An Approach to Terrain Trafficability Evaluation Based on a Neural Network for Emergency Decision-Support Systems

This paper deals with a new method to evaluate the velocity of a transport unit based on static and dynamic data on terrain relief, urban objects, and the values of hydrometeorological parameters. We considered the existing and available approaches to off-road routes based on the parameters mentioned above (considering velocity reduction depending on terrain parameters) and defined the problem statement. Then, we designed the method to evaluate the velocity on a terrain surface using an artificial neural network. Our method considers the static data (e.g., type of surface) and dynamic hydro-meteorological parameters of the terrain. To determine the efficiency of the method, we conducted the experiment using data from the Leningrad region. The experiment shows significantly increased accuracy in the velocity evaluation of the transport unit.