预期功能安全场景库复杂度量化方法研究

基于Pegasus场景分层体系,提出了一种关于场景复杂度的量化方法,以此来评定场景数据的质量。该方法确定了每层要素的决定因素,根据决定因素确定每层要素的复杂度,通过求出各层要素复杂度之和,得到场景数据的总复杂度。此外,为了防止过复杂现象,提出了母子库法和系统场景概率法,将场景要素的复杂度乘以该要素的出现概率,得到修正后的复杂度。研究结果显示,通过该方法可以构建合理可用的场景库。

A Complexity Quantification Method for Safety of the Intended Functionality Scenario Library

Based on the Pegasus hierarchical system, a method of scenario complexity calculation wasproposed to evaluate the quality of scenario data. Initially the determinants in each layer were selected and correspondingly the complexity for each layer was determined. The total complexity of the scenario data was obtained by calculating the sum of the complexity in each layer. In addition, in order to prevent over complexity, the parent-child database and the system scenario probability were proposed. The complexity of a scenario element was multiplied by the probability of its occurrence to acquire the modified complexity. Thefinal results show that the proposed method can build a reasonable and useful scene library.