无人驾驶公交车超车行为的评价与优化研究  

Research on evaluation and optimization of overtaking behavior of autonomous buses

作　　者：刘昊 陈韬 谢辉[1] LIU Hao;CHEN Tao;XIE Hui(School of Mechanical Engineering,Tianjin University,Tianjin 300072,China)

机构地区：[1]天津大学机械工程学院,天津300072

出　　处：《武汉大学学报（工学版）》2024年第4期486-496,共11页Engineering Journal of Wuhan University

摘　　要：为了提高无人驾驶公交车超车过程的安全性、舒适性及规划的合理性,针对无人驾驶公交车的超车行为提出基于逼近理想解排序法(technique for order preference by similarity to an ideal solution,TOPSIS)的主客观赋权的评价与优化模型。首先,开发了基于N-K模型的环境脆弱性模型。然后,基于规划的合理性、安全性及驾驶操控性这3个方面构建了无人驾驶公交车超车行为评价体系。最后,分析控制参数对超车行为评分的影响规律并进行优化。基于优化前后的超车行为参数,在5种测试案例下进行仿真测试。结果表明:模型参数被优化后,其对超车行为的评分会随着环境脆弱性的耦合风险的提升而增大,优化后的评分标准差达到0.6138,而优化前的评分标准差仅为0.1061,超车行为的评价区分度有所提高;优化后的超车行为评分相较于优化前平均提升了3.075倍。In order to improve the safety,comfort and rationality of planning in the overtaking process of unmanned buses,a TOPSIS-based evaluation and optimization that combines the subjective and objective weightings is proposed for the overtaking of unmanned buses.First,an environmental vulnerability model based on N-K model is developed.Secondly,the evaluation system of a driverless buses overtaking is constructed based on three aspects in the criterion layer:planning rationality,safety,and driving maneuverability.Finally,the influence rule of control parameters on the scoring of overtaking behaviours is analyzed and optimized.Based on the overtaking behaviours parameters before and after optimization,simulation tests are performed in five test cases.The results show that the score of overtaking behaviours increases with the increase of coupling risk of environmental vulnerability after parameter optimization.The standard deviation of score is 0.6138 after optimization whereas that before optimization is only 0.1061.The score of overtaking behaviours after optimization is 3.075 times higher than that before optimization.

关 键 词：无人驾驶公交车 环境脆弱性 N-K模型 TOPSIS法 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]