Automatic recognition of driving scenarios for ADAS design

In this paper, a method to characterize and automatically recognize the most common driving scenarios in on-road experiments is presented. The aim of the proposed approach is to build a suitable simulator to develop and test Advanced Driver Assistance Systems (ADAS's). Therefore, unlike most of the existing algorithms, the whole procedure takes advantage of the intrinsic off-line nature of the problem. Context-free grammars are shown to be an effective and suitable tool for modeling the driving scenarios, while experimental results are used to validate the proposed approach and show limits and potential of a real-world application.