Challenges in Design and Validation of MEMS Based Smart Sensor Systems

The emergence of MEMS and decreasing costs for microelectronics and wireless communication enable new fields of use for sensors. For instance, instead of temporal instrumentation with high performance accelerometers for condition monitoring of a machine, a permanent installation of low-cost sensors, or even a widespread sensor network, becomes feasible. Consequently, the development and validation process for such sensors has to integrate more requirements derived from the target application. This means, rather than qualifying a sensor as a high-precision measurement device for many potential use cases, the design goal is a sensor system that fits a limited number of use cases and provides sufficient precision. Also, in the case of smart sensors, which incorporate signal analysis or even artificial intelligence, validation of hardware and software must be integrated. This is particularly challenging, when adaptative or learning algorithms are implemented into the sensor system. The paper gives an overview of trends in industrial MEMS applications, the respective testing and validation strategies. Proposals for improved procedures are discussed, using current work on sensor developments.