基于萤火虫-免疫算法的CVIS测试序列优化方法

IFA-based Test Sequence Optimization Method for CVIS

为了对车路协同系统的功能进行有效测试,提高测试效果,对测试序列优化方法进行了研究。通过对车路协同系统进行功能需求分析,构建包含多交叉路口、多路段、多车的车路协同系统典型应用仿真场景,并对典型场景利用故障树分析的方法进行功能特征分析,建立基于系统功能特征的系统测试用例,研究车路协同的功能特征,生成的测试用例进行带端口标记的时间输入/输出自动机建模,并用建模工具UPPAAL完成建模实现,利用UPPAAL建模工具的测试序列生成模块,完成基于带端口标记的时间输入/输出自动机的测试序列的自动生成,得到针对车路协同系统仿真测试的测试序列。着重设计运用萤火虫算法和免疫算法对已经生成的测试序列进行优化,并且从序列覆盖度、冗余度、快速性方面对测试效果做出了评估。研究结果表明:当示例场景状态数较多时,优化算法能保证测试序列的覆盖度为100%,且能降低测试冗余82%,缩短测试时间约为10s;基于萤火虫-免疫算法的测试序列优化方法在测试冗余的降低上卓有成效,且优化算法较时间自动机而言能够更快更有效地降低测试冗余,优化算法的测试效果明显优于时间自动机生成的原始序列;通过对测试冗余的降低实现测试效率的提高具有可行性。

In order to test the function of cooperative vehicle infrastructure system（CVIS）effectively and improve its effect,the test sequence optimization method was investigated.According to the functional requirement analysis of CVIS,typical application scenarios containing multiple intersections,multiple sections and multiple vehicles were constructed.Then,functional characteristics were illustrated by the analysis of typical application scenarios with fault tree analysis（FTA）. Moreover,test cases based on functional characteristics of CVIS were established.The test sequences of simulation test in light of CVIS were generated by a modelinganalysis by dint of the method of labeled-ports time input/output automata（Lp-TIOA）and the modeling tool of UPPAAL.Besides,test sequences generated by the Lp-TIOA modeling were optimized by firefly algorithm and immunity algorithm（IFA）proposed in this paper.Meanwhile,the effect of test was assessed from the degree of coverage,redundancy and quickness.The results show that the coverage can be ensured to be 100% by the optimizing algorithm,the redundancy decreases to 82%,and test time is shorten to 10 swhen the number of scenario states is large.The test sequence optimization is fruitful in reducing the test redundancy based on the IFA and the IFA is faster and more effective to reduce the test redundancy than TIOA.It follows that the optimization result is better than the original sequence generated by timed automata.The test efficiency can be increased with the decrease of the redundancy.Therefore,the test sequence optimization algorithm is feasible.