Learning multiple gaits is non-trivial for legged robots, especially when encountering different terrains and velocity commands. In this work, we present an end-to-end training framework for learning multiple gaits for quadruped robots, tailored to the needs of robust locomotion, agile locomotion, and user's commands. A latent space is constructed concurrently by a gait encoder and a gait generator, which helps the agent to reuse multiple gait skills to achieve adaptive gait behaviors. To learn natural behaviors for multiple gaits, we design gait-dependent rewards that are constructed explicitly from gait parameters and implicitly from conditional adversarial motion priors (CAMP). We demonstrate such multiple gaits control on a quadruped robot Go1 with only proprioceptive sensors.

翻译：掌握多种步态对于足式机器人而言具有挑战性，尤其是在应对不同地形和速度指令时。本研究提出了一种面向四足机器人的端到端训练框架，用于学习多种步态，专门针对鲁棒运动、敏捷运动及用户指令需求进行优化。通过步态编码器和步态生成器同步构建潜空间，该框架能够帮助智能体复用多种步态技能以实现自适应步态行为。为学习多步态的自然行为，我们设计了步态相关奖励函数，这些奖励函数既通过步态参数显式构建，也通过条件对抗运动先验（CAMP）隐式构建。我们仅利用本体感觉传感器，在四足机器人Go1上验证了这种多步态控制方法。