ABSTRACT
In this chapter, after reviewing some of the work on calibration and sensor fusion, the authors present an efficient data fusion algorithm with tolerance to multiple faults, which not only minimizes mean square error (MSE) but also keeps the precision high by bounding maximum mismatch (MM). They illustrate the performance of their algorithms on the example of two multisensor systems: a multisensor system consisting of eight STTS751 temperature sensors and a multisensor system consisting of five MMA8451Q 3-axis accelerometers. The authors also present the background and related work on calibration, sensor fusion, and fault detection in multisensor systems. They discuss screening process to detect the potentially faulty sensors in a multisensor system. Data fusion methods can also be used to detect faulty sensors and deliver fault-tolerant measurements. The authors provide information on the proposed data fusion algorithm, which minimizes MSE while bounding MM. They also provide information on experimental results.
