协调的附肢能积累更多能量以在地面实现自扶正
Coordinated appendages accumulate more energy to self-right on the ground
AI总结：
本文通过建立混合动力学模板，揭示了翅膀-腿部协调相位通过改变机械能预算影响自扶正成败，并据此提出改进机器人自扶正性能的控制策略。
摘要：
动物和机器人在翻倒后必须在地面上实现自扶正。盘状蟑螂将翅膀压向地面，试图通过翻筋斗的方式动态自扶正。然而，由于这一动作非常费力，该动物常常无法克服势能障碍，并会不断进行尝试。在这一过程中，该动物会挥动腿部，其侧向扰动最终使其侧滚以实现自扶正。我们之前的工作开发了一种受蟑螂启发的机器人，能够进行腿部辅助、翅膀驱动的自扶正，而一项机器人仿真研究揭示，该策略的结果敏感地依赖于翅膀-腿部协调（以其运动之间的相位来衡量）。在此，我们通过开发一个模板来对由不连续接触和驱动所产生的复杂混合动力学进行建模，从而进一步阐明其原因。我们使用该模板计算了身体实现自扶正所必须克服的势能障碍、翅膀推地和腿部挥动所产生的机械能贡献，以及翅膀-地面碰撞所导致的机械能耗散。该模板揭示，翅膀-腿部协调（相位）通过改变机械能预算来强烈影响自扶正结果。协调良好的附肢运动（良好相位）比协调较差的运动（不良相位）能积累更多机械能，从而更好地克服势能障碍，更成功地实现自扶正。最后，我们展示了该模板的实际用途，即用于预测一种新的控制策略以进一步提高自扶正性能，并为机器人设计提供信息。
英文摘要：
Animals and robots must right themselves after flipping over on the ground. The discoid cockroach pushes its wings against the ground in an attempt to dynamically self-right by a somersault. However, because this maneuver is strenuous, the animal often fails to overcome the potential energy barrier and makes continual attempts. In this process, the animal flails its legs, whose lateral perturbation eventually leads it to roll to the side to self-right. Our previous work developed a cockroach-inspired robot capable of leg-assisted, winged self-righting, and a robot simulation study revealed that the outcome of this strategy depends sensitively on wing-leg coordination (measured by the phase between their motions). Here, we further elucidate why this is the case by developing a template to model the complex hybrid dynamics resulting from discontinuous contact and actuation. We used the template to calculate the potential energy barrier that the body must overcome to self-right, mechanical energy contribution by wing pushing and leg flailing, and mechanical energy dissipation due to wing-ground collision. The template revealed that wing-leg coordination (phase) strongly affects self-righting outcome by changing mechanical energy budget. Well-coordinated appendage motions (good phase) accumulate more mechanical energy than poorly-coordinated motions (bad phase), thereby better overcoming the potential energy barrier to self-right more successfully. Finally, we demonstrated practical use of the template for predicting a new control strategy to further increase self-righting performance and informing robot design.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
