Approach to the validation function of intelligent sensors based on error´s predictors

In order to eliminate measurement error propagation, an important characteristic of intelligent sensors is the validation function. With this, the intelligent sensor should be capable of noticing if its measure is right or not. A lot of literature about validation is available, but most of the time, researchers are solving a particular problem and exist lack of information to evaluate the feasibility to implement validation in small devices with low computational capability. In this paper the description of one validation function based on error predictors is presented. Using real data and based on error predictor, the intelligent sensor measured can be validated. The performance of two kinds of predictors is evaluated. The advantage of this proposal is that one method with low computational requirements can be selected.