The self-validating (SEVA) sensor carries out an internal quality assessment, and generates, for each measurement, standard metrics for its quality, including online uncertainty. This paper discusses consistency checking and data fusion between several SEVA sensors observing the same measurand. Consistency checking is shown to be equivalent to the maximum clique problem, which is NP-hard, but a linear approximation is described. A technique called uncertainty extension is proposed which causes a smooth reduction in the influence of outliers as they become increasingly inconsistent with the majority.
- Maximum clique problem
- self-validating (SEVA) sensors
- sensor fusion