To improve the reliability and stability of the highway preventive system

the key is to improve the accuracy class of dynamic weighting system

propose a dynamic weighing data processing method combining the expert system and fuzzy multi-threshold de-noising algorithm. Because of more knowledge and complex process in dynamic bending sensor weighing process

use expert system to deal with these knowledge; due to strength of interference of signal of the dynamic weighing system in different wavelet decomposition layers and different data range is different

use fuzzy multi-threshold de-noising algorithm to elimination noise of the original dynamic weighing data

makes the reconstructed dynamic weighing datas meet the requirement and in the Matlab analysis shows that the algorithm is effective. Verified through field trials

the combination of the expert system and fuzzy multi-threshold de-noising algorithm

makes the system weighing error less than 2 % and the weighing accuracy achieves the second class index.