基于驾驶员博弈仿真的车联网下节能策略评价

发布时间：2018-07-25 10:59

【摘要】：车联网技术的不断发展,为智能交通系统的实现提供了可能,在车联网逐渐普及的过程中,如何利用车联网提供的信息来帮助驾驶员提高行车效率、减少燃油消耗成为汽车领域的研究热点。从车辆方面出发进行节能驾驶的研究已经相对成熟了,这类研究主要是根据发动机的动力性和经济性去调整车辆在行驶中的速度和加速度,该方法在实际使用中常常因受到周围车辆的干扰等原因,很难达到预想的行驶效率及节能效果。从驾驶员行为及交通条件出发进行节能驾驶的研究,可以避免因为没有考虑周围其他车辆造成与实际状况不符的情况,但提出的节能驾驶策略是否有效需要进行实验验证。然而,基于实车的实际道路或者试验场实验通常存在实验成本高、周期长、场景复杂并且重复性差等问题,而传统的车辆仿真技术和交通系统仿真技术虽然具有成本低重复性高的优势,但也存在无法实现车联网下车辆对周围环境感知及车辆间信息分享等实验环境的不足。因此,本文对车联网下节能策略评价方法进行了研究,建立了面向车联网下车辆运行仿真的驾驶员博弈仿真平台,并用于评价驾驶员的节能驾驶策略。本文的主要工作和成果总结如下:1.对驾驶员博弈仿真系统进行分析,设计了驾驶员博弈仿真平台。平台主要包括道路生成与信号灯模块、驾驶员博弈行为仿真模块、人机交互模块和数据分析模块。道路模型和信号灯配时方案参考实际道路情况选择,驾驶员博弈行为仿真模块包含了GM跟驰模型及改进的Gipps换道模型,其中驾驶员行为参数根据实际道路的调查结果选取,人机交互模块可以动态显示仿真过程中各个车辆位置和信号灯实时状态,数据分析模块可以根据仿真结果统计出每个车辆对应的各项行驶指标并保存输出。2.对驾驶员博弈仿真平台的性能进行了评价。对驾驶员博弈仿真平台评价指标进行了研究,确定以平均行程速度和驾驶员分布均匀性为主要评价指标,并参考速度分布、怠速比和速度时间序列几项指标。设计仿真实验,根据确定的几项指标分别对平台进行验证,得到仿真平台性能满足设计要求的结果。3.对典型交通状况下的仿真参数选择进行了研究。设计在不同车辆数下的驾驶员博弈仿真实验,确定表征自由流、平峰流、高峰流交通状态的车辆数分别为100辆,300辆和800辆。设计在不同仿真时间下的驾驶员博弈仿真实验,确定仿真时长达到5400s时,仿真参数和统计数据总体收敛的结果。4.对评价驾驶节能策略效果的指标和参考值进行了研究。对用于评价驾驶节能策略的指标进行了分析,确定以时间收益和油耗收益为主要指标,并使用价值方法将这两个指标统一成行驶总收益这一个指标作为评价驾驶节能策略的指标。5.对固定型驾驶策略的行驶收益进行了博弈分析,揭示了各种驾驶策略在多策略环境下的不同收益情况。设计在不同的车辆数下,对五种固定型驾驶策略在不同比例分布及单类型驾驶员时的仿真实验,分析在各种条件下各类型驾驶员的收益结果,并将每种交通密度及驾驶员分布下所有驾驶员的平均收益作为评价驾驶节能策略的参考值。根据纯策略博弈和混合策略博弈分析理论上各类型驾驶员的行驶收益,对比理论值和仿真的真实值,推断出:在实际交通环境中,采用各类型驾驶策略的驾驶员收益差比理论值小,整体收益呈现一种更均衡的状态。6.对车联网下跟随策略、平均策略和组合策略的三种驾驶节能策略实际效果进行了分析,揭示了复杂交通环境下实现节能策略的难度,也说明了在多策略驾驶条件下进行策略有效性评价的必要性。设计在不同驾驶员类型比例分布及不同车辆数下的仿真实验,对分别使用跟随策略、平均策略和组合策略的实际收益进行统计,并根据统计结果评价各类型节能策略的实际效果和适用范围。根据分析结果得出,各类型驾驶员都无法在任何交通环境和交通密度下持续占优,只有根据道路的交通环境和交通密度实时调整驾驶策略才能达到节能效果的结果。
[Abstract]:The continuous development of vehicle networking technology provides the possibility for the realization of the intelligent transportation system. In the process of the gradual popularization of the vehicle network, how to use the information provided by the vehicle network to help drivers improve the driving efficiency and reduce fuel consumption has become a hot spot in the field of automobile research. It is mature. This kind of research mainly adjusts the speed and acceleration of the vehicle in driving according to the power and economy of the engine. This method is often difficult to achieve the expected driving efficiency and energy saving effect because of the interference of the surrounding vehicles in actual use. It can be avoided because the other vehicles around the world do not agree with the actual situation, but the effectiveness of the proposed energy-saving driving strategy needs to be verified experimentally. However, the actual road or test field experiments based on real cars usually have problems such as high experimental cost, long period, complex scene and poor repeatability, etc. The traditional vehicle simulation technology and the traffic system simulation technology have the advantages of low cost and low repetition, but there are also the inability to realize the experimental environment of vehicle environment perception and information sharing between vehicles. Therefore, this paper studies the evaluation method of energy saving strategy under the vehicle network, and establishes a car facing couplet. The driver game simulation platform of the vehicle operation simulation under the network is used to evaluate the driver's energy-saving driving strategy. The main work and results of this paper are summarized as follows: 1. the driver game simulation system is analyzed and the driver game simulation platform is designed. The platform mainly includes the Lu Shengcheng and the signal lamp module, and the driver's game behavior is imitated. True module, man-machine interaction module and data analysis module. The road model and signal timing scheme refer to the actual road condition selection. The driver game behavior simulation module includes the GM following model and the improved Gipps lane change model. The driver behavior parameters are selected according to the actual road investigation results, and the human-computer interaction module can be moved. The position of each vehicle and the real-time state of the signal lamp are displayed in the simulation process. The data analysis module can count the driving indexes corresponding to each vehicle according to the simulation results and evaluate the performance of the driver game simulation platform. The simulation platform evaluation index of the driver game is studied and the.2. is determined. The average travel speed and the uniformity of the driver distribution are the main evaluation indexes, and refer to the speed distribution, the idle speed ratio and the speed time series. The simulation experiments are designed to verify the platform according to the certain indexes. The simulation platform performance meets the design requirements and the simulation parameters of the typical traffic conditions are obtained by.3.. The driver game simulation experiment under different vehicle number is designed to determine the number of vehicles representing free flow, flat peak flow and peak flow traffic state, respectively, 100 vehicles, 300 vehicles and 800 vehicles. The driver game simulation experiment under different simulation time is designed to determine the simulation parameters and statistical data when the simulation is up to 5400s. The result of overall convergence.4. studies the index and reference value of the effect of driving energy saving strategy evaluation. The indexes used to evaluate the driving energy saving strategy are analyzed, and the main indexes of the time income and oil consumption income are determined, and the value method is used to unify the two indexes as the driving total income as the evaluation driving. The index.5. of driving energy strategy has a game analysis on the driving income of the fixed driving strategy, and reveals the different benefits of various driving strategies under the multi strategy environment. The simulation experiments of five fixed driving strategies under different proportion distribution and single type drivers are designed under different number of vehicles. The income of each type of driver is given, and the average income of each driver is considered as the reference value of the driving energy saving strategy. The driving income of all types of drivers in the theory of the pure strategy game and the mixed strategy game is analyzed, and the theoretical value and the real value of the simulation are compared. In the actual traffic environment, the driver income difference of each type of driving strategy is smaller than the theoretical value, and the overall income presents a more balanced state.6.. The actual effects of three driving energy saving strategies are analyzed, which reveals the difficulty of realizing the energy saving strategy in complex traffic environment. It also illustrates the necessity of evaluating the effectiveness of the strategy under the conditions of Multi Strategy driving. The simulation experiments of different driver type ratio distribution and different vehicle number are designed, and the actual income of the following strategy, the average strategy and the combination strategy are used respectively, and the actual results are used to evaluate the practice of each type of energy saving strategy. According to the results of the analysis, it is concluded that all types of drivers can not continue to dominate in any traffic environment and traffic density. Only according to the traffic environment and traffic density can adjust the driving strategy in real time to achieve the result of energy saving.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U495

【相似文献】