Teleoperated humanoid robots hold significant potential as physical avatars for humans in hazardous and inaccessible environments, with the primary goal of channeling human intelligence and sensorimotor skills through these robotic counterparts. Precise coordination between humans and robots is crucial for accomplishing whole-body behaviors involving locomotion and manipulation. To progress successfully, dynamic synchronization between humans and humanoid robots must be achieved. This work enhances recent advancements in human whole-body dynamic telelocomotion, addressing challenges in robustness. By embedding the hybrid and underactuated nature of bipedal walking into a virtual interface, termed a human walking reference, we achieve more dynamically consistent walking gait generation. Additionally, we introduce a reactive robot controller that reproduces scaled human motion, allowing the realization of telelocomotion behaviors on the full-body dynamics of a bipedal robot. Real-time telelocomotion simulation experiments validate the effectiveness of our methods, demonstrating that a trained human pilot can dynamically synchronize with a simulated bipedal robot, achieving sustained locomotion, controlling walking speeds within the range of 0.0 m/s to 0.3 m/s, and enabling backward walking for distances of up to 2.0 m. This research contributes to advancing teleoperated humanoid robots and paves the way for future developments in synchronized locomotion between humans and bipedal robots.

翻译：遥操作人形机器人在危险和难以进入的环境中作为人类的物理化身具有巨大潜力，其主要目标是借助这些机器人实体传递人类的智能与感知运动技能。人与机器人之间的精确协调对于完成涉及行走与操作的全身行为至关重要。要取得顺利进展，必须实现人与双足机器人之间的动态同步。本研究完善了近期人类全身动态遥距行走领域的进展，重点应对鲁棒性方面的挑战。通过将双足行走的混合与欠驱动特性嵌入至被称为"人类行走参考"的虚拟接口中，我们生成了更具动态一致性的行走步态。此外，我们引入了一种响应式机器人控制器，能够复现缩放后的人体运动，从而在双足机器人的全身动力学上实现遥距行走行为。实时遥距行走仿真实验验证了所提方法的有效性，结果表明：经过训练的人类操作员能够与仿真双足机器人实现动态同步，完成持续行走，将行走速度控制在0.0 m/s至0.3 m/s范围内，并实现距离达2.0米的后退行走。这项研究为推进遥操作人形机器人的发展做出贡献，并为未来人类与双足机器人之间同步行走的研究铺平道路。