Multisensor information fusion of pulsed GTAW based on improved D-S evidence theory

Abstract

The D-S evidence theory is a method widely used in information fusion technology, but it cannot solve the conflict evidence problem. This paper proposed a new method to dispose the conflict evidence problem in D-S evidence theory. The method first used evidence distance to obtain the weights of different evidences, then fuzzy inference theory was used to adjust the basic probability assignments (BPA) of every evidence according to the obtained weights, and the adjusted BPAs were fused by the D-S combine rules to obtain the fusion results in the end. The method was used in pulsed gas tungsten arc welding (GTAW) and fused the information obtained by arc, sound, and visual sensors in pulsed GTAW process. Experiment results showed that the method could effectively solve the conflict evidence problem, sufficiently fuse different sensors' information, and obtain more precise results than using a single sensor alone.