Novel robot actuators leveraging the molecular mechanics and topology of biological muscle

目前的腿式机器人无法应对复杂的协调和控制需求，这些需求需要在现实世界、个人和协作应用中取得成功和帮助，这些应用需要在家庭、工作场所、城市或野生自然环境等非结构化环境中操作。实现现实世界中可展开机器人的一个关键限制因素是驱动和实现机器人每一个动作的执行器。在过去的几十年里，机器人驱动的缺陷已经成为一个广泛认识的问题，导致对更高效、动态和强大的驱动器的需求，以显着提高腿式机器人系统的能力，然而，现有的技术仍然不足以满足期望和必要的要求。本研究计划的主要目的是通过开发功能类似生物肌肉的致动器来引领机器人致动器设计的全新方法，该致动器通过在分子水平上的许多小规模离散致动的组合工作来产生连续的，大规模的力和位移。为了利用和放大小的子驱动器的相对较小的影响，肌肉中的力产生机制（肌球蛋白蛋白）以高频周期运行，而并联和串联的子驱动器可以实现更高的力和缩短的速度。在这个研究项目中开发的肌肉驱动器将利用类似的模块化和高度冗余的小型循环驱动机构架构来产生宏观驱动。在副致动器层面，不同的致动技术，包括更传统的电磁学和气动，或新的产生力的功能，智能材料可以用来产生往复冲程，从而驱动更大的肌肉激励致动器。这种驱动方案还直接促进了接近零阻抗的反向驾驶能力，允许利用自然系统动力学来增加峰值功率，提高运动和其他控制任务的能量效率。该功能还实现了机械透明度和被动合规，这对于要求生物力学兼容性和安全性的应用至关重要，例如力辅助或增强可穿戴机器人设备。该计划将研究不同的方法，以利用机器人执行器结构中与传统方法不同的新范式的这些和其他令人兴奋的优势。这种新颖的、以生物为灵感的设计策略具有独特的潜力，可以超越现有执行器的性能限制，同时提供新的机器人功能，这对它们在更个性化和近端服务应用中的成功至关重要。随着未来机器人设计需求的扩大，以满足动态和不可预测的现实世界中的新挑战，这项研究计划为扩展和推进机器人能力的新方向提供了重要途径。

Current legged robots are unable to cope with the demands of complex coordination and control needed to succeed and assist in real-world, personal, and collaborative applications which require operation in unstructured settings such as homes, workplaces, cities, or wild natural environments. A crucial limiting factor to the realization of real-world deployable robots are the actuators which drive and enable their every movement. The deficiency in robotic actuation has been a widely recognized problem over the last few decades leading to demand for more efficient, dynamic, and powerful actuators to dramatically improve the capabilities of legged robotic systems, however, existing technology remains insufficient for desired and necessary requirements. The principle aim of this research program is to lead an entirely new approach to robotic actuator design by developing actuators that function much like biological muscles, which generate continuous, large-scale forces and displacements through combined work of many small-scale, discrete actuations at the molecular level. In order to leverage and amplify the relatively small effect of the small sub-actuations, the force generating mechanisms in muscle (myosin proteins) operate in high-frequency cycles, and while parallel and series arrangement of sub-actuators enables higher forces and shortening velocities. Muscle inspired actuators developed in this research program will leverage a comparable modular and highly redundant architecture of small-scale, cyclic actuation mechanisms to produce macro-scale actuation. At the sub-actuator level, different actuation techniques, including more conventional electromagnetics and pneumatics, or new force-generating functional, smart materials may be used to produce reciprocating strokes which can drive a larger muscle-inspired actuator. This actuation scheme also directly facilitates near zero-impedance back-drivability, allowing the exploitation of natural system dynamics for increasing peak power and improving energetic efficiency in locomotion and other control tasks. This feature also enables mechanical transparency and passive compliance essential for applications demanding biomechanical compatibility and safety, such as force-assisting or augmenting wearable robotic devices. This program will investigate different methods for exploiting these and other exciting advantages of a new paradigm in robotic actuator construction that diverges from traditional approaches. This novel, bioinspired design strategy holds unique potential to surpass the performance limits of existing actuators while simultaneously delivering new robot capabilities key to their success in more personal and proximal service applications. As future robot design requirements expand to meet new challenges in the dynamic and unpredictable real-world, this research initiative offers a vital pathway to extending and advancing the limits of robot capabilities in new directions.