Design of a pneumatic soft robotic actuator using model-based optimization

In this thesis, the design and optimization process of a novel soft intelligent modular pad (IntelliPad) for the purpose of pressure injury prevention is presented. The structure of the IntelliPad consists of multiple individual multi-chamber soft pneumatic-driven actuators that use pressurized air and vacuum. Each actuator is able to provide both vertical and horizontal motions that can be controlled independently. An analytical modeling approach using multiple cantilever beams and virtual springs connected in a closed formed structure was developed to analyze the mechanical performance of the actuator. The analytical approach was validated by a finite element analysis. For optimizing the actuator\u27s mechanical performance, firefly algorithm and deep reinforcement learning-based design optimization frameworks were developed with the purpose of maximizing the horizontal motion of the top surface of the actuators, while minimizing its corresponding effect on the vertical motion. Four optimized designs were fabricated. The actuators were tested and validated experimentally to demonstrate their required mechanical performance in order to regulate normal and shear stresses at the skin-pad interface for pressure injury prevention applications