Prediction and fault detection of environmental signals with uncharacterised faults

Many signals of interest are corrupted by faults of an unknown type. We propose an approach that uses Gaussian processes and a general "fault bucket" to capture a priori uncharacterised faults, along with an approximate method for marginalising the potential faultiness of all observations. This gives rise to an efficient, flexible algorithm for the detection and automatic correction of faults. Our method is deployed in the domain of water monitoring and management, where it is able to solve several fault detection, correction, and prediction problems. The method works well despite the fact that the data is plagued with numerous difficulties, including missing observations, multiple discontinuities, nonlinearity and many unanticipated types of fault. Copyright © 2012, Association for the Advancement of Artificial Intelligence. All rights reserved