The invention discloses a short distance driving time prediction method comprising the following steps: S1, using a traffic data acquisition device to collect data, and using a normalization method to carry out pretreatment for the collected data; S2, applying the pretreated data to train a stack autoencoder depth network, thus obtaining an optimal network structure and a corresponding parameter, and generating the optimal stack autoencoder depth network; S3, using the trained optimal stack autoencoder depth network to predict a vehicle short distance driving time. The short distance driving time prediction method can deeply excavate a non-linear relation and other inner characteristics between input variables, is high in prediction precision, and high in robustness.
Short distance driving time prediction method
2015-09-09
Patent
Electronic Resource
English
Method For Driving Ultrasonic Sensors To Measure Short Distance
European Patent Office | 2018
|Method For Driving Ultrasonic Sensors To Measure Short Distance
European Patent Office | 2018