基于博弈论的人类驾驶与无人驾驶协作换道模型

发布时间：2018-07-02 19:55

本文选题：换道模型 + 博弈论　；参考：《计算机工程》2017年12期

【摘要】：无人驾驶汽车根据自身获取信息进行决策,与人类驾驶缺乏有效的交互机制,在复杂交通环境下难以发挥优势,甚至会引起事故。为此,引入智能网联汽车的设计思想,针对车辆在行驶过程中的换道行为,基于博弈论构建一种包含车辆协同机制的换道模型。考虑驾驶员实施换道的行为特征,车辆在做出换道决策后观察分析其他车辆的驾驶风格,其结果会影响最终决策。根据安全前提下可能获得的加速空间决定支付函数,同时加入表征驾驶风格的参数,并量化得到驾驶风格分数。在SUMO仿真软件上的测试结果显示,使用博弈换道模型的车辆在不同车流密度环境下平均通行数量提高5.60%,平均通行时间减少8.44%,测试中未发生事故,表明模型具有感知、判断人类驾驶风格的能力,可在保证安全的前提下提高通行效率。
[Abstract]:Driverless vehicles make decisions according to their own information, and lack of effective interaction mechanism with human driving, so it is difficult to give full play to their advantages in complex traffic environment, and even cause accidents. This paper introduces the design idea of Intelligent Networked vehicle, aiming at the changing behavior of the vehicle in the course of driving, based on the game theory, constructs a kind of change model which includes the vehicle coordination mechanism. Considering the behavior characteristics of the driver to change the lane, the vehicle will observe and analyze the driving style of other vehicles after making the change decision, and the result will affect the final decision. The payoff function is determined according to the acceleration space that may be obtained under the condition of safety, and the driving style score is obtained by adding the parameters to characterize the driving style. The test results on sumo simulation software show that the average traffic number of vehicles using game changing model increases 5.60 in different traffic density environment, and the average traffic time decreases 8.44%, and no accidents occur in the test, which indicates that the model is perceptive. The ability to judge human driving style can improve traffic efficiency on the premise of safety.
【作者单位】：大连理工大学电子信息与电气工程学部;
【基金】：国家自然科学基金(61402076) 国家高技术研究发展计划项目(2012AA111902)
【分类号】：O225;U491

【相似文献】