Data validation and confidence of self-validating multifunctional sensor

A novel strategy using fuzzy logic coupled with polynomial predictive filters is proposed for the online measurements validation and confidence evaluation of multifunctional sensor. The predictive errors are generated by comparing the predictive outputs with actual measurements, and then some fuzzy logic rules are established to obtain the confidence evaluation, which reflects the measurements quality to some extent. Incorrect data under faults can be picked up by using the proposed method to ensure the correctness of measurements. A good validated value will replace the wrong output by a reasonable predictive value or the inner potential correlation. The data validation is accomplished and associated with a qualitative value for its reliability. The methodology presented is illustrated in a real experimental system of multifunctional sensor and results demonstrate that the proposed strategy can provide a good solution to the data validation and confidence evaluation of multifunctional sensor.