Processing, assessing, and enhancing the Waymo autonomous vehicle open dataset for driving behavior research

Recently released Autonomous Vehicle (AV) trajectory datasets can potentially catalyze research progress on AV-oriented traffic flow analysis. This paper aims to comprehensively and systematically process assess one of the open datasets, i.e., Waymo Open Dataset, with a focus car following paired trajectories. First, original dataset has been processed into user-friendly format which contains all important information related behavior AV surrounding objects. Second, data quality assessed in terms internal consistency, jerk values completeness. Results show that extracted trajectories are incomplete but generally they have better than Next Generation Simulation program (NGSIM) dataset. Third, further enhanced by using an optimization-based outlier removal method wavelet denoising method. Additionally, we tested impact outliers noise IDM calibration, revealed significant differences parameter for desired time gap T maximum acceleration a.