Research On Data Fusion Method Of Large Diameter Multi-sensor Measurement System

Study and application of multisensor data fusion theory have aimed to combine multiple source information from various sensors intelligently and obtain more detailed, complete, dependable description and decision than a single sensor. Application spans a broad range that includes military and civilian. Multisensor data fusion is one popular research field in domestic and foreign scientific and technological circles.To improve the measurement accuracy while measuring large diameters using rollers, the errors which affect measurement precision should be corrected such as that caused by the distortion of the rollers and temperature fluctuate etc. So five kinds of transducers, more than 30 in number totally are used in the developed measuring system. Through data fusion, we can measure the big diameter in high precision.In the paper, multi-sensor data fusion technology is presented, and on the base, measuring large diameters using rollers data fusion system is designed and achieved. Especially, the data level fusion and the character level fusion based on neural network are accomplished.In data level fusion, abnormal measurement data of three rollers' rotating angles is found and mended firstly. Then the locality of skidding can be identified by the correlation analysis of large-diameter measurement based on multiple rolling-wheel method. Finally, the simulation proves the validity of the given method.In feature level fusion, through three fusion methods of back propagation neural network, adaptive weight and mean, we fuse three groups measurement datum and get a group result. Finally the conclusion is proved by simulation that the back propagation neural network method is optimal.