面向汽车智能化仿真的交通模拟

发布时间：2018-01-29 22:22

  本文关键词： 交通建模 模糊决策 动态轨迹规划 交通仿真系统　出处：《吉林大学》2015年硕士论文　论文类型：学位论文

【摘要】：与传统汽车技术研发不同，汽车智能化系统与汽车行驶环境、特别是交通环境高度相关。交通和交通车辆对智能汽车或汽车智能行驶过程的环境传感感知、轨迹规划和自动控制等具有十分重要的影响，是关乎汽车智能行驶安全和通行效率的关键因素。当前基于试验场地的实车试验仍然是汽车智能化技术和产品开发、测试和验证的主流平台，但实车试验往往成本高、周期长、可重复性差且往往具有较高的危险性。虽然模拟仿真技术越来越成为了汽车产品研发的重要技术手段，但现有汽车模拟仿真平台多侧重于车辆本身的性能问题，而较少关注对汽车行驶环境和环境传感的模拟，因此难于满足对汽车智能化技术的虚拟研发需求。因此研发面向汽车智能化仿真的交通模拟仿真技术和仿真平台就具有十分重要的意义。 不同于基于多刚体力学的传统汽车模拟仿真技术，也不同于基于流体力学的传统交通模拟仿真技术，面向汽车智能化的交通仿真技术是一个较新的研究领域，其目的在于为车载环境传感与感知、轨迹规划和智能决策、智能或无人驾驶等提供具有较高置信度的、且能与智能汽车形成相互交互的实时交通环境，因此充满着技术挑战。 随着基于计算机图形学的汽车行驶环境三维建模技术的发展，随着车载像机和雷达等环境传感器的模拟仿真技术的发展，对面向智能汽车行驶的交通环境模拟提出了新的要求，包括具有交通流的随机性特征、具有车辆动力学的特征和具有主动控制的特征等，使之能与智能汽车形成相互交互。此外，模型的置信度和高效实时性也是交通车辆和环境模拟仿真的关键要求，不仅可以弥补实车试验的不足、部分替代以减少实车试验，也是提高智能汽车技术与产品研发质量和效率的关键技术。 本文通过对现有研究现状的深入调研，以汽车智能化仿真需求为导向，对面向汽车智能化仿真的交通模拟方法进行了研究。在此基础上开发了具有不同宏观特征的交通流模型和微观交通车辆模型，使交通车辆具有自主驾驶和行为决策等能力，实现了交通车辆之间，交通车与智能汽车之间的交互行为，构建出复杂而又真实的交通试验环境。 本文的主要工作和成果如下： 1．开展了对交通和交通车辆模拟方法的研究，并据此设计了其模型。该模型包括发车模型和智能交通车辆模型。发车模型包括交通流宏观参数，以产生符合需求的交通流。智能交通车辆模型能够根据道路模型以及周围交通车辆信息等自主行驶。 2.开展了对车辆跟驰理论的研究；在扩展型IDM模型基础上，通过改进加速度调整项解决了其在车辆重启动方面的不足。 3.研究了智能交通车辆的行为决策方法，建立了基于模糊推理的换道决策模型。提出以驾驶员不满积累程度为换道决策行为产生的客观评价指标。该模型考虑了换道时影响驾驶员换道愿望的主客观因素，分别以换道收益和驾驶员鲁莽程度作为评价指标，另外以潜在冲突可能性评价换道可行性。采用模糊逻辑的方法将影响换道决策的各个因素联系起来，有效地模拟了真实驾驶员的行为决策过程，使得智能交通车辆的行为决策表现具有较高的置信度，充分反映了实际交通流的动态不确定性，提高了交通模拟的置信度。 4.针对智能交通车辆的换道过程，提出了基于多目标约束的动态轨迹规划方法。该方法通过二自由度车辆动力学模型生成轨迹，不仅考虑了车辆的动力学特性，而且计算效率高，实时性好。该轨迹可以根据安全和效率的不同需求进行最优化调整，，结合交通车轨迹预估，能够在复杂的道路交通环境中快速地规划出一条兼顾安全性、舒适性、操纵稳定性和驾驶效率等多目标约束的轨迹，确保了对动态交通环境的快速响应，使得智能交通车辆在道路行驶中具有较高的真实度。 5.在交通建模理论研究的基础上，设计并编程实现了面向汽车智能化仿真的交通模拟系统。设计了辅助的道路结构和依托于道路结构的车辆管理体系。通过测试证明了理论模型和系统的有效性。
[Abstract]:Different from the traditional automobile technology development, intelligent vehicle system and road environment, especially traffic environment are highly correlated. The traffic and transport vehicles for environmental sensing perception of intelligent vehicle or vehicle intelligent, has important implications for trajectory planning and control, is a key factor related to the intelligent vehicle driving safety and traffic efficiency based on the real vehicle test. The test site is still the intelligent vehicle technology and product development, testing and validation of the mainstream platform, but the real vehicle tests are often high cost, long cycle, risk of poor repeatability and often have high. Although analog simulation technology becomes an important technical means of automotive product development however, the problem that the existing automobile performance simulation platform mainly focus on the vehicle itself, but pay less attention to the vehicle simulation environment and environmental sensing, so difficult In order to meet the demand of virtual research and development for automotive intelligent technology, it is very important to develop vehicle oriented intelligent simulation technology and simulation platform.
Different from the simulation of multi rigid body mechanics is different from the traditional car based on the fluid mechanics of traditional traffic simulation technology based on traffic simulation technology of the intelligent vehicle is a relatively new field of study, its purpose is to provide the vehicle environment sensing and perception, path planning and intelligent decision-making, intelligent or unmanned etc. provide high reliability, real-time traffic environment and the formation of mutual interaction and intelligent vehicle, so full of technical challenges.
With the development of 3D modeling technology of driving environment of computer graphics based on the car, with the development of vehicle like machine simulation technology and radar sensor environment, new requirements oriented intelligent vehicle traffic environment simulation is proposed, including random characteristics of traffic flow, with characteristics of vehicle dynamics and characteristic of active control so, to form mutual interaction and intelligent vehicle. In addition, the key requirement of simulation confidence and efficient real-time is traffic and environment model, can not only make up the inadequacy of the real vehicle test, in order to reduce the partial replacement of real vehicle tests, but also improve the key technology of intelligent vehicle technology and product development quality and efficiency.
Through in-depth research on the status quo of the existing research, the intelligent vehicle simulation oriented, oriented intelligent simulation of automobile traffic simulation method is studied. On the basis of the development of the model and the micro vehicle traffic flow model with different macroscopic characteristics, so that the vehicle has autonomous driving and behavior decision. To achieve the traffic between vehicles, interactions between traffic and intelligent vehicle traffic, build the experimental environment is complicated and real.
The main work and results of this article are as follows:
1. research on simulation method for traffic and traffic vehicles, and based on the model. The model includes the grid model and intelligent vehicle model. The grid model including the macroscopic traffic flow parameters, to meet the needs of the traffic flow. The intelligent transportation vehicle model according to the road and the surrounding traffic information model of autonomous driving.
2., vehicle following theory has been studied. Based on the extended IDM model, it has solved its shortcomings in vehicle restart by improving the acceleration adjustment.
3.鐮旂┒浜嗘櫤鑳戒氦閫氳溅杈嗙殑琛屼负鍐崇瓥鏂规硶,寤虹珛浜嗗熀浜庢ā绯婃帹鐞嗙殑鎹㈤亾鍐崇瓥妯″瀷.鎻愬嚭浠ラ┚椹跺憳涓嶆弧绉�绱�绋嬪害涓烘崲閬撳喅绛栬�屼负浜х敓鐨勫�㈣�傝瘎浠锋寚鏍�.璇ユā鍨嬭

本文编号：1474501