Driver Profiling Using Long Short Term Memory (LSTM) and Convolutional Neural Network (CNN) Methods

Driver profiling has a major impact on traffic safety, fuel consumption and gas emission. LSTM CNN based neural network models were developed to classify assess bus driver behavior characterized by deceleration, engine speed pedaling, corner turn lane change attempts. Deceleration, test scenarios performed concrete paved track while changing tests conducted commercial asphalt highway. Despite the majority of studies relying image, vehicle data additional sensor fusion, here only streams received from CAN Bus system used train proposed architectures. After parsing into meaningful characteristic parameters, different architectures trained varying number layers, neurons epoch number. Both 1D-CNN networks resulted in comparable success rates. architecture indicates better performance indices for identification aggressive driving compared behavioral modelling.