M.Sc Thesis

M.Sc StudentTibi Gal
SubjectDesign of Thin, Multi-Layered, Electrically Driven, Thermal
Expansion Based Actuators for Soft Robotic
DepartmentDepartment of Autonomous Systems and Robotics
Supervisor ASSOCIATE PROF. Amir Degani


Soft and continuum robots, can achieve complex configurations of their entire structure. Such capabilities can be useful in different tasks such as obstacle avoidance or whole arm grasping and manipulation. By bonding micron-thin polymer layers with an electrically conductive layer, we can utilize the bi-metal effect to create a soft, thin, electrically activated thermal actuator. These actuators’ softness is gained from their thickness rather than the properties of their materials. The performance of these actuators, that is, curvature to input voltage ratio, depends on the material's properties and geometry. Although analytic models of multi-layered sensor designs, based on the bi-metal effect are available, they do not consider the electro-thermal relation present in Electro Thermal Actuators (ETA). In prior works, the three-layered (tri-layer) design was overlooked due to similar performance relative to bi-layered actuators. However, this is untrue for ETAs. Unlike the bi-layer case, where both layers need to be as thin as possible to achieve high performance, the ETA's conductive layer must be thick to lower the electrical resistance to produce high temperature in low voltage.

In this work, we optimize the bi-layer actuator curvature sensitivity based on our analytic model with and without the thermal modeling, for a single module segment. The bi-layer optimization not only predicts the maximal value for the bi-layer ETA performance, but also reveals its relatively low maximal performance. We reproduce the former analytic model of the single tri-layer module while including the electro-thermal relation and validate our analytic results experimentally. Based on our model, we optimize the tri-layer ETA curvature sensitivity and achieve a significant improvement of more than 3000% over the optimal bi-layer ETA in curvature. We continue by analytically modeling the ETA tip force based on a cantilever model and experimentally validate our results. Finally, we present three, experimental, proof-of-concept, soft robot prototypes that utilize the tri-layer ETA: single segment "U-shape" crawler, multi-segment serial "snake" and a multi-segment parallel "quadrapod". These robots utilize the tri-layer ETA not only as an actuator but also as their main structure. Each robot design is aimed for a different soft robotics application that utilize the soft and continuum structure such as locomotion and manipulation.