Temperature and humidity prediction of mountain highway tunnel entrance road surface based on improved Bi-LSTM neural network

As the temperature of mountain roads is low and the temperature difference between day and night is large, the change of road surface temperature and humidity will affect the road conditions. For example, cold and wet roads are prone to ice condensation, which will reduce the pavement performance and affect driving safety, greatly increasing the accident risk at the tunnel entrance and exit. Therefore, in this paper, the temperature and humidity of the road surface at different distances and times are measured on the spot to explore their spatio-temporal variation patterns, and the improved distribution estimation particle swarm optimization (DEPSO-Bi-LSTM) model is used to predict them. Compared with LSTM and Bi-LSTM, the results show that the R-2 predicted by this model is 0.92 and 0.86. Because of its advantages in feature extraction and multi-dimensional data processing, its performance is better than other algorithms. The presented method can provide technical references for temperature prediction of the pavement and the development of an early-warning system for icy pavements in cold regions.