Dog galloping on rough terrain exhibits similar limb co-ordination patterns and gait variability to that on flat terrain.
Dog galloping on rough terrain exhibits similar limb co-ordination patterns and gait variability to that on flat terrain.
复制标题
在崎岖地形上奔跑的狗表现出与在平坦地形上相似的肢体协调模式和步态变化。
DOI：
10.1088/1748-3190/abb17a
复制
发表时间：
2021
影响因子：
3.4
通讯作者：
Wilshin S
中科院分区：
文献类型：
--
作者：
Wilshin S
了解动物在运动过程中如何调节它们的步态可以提供生物学见解，并启发机器人控制器。为什么动物以最快的速度奔跑仍然没有完全解释清楚。假设的驰振原因包括它能够募集脊柱肌肉骨骼结构，如碰撞理论所预测的，它使能量损失最小化，或者它提供了延长的飞行阶段，有更多的时间用于腿部放置，从而增强或提供必要的机动性[亚历山大1988 Am. 28 237-45; Ruina，Bertram和Srinivasan 2005 J. Theor. 237 170-92; Usherwood 2019 J. Exp.部分A 333 9-19; Hildebrand 1989 Bioscience 39 766-75]。后一种假设在机器人学中有一定的意义，基于多稳定性概念的控制器已经获得了一些牵引力。在这里，我们通过研究狗在平坦和粗糙地形上的步态动力学来检验这一假设。这一假设预测，在粗糙地形的疾驰步态期间，将噪声注入到地面接触的时间和位置中将导致各向同性更嘈杂的疾驰步态，以平坦地形上使用的疾驰为中心。我们发现，狗的步态在粗糙的地形上的腿摆动时间是不一致的平均步态的变量，并限制这种变化的上限值是不太可能的生物相关的。然而，平均步态的位置实际上仅移动少量。因此，我们发现对这一假设的支持有限。这表明，实现具有严格调节的目标奔马步态仍然是期望的行为，并且步态的大量可变性不是奔马的期望特征。对于机器人技术，我们的研究结果表明，在粗糙的地形上出现的动物环境动力学并没有表现出均匀更广泛的吸引力盆地。未来的机器人工作可以测试控制器是否允许或不允许平均步态和步态变异性的变化产生更经济和/或稳定的疾驰。
Understanding how animals regulate their gait during locomotion can give biological insight and inspire controllers for robots. Why animals use the gallop at the highest speeds remains incompletely explained. Hypothesized reasons for galloping include that it enables recruitment of spinal musculoskeletal structures, that it minimizes energy losses as predicted by collisional theory, or that it provides extended flight phases with more time for leg placement and hence enhances or provides necessary maneuverability [Alexander 1988 Am. Zool. 28 237–45; Ruina, Bertram and Srinivasan 2005 J. Theor. Biol. 237 170–92; Usherwood 2019 J. Exp. Zool. Part A 333 9–19; Hildebrand1989 Bioscience 39 766–75]. The latter-most hypothesis has implications in robotics, where controllers based on the concept of multistability have gained some traction. Here we examine this hypothesis by studying the dynamics of dog gait on flat and rough terrain. This hypothesis predicts that injection of noise into timing and location of ground contacts during the galloping gait by rough terrain will result in an isotropically more noisy gallop gait, centered around the gallop used on flat terrain. We find that dog gait in terms of leg swing timing on rough terrain is not consistently more variable about the mean gait, and constrain the upper limits of this variability to values that are unlikely to be biologically relevant. However the location of the mean gait indeed only shifts by a small amount. Therefore, we find limited support for this hypothesis. This suggests that achieving a target gallop gait with tight regulation is still the desired behavior, and that large amounts of variability in gait are not a desired feature of the gallop. For robotics, our results suggest that the emergent animal-environment dynamics on rough terrain do not exhibit uniformly wider basins of attraction. Future robotics work could test whether controllers that do or do not allow shifts in mean gait and gait variability produce more economical and/or stable gallops.