Forecast: a neural system for diagnosis and control of highway surfaces

A study was conducted to characterize a neural network that can forecast the temperature of road surfaces. The neural network features backpropagation algorithms and online parameter tuning features. It is also capable of achieving a multiple-step prediction while shifting back other input data by the unit. Results confirmed the reliability of the neural network in forecasting road surface temperatures.