The modelling of micro electro mechanic switches using HFSS and artificial neural networks

This thesis is about some kinds of RF MEMS, which is a newly developing technology. The thesis contains mixture of mechanical, electronical and materials fields. In the first section, a short entrance is made to RF MEMS switches and microelectronical RF switch technologies after telling the reasons of using RF MEMS. In the second part, RF MEMS switches are investigated in detail and detailed working techniques and production steps are presented. In the third part, microwave characteristics and S parameters of the RF MEMS parallel and serial switches that we produced in HFSS program are calculated. The coming section includes some necessary information about artificial neural networks technology. In the last section, we applied the variables that are produced through the HFSS program in artificial neural networks according to different network models and training algorithms and used graphics to compare the concluded results with the expected results. Keywords: RF MEMS, switch, artificial neural networks xvii