Towards Complexity Level Classification of Driving Scenarios Using Environmental Information

A driver's driving performance and behavior are highly related to their cognitive workload, which corresponds as well to the complexity of their driving scenarios. Their performance might be impacted when driving through downtown on a rainy night, compared with driving on a remote countryside road on a sunny afternoon. Therefore, advanced intelligent vehicles should have the ability to assess the complexity level of the current driving scenario and therefore understand a driver's cognitive workload to prevent potential hazard situations. In this study, environmental information extracted from the OpenStreetMap (OSM) domain, surrounding vehicle information derived from the video, and prior environmental knowledge such as weather, time and driving location are employed to explore the capability of classifying the complexity level of driving scenarios. A 0.855 F1-score was achieved using the random forest algorithm and combining all features extracted from the environmental information. Conclusions with potential benefits of the development of intelligent vehicles and future study directions are also discussed.