Liwei Lin (Advisor)

We present a systematic MEMS structural design approach via a “trial-and-error” learning process by using the deep reinforcement learning framework. This scheme incorporates the feedbackfrom each “trial” to obtain sophisticated strategies for MEMS design optimizations. Disk-shaped resonators are selected as case studies and three remarkable advancements have been realized:1) accurate overall performance predictions (97.9%) via supervised learning models; 2) efficientMEMS structural optimizations to guarantee targeted structural properties with an excellent generation accuracy of 97.7%; and...

In this work, we report a thermoelectric energy harvester based on an ion-conductive ionogel withthree distinctive features as compared with the state-of-art systems: (1) selfhealable in both thematerial and device operations under large deformations; (2) ultrahigh Seebeck coefficient withthe measured value of 7068 μV/K for the material, and 298 μV/K for a prototype energy harvesterin operations; and (3) stable performances under large bending states with little...

Advanced tactile feedback systems are important tools in the field of human–machine interfaces. In this work, an airflow-assisted corona charging process is utilized to charge films made of electret material for the construction of a sandwich-structured flexible actuator system. With a voltage as low as 20 V, this flexible actuator can stimulate skin sensations for basic tactile feedback functions. Under a driving voltage of 200 V, the system can generate an output force of ≈55 mN, which is larger than that of the output force by cellphones under the vibration mode. Utilizing these...