Genetic Algorithm Optimization to Maximize Sensitivity of Triboelectric MEMS Accelerometers

Abstract

Abstract. A genetic algorithm is used to optimize the sensitivity of a triboelectric MEMS accelerometer fabricated using CMOS-compatible processes. The optimization focuses on the structural parameters of an awl-shaped serpentine microspring (ASSM) connected to the proof mass. Constraints are defined by microfabrication standards and the available chip area. Three different constraint problems are considered, and the algorithm yields designs with maximized sensitivity in each case. Simulations under various sinusoidal vibration scenarios show that the optimized designs significantly outperform the previous version, achieving a 2.72× improvement at 1000 Hz and up to 20× at 500 Hz.