A Method for Extracting Cut-in Scenario Based on Real-World Road Data Collected by ASEva

Based on the real-world driving data collected by the ASEva scene data acquisition system of ZEER Company, this paper proposes a method to extract dangerous cut-in scenarios from the driving data. This method analyzes the original data of ASEva to obtain effective sample data corresponding to a certain moment, such as the ego vehicle data, lane line data, and the target object data. Based on these data and algorithms, the front key vehicle is extracted, while excluding the misjudgments caused by excessive curvature due to the big curves on the road or the curvature caused by the turning of the ego vehicle. Finally, it is determined whether the latitudal absolute velocity of the target vehicle meets the threshold of the dangerous cut-in scenario. If the parameters of the target vehcile meet the cut-in scenario, the segment data between the start and end time is extracted. Finally, the accuracy of the extraction of dangerous cut-in scenarios is verified through the scenario generation method.