Optimal Design of a Diagnosis System for Faulty Sensors

A problem of optimal design of a diagnosis system for faulty sensors is formulated as a problem of minimizing the cost required to construct the allocation of sensors which realizes the required upper limit of probability of false-alarm, lower limit of systematic error and upper limit of probability of mis-alarm for the given set of sensors to be diagnosed. A concept of critical detectability is introduced corresponding to the degree of freedom of the sensor system and a method is presented to solve this problem by use of this concept and the branch-and-bound algorithm.