Multi-Agent Temperature State Recognition Based on D-S Evidence Theory

Body temperature state pattern is the most important index, which can reflect normal and abnormal running state of the high power loading devices for the vehicle on subway, in order to timely, quickly and accurately recognize body temperature state of the objective device, and reduce false alarming probability, make use of neural network system intelligent Agent and expert system intelligent Agent which simultaneously recognize body temperature state of the objective devices and get probability pattern recognition, and applies D-S evidence theory and weight factor into steps which fuse several intelligent agents into one union intelligent agent, finally make a intelligent decision result. The practical results show that this method greatly strengthens reliability of recognition ability for device temperature state pattern, and also is easy to expandable.