Human-like Decision Making for Autonomous Lane Changing Using Deep Learning

Machine learning technique has been shown to greatly improve the intelligence level of the decision-making system of intelligent vehicle. However, due to the complexity and uncertainty of the real-world traffic environment, its reliability is difficult to guarantee. To address this issue, a method for human-like autonomous lane changing decision-making that adopts hierarchical decision-making is proposed. We built a sample database by carrying out simulation driving experiments, based on the data-driven and deep learning LSTM method to generate lane changing intention, and further verify the rationality of the intention through constraint rules. The simulation experimental results show that this method can safely and effectively realize the autonomous lane changing of intelligent vehicle, which can provide a reference for the research of intelligent vehicle decision-making technology.