Modeling and design of fiber reinforced pneumatic actuators for soft robotics applications

Soft robotics is an emerging branch of robotics where the objective is to make robot structures soft and compliant. These robots trade accuracy and load bearing ability with adaptability and safety. Such robots require soft actuators to generate force and/or motion. Fiber reinforced pneumatic actuators are one such example of soft actuators that are commonly used in soft robots. These actuators have the advantages of simple fabrication, high power to weight ratio, and an ability to produce a variety of deformation behavior. Fiber Reinforced Elastomeric Enclosures, abbreviated as FREEs, are one such embodiment of fiber reinforced actuators that have two sets of inextensible fibers bonded on to an elastomeric tube. The orientation of these fibers determine the deformation behavior of FREEs, which include contraction, extension, rotation, bending, and spiral motion. This dissertation deals with the mathematical modeling of FREEs and its use in designing soft robots. Based on drawing parallels with inflatables, a mathematical model that solves a constrained volume maximization (CVM) problem to determine the deformation behavior of FREEs is presented. This model can predict the kinematics and kinetostatics of FREEs for any actuation pressure and loading conditions. This model is computationally simpler than the current models in literature and can also capture variations in both fiber winding and actuator geometry. The model is validated using multiple prototypes of FREEs spanning a range of fiber angles. There is a need for design methods that can systematically guide a soft roboticist towards choosing or designing an actuator for desired specifications. Towards this, a shape matching problem for FREEs is presented in this dissertation, where the objective is to design a FREE such that it matches the shape of a target design curve upon actuation. The design method uses piecewise smooth arcs or helices to approximate a planar or spatial curve, respectively. Following which, it solves an inverse of CVM problem to determine the actuator parameters. A nondimensional chart based approach is also shown for designing a FREE for matching two different planar curves. Multiple example curves are used to demonstrate and validate the design method. This method can be used for designing actuators for wearable robots, assistive devices as well as locomotion robots. Two specific applications of soft robots are shown in this dissertation. The first one is a soft orthotic sleeve for Lofstrand crutch users. This sleeve helps in distributing the palm load to the forearm, thereby reducing the risk of wrist injury as well as improving the wrist posture. The second application is of a pipe climbing soft robot that uses bending and extending type actuators to generate motion on the outside of a pipe. This robot can have applications in nuclear and oil and gas industries for pipe inspection and maintenance