A skeletal muscle consists of a number of independent muscle fibers, each producing a mechanical output efficiently in a compact body. Collective behaviors of a bundle of muscle fibers exhibit unique features and functionality that today's actuators are unable to produce. Unlike conventional electric motors, for example, muscles are mechanically flexible and adaptable to load conditions. Muscles can vary effective stiffness and output impedance in a wide range. Furthermore, a variety of muscles having a wide range of size, power, and degrees of freedom can be built from the same module, i.e., the muscle fiber.
Inspired by this skeletal muscle architecture, this project investigates cellular artificial muscles using actuator materials, in particular, PZT (Lead Zirconate Titanate). Like a muscle fiber, a PZT cellular unit is an independent unit that is designed for optimal efficiency and functionality. Using an effective flexure, each PZT cell can produce large displacement and force comparable to a skeletal muscle fiber. Such an optimal unit is inevitably small to drive a large load. Arranging the PZT cells in series, parallel, and antagonistic configurations, a bundle of PZT units collectively exhibit unique features and functionality that a single bulk PZT actuator cannot produce. These include: A). Variable stiffness: Switching individual cellular units ON or OFF creates a significant change to the aggregate stiffness of the multi-cellular system. B). Tailored force-displacement characteristics: By exploiting the ON-OFF nonlinearity of individual units and superimposing the multitude of the nonlinear functions, we can tailor the aggregate force-displacement curve to task goals and environment conditions. This allows us to physically implement the "work loop" characteristics of animal motion at the actuator level. C). Variable resonant frequencies: The multi-cellular PZT actuator forms a network of mass-spring systems due to the combined effect of a flexure and PZT stack at each unit. This multi d.o.f. system has resonant frequencies at which the output displacement becomes much larger than its static stroke, making cyclic motion, such as flapping and running, very efficient. Furthermore, ON-OFF switching of individual units causes a change to the mass distribution and thereby makes the resonant frequency highly tunable. D). Energy harvesting: The new actuator is completely backdriveable having negligibly small friction. Low friction and tunable resonance capability enables a multi cell array to capture energy at a resonant frequency that is the most effective for the environmentally imposed forcing.
A car is powered by various types of engines having 4, 6, and more cylinders. As the number of cylinders increases, we get more power, less vibration, and smoother ride. A collection of multiple cylinders can produce desired features that a single bulky cylinder cannot. In this project we studied collective behaviors of piezoelectric actuators, a type of electric actuator producing force in response to applied voltage, to explore unique features that a single piezo actuator cannot produce. We called this a multi-cellular piezo actuator, and have obtained three interesting outcomes. First, similar to the car engines, we can build a multi-cellular piezo actuator with 4, 6, and more "piezo cylinders", collectively pushing gear teeth of a rod, like rotating an engine crank shaft. Arranging these piezo cylinders in a particular configuration, we can harmonize the movement of the actuator. Although each piezo cylinder produces jerky motion due to pronounced nonlinear force generation, the multiple piezo cylinders can cancel out those undesirable characteristics, producing a smooth, harmonized motion. Second, the new multi-cellular piezo actuator can do more than simply producing force and displacement. Similar to biological muscles, it can be a soft, compliant muscle or a stiff, rigid one depending on the situation. We have found effective algorithms to vary the stiffness of the actuator simply by coordinating the multiple piezo cylinders in a particular manner. In other words, we can virtually change the spring by software, through which the force is generated. Furthermore, we can vary the resonant frequency so that large amplitude of cyclic motion can be produced with small energy consumption. These features are similar to animal’s running, where legs are resonating at a particular frequency and compliant feet can absorb impacts from the ground. Third, like cylinders of a car engine, the size of piezo cylinders is an important factor that determines output power, efficiency, and reliability. It should be neither too large nor too small. We have made careful analysis to find an optimal form factor of piezo cylinders, by taking into account basic characteristics of piezoelectric materials, a strain amplification mechanism, and structural material properties. To attain the best performance, we should build piezo cylinders with the optimal form factor, and use multiple cylinders to construct a larger actuator. This means that modular design is effective for improving actuator power and efficiency. Furthermore, modular design allows for mass production, and various sizes of actuators can be produced from the same modules. The modular architecture also improves reliability; although some piezo cylinders are not functional, other cylinders can fill in to continue a mission. Finally, we have found that the new piezo actuator will be particularly useful for those applications where constant loads must be borne for a long time. As a capacitive actuator, piezo actuators do not consume any energy for holding a large load. In other words, they have a built-in braking capability, which consumes no energy for bearing a load. See more details in our publications.
