The highD dataset is a large-scale naturalistic vehicle trajectory dataset collected from German highways using drone-based measurements. It provides a comprehensive dataset of 16.5 hours of recordings, including 110,000 vehicles, a total driven distance of 45,000 km, and 5600 recorded complete lane changes. The dataset is designed for scenario-based validation of highly automated driving systems and includes detailed information on vehicle trajectories, maneuvers, and traffic scenarios. The dataset is available online at http://www.highD-dataset.com and includes video recordings, annotated data, and trajectory information. The highD dataset is compared with other datasets such as NGSIM, which is also used for traffic simulation and automated driving research. The highD dataset offers advantages in terms of naturalistic driving behavior, static and dynamic scenario description, and data privacy protection. However, it has limitations in terms of flexibility and effort effectiveness compared to onboard measurement methods. The dataset includes post-processed trajectories, annotations, and maneuver classifications, making it a valuable resource for research on automated driving systems. The highD dataset is intended for use in safety validation, impact assessment, and research on traffic simulation models, traffic analysis, driver models, and road user prediction models.The highD dataset is a large-scale naturalistic vehicle trajectory dataset collected from German highways using drone-based measurements. It provides a comprehensive dataset of 16.5 hours of recordings, including 110,000 vehicles, a total driven distance of 45,000 km, and 5600 recorded complete lane changes. The dataset is designed for scenario-based validation of highly automated driving systems and includes detailed information on vehicle trajectories, maneuvers, and traffic scenarios. The dataset is available online at http://www.highD-dataset.com and includes video recordings, annotated data, and trajectory information. The highD dataset is compared with other datasets such as NGSIM, which is also used for traffic simulation and automated driving research. The highD dataset offers advantages in terms of naturalistic driving behavior, static and dynamic scenario description, and data privacy protection. However, it has limitations in terms of flexibility and effort effectiveness compared to onboard measurement methods. The dataset includes post-processed trajectories, annotations, and maneuver classifications, making it a valuable resource for research on automated driving systems. The highD dataset is intended for use in safety validation, impact assessment, and research on traffic simulation models, traffic analysis, driver models, and road user prediction models.