Learning to Walk and Fly with Adversarial Motion Priors

TL;DR

This paper introduces a novel approach enabling robots to seamlessly switch between walking and flying by using adversarial motion priors, which learn from motion datasets and adapt through reinforcement learning, without complex reward engineering.

Contribution

The work presents a new method for multimodal locomotion in robots, combining adversarial motion priors with reinforcement learning for automatic mode switching.

Findings

Successful imitation of human-like gait and aerial motions

Automatic emergence of mode-switching behavior

Potential applications in search, rescue, and exploration

Abstract

Robot multimodal locomotion encompasses the ability to transition between walking and flying, representing a significant challenge in robotics. This work presents an approach that enables automatic smooth transitions between legged and aerial locomotion. Leveraging the concept of Adversarial Motion Priors, our method allows the robot to imitate motion datasets and accomplish the desired task without the need for complex reward functions. The robot learns walking patterns from human-like gaits and aerial locomotion patterns from motions obtained using trajectory optimization. Through this process, the robot adapts the locomotion scheme based on environmental feedback using reinforcement learning, with the spontaneous emergence of mode-switching behavior. The results highlight the potential for achieving multimodal locomotion in aerial humanoid robotics through automatic control of walking and flying modes, paving the way for applications in diverse domains such as search and rescue, surveillance, and exploration missions. This research contributes to advancing the capabilities of aerial humanoid robots in terms of versatile locomotion in various environments.