Safety-critical systems often incorporate fault-tolerance techniques in their design, particularly the technique of redundancy. The idea of consistency checking found in redundancy techniques can be more broadly applied by using knowledge about system parameters and their relationships. Here we present a model of data fusion, which detects failures and provides good estimates of plant parameters by checking sensor data for consistency. We illustrate our approach with a boiler system example, proving that the water level in the boiler is always within its safe range.
|Journal||Journal of Computer and Software Engineering|
|Publication status||Published - 1995|