基于OEDR和ODC的自动驾驶汽车实际道路测试评价技术研究

Research on the Test and Evaluation Technique of Real Roads for Automated Driving Vehicles Based on OEDR and ODC

本文中提出了基于自动驾驶目标和事件探测与响应(OEDR)的测试道路分类方法,构建基于设计运行范围的道路元素基础模型,改进Jaccard Similarity有效相似度算法用于测试道路符合性评估。本文中还根据测试道路选取,提出了基于设计运行条件(ODC)边界的测试方法,结合乘员主观评价形成自动驾驶汽车实际道路测评体系。选取我国典型道路开展的试验结果表明,本文中提出的测评体系具有可行性和普适性,可作为一种标准化方法广泛应用。

In this paper,a classification method of test roads is proposed based on the object and event de-tection and response(OEDR)task of the automated driving system,the road element fundamental model is con-structed based on the operation design domain,and the effective Jaccard similarity method is modified for the con-formity assessment of test road. In addition,a test and assessing method based on the boundary of operation designcondition is also put forward according to the selection of test road,and combined with occupant subjective evalua-tion,a test and evaluation system for automated driving vehicles on real road is built. The results of test on typicalroads in China show that the test and evaluation system proposed has good feasibility and universality,being able tobe widely applied as a standardized method.