envelope. To prevent slippage on the legs, all forces must be contained inside the friction cone, indicated in light red in (a).
• Figure 5: Terrain probing: the blue spheres are the next probing footholds, and the red spheres indicate the alternative probing positions for the current probing leg (right-hind). The orange arrows indicate the GRF envelope computed by the optimization. The limits of the support polygon constraint are indicated with the red polygon. The smaller blue sphere is the projection of the robot's CoM on the terrain in the direction of gravity. The green spheres are the probing-around positions to ensure safety in the region around the probing point. The contact forces are indicated with the green arrows.