Quadrupeds achieve rapid and highly adaptive locomotion owing to the coordination between their legs and other body parts such as their trunk, head, and tail, i.e. body-limb coordination. Therefore, a better understanding of the mechanism underlying body-limb coordination could provide informative insights into the improvement of legged robot mobility. Sprawling locomotion is a walking gait with lateral bending exhibited in primitive legged vertebrates such as salamanders and newts. Because primitive animals are anticipated to possess the essence of quadruped motor control, their locomotion helps better understand body-limb coordination mechanisms. Previous studies modeled neural networks in salamanders and employed it to control robots and investigate and emulate sprawling locomotion. However, these models predefined the relationship between the legs and the trunk, such that how body-limb coordination is attained is largely unknown. In this article, we demonstrate that sensory feedback facilitates body-limb coordination in sprawling locomotion and improves mobility through mathematical modeling and robot simulations. Our proposed model has cross-coupled sensory feedback, that is, bidirectional feedback from body to limb and limb to body, which leads to an appropriate relationship between the legs and the trunk without any predefined relationship. Resulting gaits are similar to the sprawling locomotion of salamanders and achieve high speed and energy efficiency that are at the same level as those of a neural network model, such as conventional models, optimizing the relationship between the legs and the trunk. Furthermore, sensory feedback contributes to the adaptability toward leg failure, and the bidirectionality of feedback facilitates parameter tuning for stable locomotion. These results suggest that cross-coupled sensory feedback facilitates sprawling locomotion and potentially plays an important role in the body-limb coordination mechanism.

中文翻译：

---

在伸展运动中，身体和四肢之间具有交叉耦合的感觉反馈的分散控制。

四足动物由于其腿部与身体其他部位（例如躯干，头部和尾巴）之间的协调（即肢体-肢体协调）而实现了快速且高度自适应的运动。因此，对肢体-肢体协调机制的更好理解可以为腿式机器人移动性的改善提供有益的见解。伸展运动是步态行走的步态，在primitive腿和vertebrate等原始有腿脊椎动物中表现出侧向弯曲。因为预计原始动物将具有四足动物运动控制的本质，所以它们的运动有助于更好地理解肢体协调机制。先前的研究在sal中对神经网络进行建模，并将其用于控制​​机器人并研究和模拟蔓延运动。然而，这些模型预定义了腿部与躯干之间的关系，因此很大程度上未知如何实现肢体协调。在本文中，我们证明了感觉反馈通过数学建模和机器人仿真促进了肢体协调运动，并提高了运动能力。我们提出的模型具有交叉耦合的感觉反馈，即从身体到肢体以及从肢体到身体的双向反馈，这导致了腿和躯干之间的适当关系，而没有任何预定义的关系。最终的步态类似于of的运动，并实现了与神经网络模型（如常规模型）相同的水平的高速和高能效，从而优化了腿部和躯干之间的关系。此外，感觉反馈有助于适应腿部故障，并且双向反馈有助于稳定运动的参数调整。这些结果表明，交叉耦合的感觉反馈促进了运动的扩展，并可能在肢体协调机制中发挥重要作用。