A quantitative model for the risk evaluation of driver-ADAS systems under uncertainty

In this paper, a quantitative model is proposed to assess the probability of accidents occurring in driver-Advanced Driver Assistance Systems (ADAS) under uncertainty using Valuation-Based System (VBS). Two kinds of uncertainties are analyzed: data uncertainty related to the states of components, and model uncertainty related to the system structure. The components and the system structure are modeled using variables, spaces of variables, and a set of valuations represented by basic probability assignments (bpas). Besides, the positive influence of learning and cooperation processes is also quantified. Finally, the proposed method is applied to a real use case: the Car Navigation System (CNS).