Metaheuristics in the automated design of CMOS-MEMS sensors for planetary exploration

CMOS-MEMS is a micro-fabrication platform in which microsystems can be monolithically developed along with signal processing stages out of a formerly conventional CMOS ASIC chip. This is achieved by means of adding a variety of materials to (or removing them from) the previously fabricated CMOS structures. In typical MEMS sensing applications (i.e. inertial measurement, gas detectors) as well as in Integrated Circuits, the designer deals only with geometry-related and topological design issues. Bio-inspired meta-heuristics such as Genetic Algorithms are proposed in order to assist the microsystems developer to achieve a desired performance goal at particular environmental and power consumption conditions, optimizing design parameters by computational means. Solid-state electronic sensors such as gas detectors, while on Earth are usually intended to operate at certain temperature and particle concentration ranges, in space or planetary scenarios the microsystem would need to outperform low particle concentrations and harsh environmental conditions. In conventional CMOS fabrication, the design rules restricts what materials are allowed and how they interconnect to form electronic and mechanic structures, after properly modelling of the objective function (design goals) in terms of the topological layout, an automated parameter selection via evolutionary computation techniques might be implemented meeting the defined optimality criteria.