An Efficient Method for Bridge Condition Assessment from Dynamic Response of a Passing Vehicle Based on KNN-SVM Approach

When there isn't much money for bridge repair, transportation network management needs to include bridge condition evaluation. This is because resources need to be used carefully. Putting rehabilitation efforts first helps keep unnecessary repairs and strengthening to a minimum, as well as the risks and expenses of waiting too long to fix things that could cause major damage or asset loss. Preprocessing, feature selection, and model training are all parts of the bridge assessment approach. Normalising during preprocessing lowers the number of false positives that come from true system changes, which makes the evaluation more reliable. The mRMR technique uses filters to find the most useful information for classification problems. They employ a hybrid KNN-SVM architecture for training, which takes the best parts of both methods and then replaces them with a version that is relevant to the task. The proposed method works better than standard methods, with a success percentage of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{9 7. 2 3 \%}$</tex>. These results suggest that better preprocessing, strong feature selection, and hybrid learning models can make Bridge Condition Assessment more accurate, help with planning maintenance, and make infrastructure safer.