Designing Soft Continuum Robots for Sensor-Enabled Control

Soft robots, which leverage flexible, stretchable, and smart materials, are relevant to numerous applications that traditional robots struggle with, such as search-and-rescue, human-robot interaction, and exploration. Since soft robots are composed of soft materials, they are inherently more robust to impacts and falls than their rigid counterparts. Additionally, soft structures are inherently safer for human-robot interaction. While clever use of soft materials offers many advantages, it complicates the control of soft robotic systems. Many of the control strategies that have been established for traditional robotic systems cannot be readily used for soft systems due to the difficulties in modeling soft systems. These control strategies require sensory feedback that can reliably provide the state of the system. However, obtaining sensory feedback from soft robotic systems is non-trivial. It has only been in the past few years that soft sensor technology has begun integrating with soft structures to try to provide the proprioceptive data needed to implement control strategies. This thesis focuses on the use of sensory feedback to compensate for the complex behavior of a soft system. In order to accomplish sensory feedback, multiple soft sensor types were investigated and integrated into soft robotic systems. Simplified analytical models were developed to help design soft systems and to interpret the state from the collected sensory data for use in feedback controllers. These simplified models also allowed the implementation of feedforward controllers. Additionally, this body of work demonstrates how sensory feedback can be used to inform feedforward controllers of certain model parameters.