Research on road bridge structural health monitoring and vehicle load prediction model integrating multi-source data

Objective: To solve the problems of road bridge structural health monitoring and vehicle load prediction, a spatiotemporal attention fusion neural network algorithm is proposed to improve the accuracy of monitoring and prediction. Methods: The algorithm uses LSTM and CNN to extract temporal and spatial features respectively, strengthens key information through the spatiotemporal attention mechanism, realizes the deep fusion of multi-source data, and is experimentally verified on the actual road bridge dataset. Support vector machine and ordinary neural network are used as comparison methods, and the accuracy and mean square error are used for evaluation. Results: In structural health monitoring, the accuracy of the proposed algorithm reaches 92.8%, which is more than 11% higher than the traditional method; the mean square error of vehicle load prediction is 13.1, which is 22.3% lower than the comparison algorithm. Ablation experiments show that the spatiotemporal attention mechanism and data fusion strategy significantly improve the model performance, providing an efficient solution for road bridge safety monitoring and load prediction.