Integrated Magnetorheological Actuator Arrays for Robotic Hands

机器人将在不久的将来协助或与人类合作，完成重复性的任务、处理重负荷或提高精细任务的精度。然而，在实际尺寸的范围内开发出与人类肌肉性能相当的致动器系统仍然是一个重大挑战，如果不是一个重大障碍的话。例如，人手包含29个主要关节，123条韧带和34块肌肉，这还不包括神经和动脉网络。目前还没有机器人解决方案能在如此小的集成封装中提供如此高的性能。

Robots will soon be assisting or collaborating with people for repetitive tasks, heavy load handling, or improving precision on delicate tasks. However, developing actuator systems that can perform as well as human muscles within a practical size envelope remains a major challenge, if not a significant roadblock. The human hand for instance, packages 29 major joints, 123 ligaments, and 34 muscles, and this does not even account for its network of nerves and arteries. No present robotic solution offers this level of performance in such a small, integrated package. This Discovery Grant proposal aims at developing highly integrated, high-bandwidth and high-torque-density actuator arrays for collaborative robots (e.g.: robotic hand) based on magnetorheological (MR) clutches. The research builds on recent results demonstrating that MR actuators have game changing potential for aerospace applications which require large torques (e.g.: 100 Nm). However, the potential for small, highly integrated systems such as robotic hands, which require low torques (e.g.: 0.1 Nm), have yet to be studied. A first research axis of this proposal will focus on optimal hardware for small-scale, highly integrated systems. The work will develop MR clutches exploiting novel fluid behavior (e.g.: squeeze-strengthening) to boost torque output. Additive manufacturing as well as PCB-based constructions will be investigated as means to maximize packaging density while simplifying manufacturing and assembly. A second research axis of this proposal will focus on system-level design and control strategies in the context of developing a robotic hand for collaborative robotics. The work will develop a hand-like gripper using one motor feeding a multitude (e.g.: 10) of miniature MR clutches coupled to a tendon architecture to actuate each robot finger. Control schemes minimizing energy consumption and maximizing dynamic response will be developed. Finally, a fully-functional prototype will be developed and its performance characterize experimentally to assess the technological potential of the proposed approach. Altogether, the work will provide robotic technologies that will allow safe interaction between man and machine. Devices such as exoskeletons and manufacturing robots working in close collaboration with humans will become a reality that will benefit Canadians in several aspects of their lives.