Research on Driving Behaviors Based on Machine Learning Algorithms

Research on driving behaviors with machine learning methods is one of the major fields in intelligence vehicles, which includes varieties of machine learning analytic and simulation methods. Here our team propose three different machine learning approaches to tackle this problem with real drivers collected data set, with on road performances data we compared those methods with their accuracy rate, results showed that the three models we are purposing, DWT (Discrete Wavelet Transform) - SVM (Support Vector Machine), DTW (Dynamic Time Warping) - kNN (k Nearest Neighbors) and LSTM (Long short-term Memory) all have their merits, which provided certain insights and theoretical supports for further studies.