Learning Whole-Body Control for a Salamander Robot

2026-03-17

AI Summary

Existing salamander robots predominantly rely on central pattern generators or model-based control, lacking a unified, learning-based, joint-level whole-body controller that generalizes to high-degree-of-freedom physical platforms—particularly for amphibious multimodal locomotion. This work proposes a reinforcement learning–based proprioceptive control framework that directly maps velocity commands and proprioceptive states to joint actions, enabling coordinated locomotion to emerge naturally. Through systematic simulation-to-reality alignment and transfer strategies, the approach achieves, for the first time, learned whole-body control on a high-DoF salamander robot capable of seamlessly transitioning between terrestrial walking and aquatic swimming. The controller’s stability and generalization are validated on both flat and uneven real-world terrains.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Amphibious legged robots inspired by salamanders are promising in applications in complex amphibious environments. However, despite the significant success of training controllers that achieve diverse locomotion behaviors in conventional quadrupedal robots, most salamander robots relied on central-pattern-generator (CPG)-based and model-based coordination strategies for locomotion control. Learning unified joint-level whole-body control that reliably transfers from simulation to highly articulated physical salamander robots remains relatively underexplored. In addition, few legged robots have tried learning-based controllers in amphibious environments. In this work, we employ Reinforcement Learning to map proprioceptive observations and commanded velocities to joint-level actions, allowing coordinated locomotor behaviors to emerge. To deploy these policies on hardware, we adopt a system-level real-to-sim matching and sim-to-real transfer strategy. The learned controller achieves stable and coordinated walking on both flat and uneven terrains in the real world. Beyond terrestrial locomotion, the framework enables transitions between walking and swimming in simulation, highlighting a phenomenon of interest for understanding locomotion across distinct physical modes.

Problem

• whole-body control
• salamander robot
• amphibious locomotion
• sim-to-real transfer
• reinforcement learning

Innovation

• Reinforcement Learning
• Whole-Body Control
• Sim-to-Real Transfer
• Amphibious Locomotion
• Salamander Robot