Convolutional Neural Network for Driving Maneuver Identification Based on Inertial Measurement Unit (IMU) and Global Positioning System (GPS)

Abstract

Identification and translation of different driving maneuver are some of the key elements to analysis driving risky behavior. However, the major obstacles to maneuver identification are the wide variety of styles of driving maneuver which are performed during driving. The objective in this contribution through the paper is to automatic identification of driver maneuver e.g., driving in roundabouts, left and right turns, breaks, etc. based on Inertia Measurement Unit (IMU) and Global Positioning System (GPS). Here, several machine learning (ML) algorithms i.e., Artificial Neural Network (ANN), Convolutional Neural Network (CNN), K-nearest neighbor (k-NN), Hidden Markov Model (HMM), Random Forest (RF), and Support Vector Machine (SVM) have been applied for automatic feature extraction and classification on the IMU and GPS data sets collected through a Naturalistic Driving Studies (NDS) under an H2020 project called SimuSafe 1. The CNN is further compared with HMM, RF, ANN, k-NN and SVM to observe the ability to identify a car maneuver through roundabouts. According to the results, CNN outperforms (i.e., average F 1-score of 0.88 both roundabout and not roundabout) among the other ML classifiers and RF presents better correlation than CNN, i.e., MCC = −0.022.

On-Board Unit (OBU)-Supported Longitudinal Driving Behavior Monitoring Using Machine Learning Approaches

“…For instance, some scholars used acceleration and angular velocity data collected through mobile phones and classified behaviors such as acceleration, deceleration, and turning [ 6 , 7 , 16 , 17 , 18 ]. Some scholars used in-vehicle devices embedded with IMU sensors to acquire vehicle motion data, applying machine learning methods for driving behavior monitoring [ 19 , 20 , 21 , 22 , 23 , 24 ]. Bonfati et al fused the CAN bus information of a vehicle and IMU sensor information to detect the driving behavior of the vehicle [ 25 ].…”

“…Some scholars applied a dynamic time warping (DTW) algorithm to detect and identify driving behaviors [ 16 , 17 , 39 ]. Other scholars applied convolutional neural networks (CNN) with multiple sources of fused data for identifying car maneuvers, and comparisons with other models such as a HMM, RF, artificial neural network (ANN), KNN, and SVM were made [ 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 ]. Junior et al used four algorithms, including multilayer perceptron (MLP), SVM, RF and Bayesian network, to identify driving behaviors such as fast lane changes, turns, braking, and acceleration, and the results showed that RF and MLP demonstrated better performances [ 18 ].…”

Analysis of driving style using self-organizing maps to analyze driver behavior

“…The multilevel models of driving style assessment described in [12] take into account behavioral aspects (driver's emotions and driver's condition) to ensure greater safety and comfort. Machine learning methods for driving style classification are reflected in the works [13]- [20]. The work [13] shows the expediency of using the support vector machine to identify the driving style.…”

“…CNN, long short-term memory (LSTM) network, and pre train-LSTM [15], [17], [20], random forest [16] - [14], [17], [20]/+ [15], [16] -+ + Application of self-organizing maps [21] + + Classification method, which is enabled by ensemble learning [23] --+ +…”

Machine Learning Applied to Identify Corrosive Environmental Conditions

“…Machine learning approaches have been demonstrated to have an excellent predictive power in recognizing patterns in data (Paolanti and Frontoni, 2020). Among a wide range of algorithms, convoluted neural networks (CNNs), support vector machine (SVM), and random forest (RF) are popular approaches used to identify underlying associations in images (AgajanianOluyemiVerkhivker, 2019;Ahmed, 2020). These techniques are increasingly being introduced in material science to identify concealed property relationships (Bulgarevich et al, 2018;Himanen et al, 2019;Xia et al, 2020).…”