Driver behavior has become an essential consideration throughout the current smart transportation system. Most drivers are unfamiliar with this new technology and very ignorant of their driving style, which leads to problems like a violation of pedestrian safety. Driver behavior analysis (DBA) is intended to improve passengers’ safety from harsh driving by analyzing the behavior of selected drivers while driving. Therefore, this paper aims to develop a model to detect aggressive driving to minimize traffic safety violations. This paper presents a method using SimpleRNN, LSTM, and GRU individually to detect drivers’ actions from statistical numeric data where LSTM performs better. A public dataset that uses numeric sequential data to predict driver's behavioral conditions. This paper also compares the proposed DBA algorithm with previously used deep learning and machine learning models to detect driver actions by accuracy and precision. After many epochs, the paper concludes that the LSTM model yields greater success than most other models and achieves 0.961 accuracies.
Driver Behavior Analysis Based on Numerical Data Using Deep Neural Networks
Lect. Notes in Networks, Syst.
Proceedings of International Conference on Data Science and Applications ; Chapter : 16 ; 211-219
2021-11-23
9 pages
Article/Chapter (Book)
Electronic Resource
English
Driver behavior analysis (DBA) , Smart transportation system (STS) , Long short-term memory (LSTM) , Recurrent neural network (RNN) , Gated recurrent unit (GRU) Engineering , Computational Intelligence , Artificial Intelligence , Data Structures and Information Theory , Statistics, general , Systems and Data Security
Analysis of Recurrent Neural Networks for Probabilistic Modeling of Driver Behavior
Online Contents | 2017
|