Date of Award
Doctor of Philosophy (PhD)
Mechanical Engineering & Applied Mechanics
A conventional robot with fixed form cannot change its morphology to respond to changes in its environment or suit a specific task. In contrast, a self-reconfigurable modular robot can rearrange its modules to adapt to a task or its environment. Module miniaturization would enable applications requiring fine resolution such as dexterous manipulation and accurate three-dimensional physical representation. As the module's actuators limit its minimum size, we present two novel methods for enabling module miniaturization: external actuation and dielectric elastomer actuation.
With external actuation, modules reconfigure using forces provided by a controlled environment rather than internal actuators. The environment imposes deterministic forces on the modules, enabling reliable reconfiguration in known time. This thesis presents two external actuation methods: inertial forces and forces due to gravity. In the inertial case, we prove that using experimentally verified module motion primitives allows a configuration discretized into module groups to realize arbitrary configurations. In the gravity case, we demonstrate a system with 14mm length-scale units using external actuation for module motion and bonding.
External actuation is suitable for arbitrary shape formation in the presence forces from a controlled environment. Achieving field modular robots comprising miniature modules requires an efficient actuation technology with scale-invariant performance. We investigate dielectric elastomer actuation as it has the potential to achieve large actuation strain and large work and torque output per mass. We present the design and empirical characterization of a dielectric elastomer actuated antagonistic bender module and demonstrate several modular robot arm configurations.
White, Paul J., "MINIATURIZATION METHODS FOR MODULAR ROBOTICS: EXTERNAL ACTUATION AND DIELECTRIC ELASTOMER ACTUATION" (2011). Publicly accessible Penn Dissertations. Paper 381.