ANN based modeling, testing and diagnosis of MEMS
Vanco B. Litovski, Miona Andrejević Stošović, Mark Zwoliński · 2005
New concepts of simulation, testing and diagnosis of MEMS are proposed, intended to boost the time to market and dependability of such systems. Black-box modeling of nonelectronic parts is introduced using artificial neural networks, so enabling radically faster simulation without concurrent algorithms and parallel computation. A lumped model of the capacitive transducer, being the part of a micro-electro-mechanical capacitive pressure sensing system, is created using an ANN. Faults are then introduced to the sensing system and simulation of the fault-free and faulty circuits are demonstrated.