A Systematic Review on Autonomous Vehicle: Traffic Sign Detection and Drowsiness Detection

Abstract

AbstractThe automatic detection of traffic signs is necessary for assisted driving, autonomous driving, and driving safety. Traffic sign play a significant task for advanced driver assistance systems (ADAS) also for autonomous driving vehicles and also driver drowsiness detection are an important part. Due to fatigue and drowsiness of the drivers, each day more number of fatalities and deaths are massively increases. In order to avoid these problems, developed a traffic signs detection and drowsiness detection based on machine learning and deep learning techniques. Histogram of Oriented Gradients (HOG), Adaptive Momentum Estimation (ADAM) optimizer features, Random Forest (RF), Region-based Convolutional Neural Network (R-CNN), Long Short Term Memory (LSTM), and Support Vector Machine (SVM) method are used. German Traffic Sign Detection Benchmarks (GTSDB) dataset is used for classification and detection and it consist of 164 classes grouped into 8 categories. The proposed methods achieve the better results in conditions of accuracy comparable performance with the state of the art.KeywordsAutonomous vehicleConvolutional Neural NetworkDrowsiness detectionDeep LearningTraffic sign detection