Compliant Control of Quadruped Robots for Assistive Load Carrying

TL;DR

The paper tackles robust, compliant load carrying with a quadruped by estimating the external base wrench from proprioception and applying an admittance-based base controller, augmented by a Control Barrier Function (CBF) to guarantee collision avoidance with a leader. It presents a model-based framework that decouples base motion from footstep planning, enabling a single acceleration command to track user input, maintain compliance, and avoid collisions via a 1-D CBF along the line to the leader. A convex MPC (ConvexMPC) for base motion and a reactive GPGD-based controller manage footstep planning and contact forces, with High-Performance Interior Point Method (HPIPM) solving the MPC QP and a QP for foot contact forces. The approach is validated in MuJoCo simulations with latency and on a Go2 quadruped, including push-recovery, assistive load carrying, and human-robot and robot-robot collaborative scenarios, demonstrating accurate wrench estimation, disturbance rejection, and safe collaboration. This work advances practical, scalable compliant control for legged robotic teams in industrial, search-and-rescue, and collaborative settings by enabling proprioception-only force estimation, admittance-based base control, and CBF-based safety guarantees.

Abstract

This paper presents a novel method for assistive load carrying using quadruped robots. The controller uses proprioceptive sensor data to estimate external base wrench, that is used for precise control of the robot's acceleration during payload transport. The acceleration is controlled using a combination of admittance control and Control Barrier Function (CBF) based quadratic program (QP). The proposed controller rejects disturbances and maintains consistent performance under varying load conditions. Additionally, the built-in CBF guarantees collision avoidance with the collaborative agent in front of the robot. The efficacy of the overall controller is shown by its implementation on the physical hardware as well as numerical simulations. The proposed control framework aims to enhance the quadruped robot's ability to perform assistive tasks in various scenarios, from industrial applications to search and rescue operations.

Figures (20)

• Figure 1: Velocity response of the robot in $x$ and $y$ axis after push along +$y$ axis (simulation).
• Figure 2: Base height adjustment during push experiment (simulation).
• Figure 3: Applied and estimated force on the robot (simulation).
• Figure 4: Vertical base load estimation comparison for $F_z=50$ N (simulation).
• Figure 5: Base velocity for the applied circular base force (simulation).