Self-compensation to build reconfigurable measurement systems

Intelligent sensors are the core components of reconfigurable measurement systems (RMS). The intelligent sensor's generic functions are compensation, processing, communication, validation, integration and data fusion. Their design involves using self-adjustment (or compensation) algorithms for eliminating or at least diminishing major types of error, such as offset, gain variation, and non-linearity, with good accuracy [1]. In addition, the design must make the readjustment process as simple as possible [2]. A methodology for designing intelligent sensors with selfcompensation which can be reconfigured to measure any variable is presented in this paper. Additionally, we analyze several compensation techniques using numerical algorithms and one based on artificial neural networks theory. The methodology is applied to reconfigure intelligent sensors for temperature and distance measurements.