我们最近发表在《Mechanism and Machine Theory》的期刊论文中，以Bennett机构为代表，探索了过约束机构在机器人足式肢体设计方面的工程潜力，用于研制具有全向行走能力的机器腿，搭建了一台性能优越的四足机器人原型样机，开拓了以过约束机器人学为核心的全新发展方向。该论文第一作者顾宇平是南方科技大学与香港大学联合培养博士研究生，合作作者包括南方科技大学硕士研究生冯世豪和郭宇芹、设计学院助理教授万芳、机械与能源工程系讲席教授戴建生、香港大学计算机系副教授潘佳，通讯作者是南方科技大学机械与能源工程系助理教授宋超阳
Our Study Investigates the Use of Overconstrained Linkages for Omni-Directional Locomotion in Legged Robots. We’ve explored the untapped potential of overconstrained linkages, specifically the Bennett linkage, in creating robotic legs for omni-directional locomotion. Our research resulted in a prototype quadruped robot with impressive performance, and we’re now paving the way for the future of overconstrained robotics.
doi: https://doi.org/10.1016/j.mechmachtheory.2022.105018
在机器人领域，四足步态一直以来都备受瞩目，其肢体机构设计也是技术进步的关键。然而，在过约束机构中作为最具代表性的经典空间机构，蕴藏着未被发掘的工程应用潜力，我们的研究着手探索过约束机构用于足式机器人肢体设计这一尚未开拓的前沿领域。我们的研究深入研究以Bennett机构为代表的过约束机构的工程设计，探索一种在四足机器人领域具有全向地面运动能力的足式肢体创新设计方法。我们的目标是在机器人腿部领域确立这类空间机构在机构学理论和工程设计方面的应用潜在。
In the world of robotics, legged locomotion has always been a subject of fascination and technological advancement. However, there’s an untapped potential lying within overconstrained linkages, and in our research, we set out to explore this uncharted territory. Our study delves into the design of overconstrained linkages, specifically the Bennett linkage, as a novel approach to creating robotic legs with omni-directional ground locomotion capabilities. The goal was to establish the theoretical foundations and engineering advantages of these unique linkages in the realm of legged robotics.
重新定义机器人腿｜Reinventing Robotic Legs
我们的研究采用了一种全新的腿部设计视角，提出了一种空间链杆和关节的参数化设计，允许通过3D打印来制造这些过约束机构。这些机构配备了同轴驱动，将链杆连接到一对伺服执行器，形成一个可重构的腿部模块。
Our research takes a fresh perspective on leg design, breaking free from the conventional norms. We propose a parametric design for the spatial links and joints, allowing for the fabrication of overconstrained limbs via 3D printing. These limbs are then equipped with coaxial actuation, and the linkage is attached to a pair of servo actuators to form a reconfigurable leg module.
多目标优化｜Multi-Objective Optimization
为了优化设计参数并最大化性能，我们采用了多目标优化方法。我们的分析考虑了可操纵性指标和力传递，确保这些过约束机构能够实现全向地面运动。
To refine the design parameters and maximize performance, we employ a multi-objective optimization approach. Our analysis takes into consideration manipulability metrics and force transmission, ensuring that these overconstrained legs can achieve omni-directional ground locomotion.
原型样机｜Prototype Success
我们成功搭建了首台过约束足式机构原型四足机器人样机，该机器人展示出令人印象深刻的全向运动能力，并仅使用了极少数量的执行器，就实现了最小的转弯半径，仅为其身体长度的0.2倍。
Our efforts resulted in the creation of a prototype quadruped robot, equipped with these novel overconstrained legs. This robot showcased impressive omni-directional locomotion capabilities and achieved a minimal turning radius of just 0.2 times its body length while using a minimal number of actuators.
不仅限于Bennett机构｜Beyond the Bennett Linkage
我们的探索不仅限于Bennett四杆机构；我们扩大了研究范围，包括所有的过约束5R和6R机构。这一扩展为未来的过约束机器人开辟了令人兴奋的可能性。
We didn’t stop at the Bennett linkage; we expanded our exploration to encompass all overconstrained 5R and 6R linkages. This expansion opens up exciting possibilities for the future of overconstrained robotics.
为前行指路｜Pioneering the Path Forward
我们的研究是机器人腿部设计领域的一项开创性工作。我们充分利用过约束机构，尤其是Bennett机构的独特能力，创造了一类在全向运动方面表现出色的机器人腿部。我们的研究发现不仅仅是效率的提升；它代表了机器人领域的一次飞跃，对未来的机器人肢体设计发展具有广泛的指导意义。
In conclusion, our research is a pioneering effort in the field of robotic leg design. We’ve harnessed the unique capabilities of overconstrained linkages, particularly the Bennett linkage, to create a new class of robotic legs that excel in omni-directional locomotion. The findings from our study are not just limited to efficiency; they represent a leap forward in the world of legged robots, with broad implications for the future of robotics.
揭示未来｜Unraveling the Future
随着我们的前行，我们承认还有许多待探索的领域。过约束几何形状可能在控制和算法设计方面存在挑战，但我们相信现代计算工具和机器学习技术可以克服这些困难。我们未来的工作将集中在具备过约束机构的四足机器人的动力学和优化控制上，以及开发具有人工智能能力的过约束机器人设计算法。我们在过约束机构领域的探索才刚刚开始。
As we move forward, we acknowledge that there is much more to explore. Overconstrained geometry may present challenges in terms of control and algorithm design, but we believe modern computational tools and machine learning can overcome these hurdles. Our future work will focus on the dynamics and optimized control of quadrupeds equipped with overconstrained linkages, as well as the development of generative design algorithms, leveraging the power of artificial intelligence in the creation of overconstrained robots. Our journey into the world of overconstrained robotics has only just begun.
Yuping Gu, Shihao Feng, Yuqin Guo, Fang Wan, Jian S. Dai, Jia Pan, and Chaoyang Song* (2022). “Overconstrained Coaxial Design of Robotic Legs with Omni-directional Locomotion.” Mechanism and Machine Theory, 176:105018.
doi: https://doi.org/10.1016/j.mechmachtheory.2022.105018
