A Generalized Approach for the Control of Micro- Electromechanical Relays

MEMS (Micro-Electromechanical Systems) is an area of research and applications that is becoming increasingly popular. It's mainly concerned with integrating micro-mechanical transducers with micro-electronic circuits on common substrates, traditionally silicon, through micro-fabrication. Instead of traditionally having the transducer and the communicating (or control) circuit as two separate entities, MEMS miniaturizes and combines them on a single chip, giving it several advantages, saving space, money, and increasing the sensitivity and accuracy of the integrated system. A micro-electromechanical relay is a type of MEM devices that is becoming increasingly important in a wide range of industries such as the computer industry, the medical industry and the automotive industry, to name a few. However, micro-relays, both electrostatic and electromagnetic, share a common dynamic structure that causes an unfavorable phenomenon called pull-in in which the movable electrode comes crashing down to the fixed electrode once it reaches a certain gap spacing, possibly damaging the relay and creating undesirable output effects. To eliminate this phenomenon and have better control over the switching of the micro-relays, improving transient response and output error, a feedback control scheme is desired. In this work, it is shown that voltage-controlled electromechanical micro-relays have a common dynamic structure allowing for the formulation of a generalized model. It is also shown that openloop control of MEM relays naturally leads to pull-in during closing. An attempt has been made to control the relays eliminating this phenomenon and tracking a command signal that dictates the motion of the movable electrode over time with improved transient response. In doing so, two control schemes were adopted, a Lyapunov-based and a feedback linearization-based one. Simulation results clearly show the superiority of the closed-loop control compared to the openloop one. It's also shown that the Lyapunov-based controller was limited in the extent to which it