Learning Dynamic Pick-and-Place for a Legged Manipulator

2026-05-15

AI Summary

This study addresses the challenge of coordinating agile locomotion and precise manipulation in legged mobile manipulators during dynamic walking, particularly under varying payload conditions. The authors propose a hierarchical reinforcement learning framework that integrates explicit mass estimation with adaptive whole-body control, enabling a quadrupedal robot equipped with a 6-DOF arm to perform continuous and rapid pick-and-place tasks while walking dynamically. This approach overcomes prior limitations restricted to light payloads and slow, segmented operations, demonstrating robust performance with heavier loads—achieving an 86.05% success rate with a 2.3 kg payload in simulation—and an extended vertical workspace in real-world experiments (0–1.1 m), where it attained an average success rate of 73.3% with a 1.3 kg payload in 4.06 seconds per task.

Abstract

Legged manipulators extend robotic capabilities beyond static manipulation by integrating agile locomotion with versatile arm control. However, achieving precise manipulation while maintaining coordinated locomotion remains a major challenge. This work presents a hierarchical reinforcement learning framework for dynamic pick-and-place tasks using a quadruped equipped with a 6-DOF robotic arm. The framework incorporates an explicit mass estimation module enabling adaptive whole-body control for objects with varying weights. In simulation, the system achieves an 86.05% success rate with payloads up to 2.3 kg. The approach is further validated through real-world experiments across six representative scenarios with controlled variations in object physical properties (size and mass) and task heights. Specifically, within a wide vertical workspace ranging from ground level to 1.1~m-high tabletops, the system demonstrates an average success rate of 73.3% for payloads up to 1.3 kg, with an average execution time of 4.06 s. Unlike prior works that handle lightweight objects and execute pick-and-place motions with slow, piecewise motions, the proposed framework exploits concurrent locomotion and manipulation for dynamic, continuous execution. These results demonstrate the potential of quadrupedal mobile manipulators for adaptive, whole-body pick-and-place with heavier payloads and extended workspaces.

Problem

• dynamic pick-and-place
• legged manipulator
• whole-body control
• adaptive manipulation
• mobile manipulation

Innovation

• hierarchical reinforcement learning
• dynamic pick-and-place
• legged manipulator
• mass estimation
• whole-body control