Design and Modeling of Soft Growing Robots
Soft and bio-inspired robots often take inspiration from the forms and behaviors of organisms found in nature. Robots have been designed that imitate caterpillars, amoebas, elephants, octopuses, and more, leveraging the adaptability of the organism's features. Yet few designs have considered the behavior of continuous, indeterminate growth seen in plants and some cells. This dissertation examines one such design, a soft robot that extends from the tip using a continuous stream of material everted by internal pressure, allowing the robot to increase in length. This robot essentially grows by adding new material at the tip. Because artificial growth as a form of movement is relatively unstudied, this thesis focuses on understanding the benefits and constraints in a system like this by looking at the design, modeling, and application of the robot.
We first looked at growth as a new degree of freedom. Unlike other forms of movement, growth is achieved by transporting new material to the tip so that it can be added to the length. To understand this difference, we develop a quasi-static model relating the driving force to the resulting growth. We show that environment friction has little to no effect on growth and that cost of material transport is a function of the previous path. Building on this model, we then demonstrate two key capabilities of growth: movement through constrained, sticky, or slippery environments and construction of usable structures to transport materials, guide tools, or create supports. Next, we consider shape change of the robot and develop a new kinematic model derived from geometric constraints that describes the robot motion and body shape given the shape of a tendon actuator. We validate the model on static and active shapes, and show how the model can be used to design actuators to match a target shape. Finally, we use the models and features learned from the earlier portions of the thesis to inform the design of soft growing robots for reconfigurable and deployable antennas. We use robotic growth to create a monopole antenna can change frequency by changing its length, and we create polarization change in a helical antenna using tendon actuation.