Kristofer S.J. Pister (Advisor)

Generating comprehensive task schedulers and specialized low-level controllers for robots in complex environments often requires system and environment knowledge, which can result in long design times. Real-life experiments for the development, tuning, and validation of such controllers can be costly for a microrobot. To address these issues, we propose the distillation of computation-intensive expert policies (that use Model-Based Reinforcement Learning) into small sequence models trained auto-regressively on model predictive control (MPC) trajectories. Previously, we modeled the long-...

Our goal is to build a family of autonomous silicon robotic insects with actuating, computing, and power capabilities integrated. A silicon-on-insulator (SOI) device is used to house all three components. These robots use electrostatic actuators driving silicon linkages, all fabricated in the device layer of the wafer. By using electrostatic actuation, these actuator linkage systems have the advantage of being low power compared to other methods of actuation on microscale granting robot autonomy through low-power energy harvesting. Computation and communication are carried out with Single...