Safety-critical Motion Planning for Collaborative Legged Loco-Manipulation over Discrete Terrain

TL;DR

The results demonstrate that the approach successfully guides the team through obstacle courses, requiring planar positioning and height adjustments, and all happening on discrete terrain such as stepping stones.

Abstract

As legged robots are deployed in industrial and autonomous construction tasks requiring collaborative manipulation, they must handle object manipulation while maintaining stable locomotion. The challenge intensifies in real-world environments, where they should traverse discrete terrain, avoid obstacles, and coordinate with other robots for safe loco-manipulation. This work addresses safe motion planning for collaborative manipulation of an unknown payload on discrete terrain while avoiding obstacles. Our approach uses two sets of model predictive controllers (MPCs) as motion planners: a global MPC generates a safe trajectory for the team with obstacle avoidance, while decentralized MPCs for each robot ensure safe footholds on discrete terrain as they follow the global trajectory. A model reference adaptive whole-body controller (MRA-WBC) then tracks the desired path, compensating for model uncertainties from the unknown payload. We validated our method in simulation and hardware on a team of Unitree robots. The results demonstrate that our approach successfully guides the team through obstacle courses, requiring planar positioning and height adjustments, and all happening on discrete terrain such as stepping stones.

Figures (5)

• Figure 1: A team of quadrupedal robots transporting a payload. The robots navigate a challenging obstacle course on discrete terrain, stepping on stones, crawling under the horizontal wall, and adjusting their height to pass over the vertical bar. More details in https://youtu.be/MuJY9rYxTO4.
• Figure 2: Diagram of the Proposed Method. The global planner defines the team’s trajectory while avoiding obstacles. A decentralized loco-manipulation controller operates on each robot that uses local MPC to plan future footholds, followed by MRA-WBC to track the trajectory and manage uncertainties from the unknown payload.
• Figure 3: Robot team navigating an obstacle course with payload and disturbances. The sequence shows (a) trajectory deviation to avoid an obstacle, (b) crawling under a low wall, and (c) adjusting height to clear a vertical bar. Through all these situations, the local MPC generates future safe foothold locations for stable locomotion on stepping stones.
• Figure 4: Single MPC Performance. The figure shows the team navigating a stepping stone with a vertical bar using a single MPC planner. The MPC generates footholds outside the safe region and a high-risk trajectory, increasing the likelihood of obstacle collisions and locomotion failure.
• Figure 5: Hardware Experiment Snapshots. (a) The robot team lowers their height to pass beneath the black box obstacle, and (b) they return to normal operating height after passing through the obstacle. All this coordination is achieved through the use of the global MPC.