With the continuous development of autonomous driving technology, the test of self-driving cars has become an important part of it. Among the use case-based, public road-based, and scenario-based test methods, the scenario-based test method is better than the other two methods in terms of freedom and operability. This paper proposes a pedestrian accident scenario construction method based on traffic accident data, selects 390 pedestrian accidents that meet the requirements from the traffic accident data, and selects the appropriate data composition by analyzing the definition, structure, function and test scenario of the AEB system. The parameter set of the scenario, through which the corresponding test scenario can be generated in a large amount, and then the analytic hierarchy process is used to sort the selected parameters, and the degree of influence of each parameter on the scenario construction is obtained, thereby rapidly generating an ideal testing scenarios. Finally, based on a real accident data, a test scenario is created in the virtual test software PreScan, and some parameters in the scenario are replaced to verify the feasibility of the scenario construction method. This article provides a direction for the development of a large number of virtual test scenarios and the selection of ideal scenarios.

CICTP