Road speed prediction method based on machine learning by analyzing road environment data and recording medium thereof

摘要

In the method of predicting the road speed, which is the speed of vehicle flow passing through a specific road at any point in the present invention, the step of collecting road environment data of a road for which the road speed is to be predicted (hereinafter referred to as a 'target road') in real time , generating a prediction data set necessary for predicting road speed by normalizing the collected road environment data, predicting the primary prediction speed through a neural network learning model using the generated prediction data set as input data, the target road To reflect the general vehicle flow, applying the past average speed of the target road to the first predicted speed, to reflect the vehicle flow that changes rapidly due to an event occurring on the target road, to the first predicted speed and estimating the final prediction speed through error correction by applying the event weight and applying the event weight and the past average speed.