As a most common driving behavior in traffic flow, lane-changing behavior is one of the most important factors that causes traffic accidents and traffic congestion. Existing studies mainly used single type of data collected from Internet of Things sensors, and didn’t take every aspect into consideration to fully evaluate driver behavior. In this work, we proposed a machine learning-based approach for lane-changing behavior evaluation using data collected from multiple sensors. The lane change behavior data were collected by the motion sensors embedded in the smartphone and multiple cameras installed inside the vehicle. A list of features was extracted from the collected data and a decision tree was constructed to evaluate lane-changing behaviors. The collected lane-changing behavior data was split into training and testing data set to build and validate the decision tree. The validation results show that the proposed model can accurately evaluate different lane-changing behaviors. Besides, this study provides theoretical support for monitoring and early warning of lane-changing behavior for connected vehicles.

CICTP