Optimal Design of Miniature Flexural and Soft Robotic Mechanism

Compliant mechanisms are flexible structures that utilize elastic deformation to achieve their desired motions. Using this unique mode of actuation, the compliant mechanisms have two distinct advantages over traditional rigid machines: (1) They can create highly repeatable motions that are critical for many high precision applications. (2) Their high degrees-of-freedom motions have the potential to achieve mechanical functionalities that are beyond traditional machines, making them especially appealing for miniature robots that are currently limited to only having simple rigid-body-motions and gripping functionalities. Unfortunately, despite the potential of compliant mechanisms, there are still several key challenges that restrict them from realizing their full potential. To facilitate this discussion, we first divide the compliant mechanisms into two categories: (1) the stiffer flexural mechanisms that are ideal for high precision applications, and (2) the more compliant miniature soft robots that can reshape their geometries to achieve highly complex mechanical functionalities.;The key limitation for existing flexural mechanisms is that their stiffness and dynamic properties cannot be optimized when they have multi-degrees-of-freedom. This limitation has severely crippled the performance of flexural mechanisms because their stiffness and dynamic properties dictate their workspace, transient responses and capabilities to reject disturbances. On the other hand, miniature soft robots that have overall dimensions smaller than 1 cm, are unable to achieve their full potential because existing works do not have a systematic approach to determine the required design and control signals for the robots to generate their desired time-varying shapes. This thesis addresses these limitations by developing two design methodologies: The first methodology is developed for synthesizing optimal flexural mechanisms, while the second is a universal programming method for designing and controlling miniature soft robots that can generate desired time-varying shapes. The first methodology is implemented by first employing a kinematic approach to select suitable parallel-kinematic configurations for the flexural mechanisms, i.e. determine their required number and type of sub-chains. Subsequently, a structural optimization approach is used to automatically synthesize and optimize the sub-chains' structural topology, shape and size. In order to integrate the kinematic and structural optimization approaches, a new topological optimization algorithm termed the mechanism-based approach has been created. In comparison with existing algorithms, a notable benefit for the mechanism-based approach is that it can eliminate infeasible solutions that have no physical meanings while having a flexible way to change its topology during the optimization process. This algorithm has been shown to be able to develop various devices such as a mu-gripper, a compliant prismatic joint, and a compliant prismatic-revolute joint. A generic semi-analytical dynamic model that can accurately predict the fundamental natural frequency for compliant mechanisms with parallel-kinematic configurations has also been developed for the proposed integrated design methodology.;The effectiveness of the first methodology is demonstrated by synthesizing a planar-motioned X-- Y-- thetaz flexure-based parallel mechanism (FPM). This FPM has a large workspace of 1.2 mmx 1.2 mmx6°, bandwidth of 117 Hz, and translational and rotational stiffness ratios of 130 and 108, respectively. The achieved stiffness and dynamic properties show significant improvement over existing 3-degrees-of-freedom, centimeter-scale compliant mechanisms that can deflect more than 0.5 mm and 0.5°. These compliant mechanisms typically only have stiffness ratios and bandwidth that are less than 50 and 45 Hz, respectively. The stiffness and dynamic properties of the optimal FPM were validated experimentally and they deviated less than 9% from the simulation results.;Our second design methodology is a universal programming methodology that can magnetically program small-scale materials to generate a series of desirable time-varying shapes. More specifically, this method allows scientists and engineers to automatically generate the required magnetization profile and actuating magnetic fields for the robots. The effectiveness of the second methodology is demonstrated via creating various miniature devices that are difficult to realize with existing technologies, and this includes a spermatozoid-like undulating swimmer and an artificial cilium that could mimic the complex beating patterns of its biological counterparts. In comparison to existing previous works that rely solely on human intuition and can only program these materials for a limited number of applications, our universal methodology has the potential to allow scientists and engineers to fully capitalize shape-programming technologies.;We envision that the first methodology can inspire engineers to develop a variety of high precision machines that have optimal performances while the second methodology has paved the way for novel miniature devices that are critical in robotics, for smart engineering surfaces or materials, and for biomedical devices. (Abstract shortened by ProQuest.).