论文总字数：30040字

摘 要

Abstract II

第一章 绪论 1

1.1课题研究的背景与意义 1

1.1.1研究背景 1

1.1.2研究意义 2

1.2国内外研究现状 2

1.2.1无人驾驶技术 2

1.2.2三维场景建模 3

1.3研究的主要内容与方法 3

1.3.1研究内容 3

1.3.2研究方法 4

第二章 构建无人驾驶场景的理论与关键技术 5

2.1 三维场景建模方法 5

2.1.1基于三维激光扫描技术的建模方法 5

2.1.2基于图像的三维场景建模方法 5

2.1.3基于CAD的三维场景建模 6

2.1.4基于程序的三维场景建模 6

2.2 虚拟现实技术 7

2.2.1 虚拟现实技术及其特征 7

2.2.2 虚拟现实技术在汽车行业中的应用 7

2.3 图形空间变换 8

2.4 纹理映射 11

2.4.1 仿射纹理映射与视点纹理映射 11

2.4.2 纹理映射常用技术 12

2.5 本章小结 13

第三章 3Dmax中构建城市环境无人驾驶场景 14

3.1道路及交通设施建模 14

3.1.1 三维地形建模 15

3.1.2 道路建模 16

3.1.3 交通设施建模 20

3.2 3DMAX建筑建模 22

3.3 3DMAX环境与相机设置 26

3.3.1 天空与环境背景 26

3.3.2 相机 27

3.4 3DMAX光照与渲染 27

3.4.1 3DMAX光照处理 27

3.4.2 渲染 29

3.5 小结 31

第四章 总结与展望 32

4.1分析场景可行性与代表性 32

4.2研究总结 32

4.3研究展望 33

致谢 34

参考文献 35

摘 要

随着以5G等信息技术的飞速发展，城市智能信息化水平日渐提高，推动着汽车行业转型升级，无人驾驶技术便成为行业的研究热点。为保证无人驾驶汽车能够在城市道路环境下正常行驶，需对其进行大量的道路测试。但城市交通复杂多变，现场试验耗时且需要花费较大的人力和物力，而虚拟测试可以部分取代现场测试，为实测提供方法和理论指导，且不受环境限制，具有可重复性，沉浸性好等优势。因此利用虚拟场景对无人驾驶汽车进行测试具有很大可行性。

目前学者对于无人驾驶的车辆定位，环境感知和自动控制等技术研究较多，比如利用激光雷达或单目相机图像识别进行环境感知、利用加权算法选择最优路径；但是对于无人驾驶汽车运行的外部环境，尤其是市区道路环境的关注较少。而在城区环境下进行路试，是无人驾驶汽车驶入城市的必由之路，故本文着力构建无人驾驶汽车城市路试场景，为无人车仿真测试和高精度地图提供参考。

在了解无人驾驶关键技术、国内外研究现状，分析无人驾驶汽车城区路试场景应有要素，并掌握3DMAX建模功能与步骤的基础上。本文选取相应市区道路与临街环境，参考全景地图、实地数据和道路交通标准，设计无人驾驶汽车路试区域三维模型布局和比例。利用3DMAX建模车道、斑马线、安全岛、红绿灯、植被、建筑物等场景要素，模拟真实城市道路场景，并针对一处典型场景进行全景合成，为模拟无人驾驶场景提供一种可参考的搭建方式。最后论述该场景的代表性和作为无人驾驶路试场景的可行性，并指出场景存在的不足、展望深入研究的方向，结合现有道路交通标准与城市交通环境，提出无人驾驶汽车与城市交通相互适应的发展建议。

关键词：虚拟仿真；无人驾驶汽车；三维建模；城市场景；3DMAX

Abstract

With the rapid development of 5g and other information technology, the level of urban intelligent information is improving day by day, which promotes the transformation and upgrading of the automobile industry. Driverless technology has become a research hotspot in the industry. In order to ensure that the driverless vehicle can run normally in the urban road environment, a large number of road tests are needed. However, the urban traffic is complex and changeable, the field test is time-consuming and needs a lot of manpower and material resources, and the virtual test can partially replace the field test, providing methods and theoretical guidance for the actual measurement, and is not limited by the environment, with the advantages of repeatability and good immersion. So it is very feasible to test driverless vehicle with virtual scene.

At present, scholars have done a lot of research on the technology of location, environment perception and automatic control of driverless vehicles, such as using lidar or monocular camera image recognition for environment perception; less attention has been paid to the external environment of driverless vehicles, especially the urban road environment. Road test in urban environment is the only way for driverless vehicles to enter the city, so this paper focuses on the construction of driverless vehicle urban road test scene, providing reference for driverless vehicle simulation test and high-precision map.

On the basis of understanding the key technologies of driverless vehicle, the research status at home and abroad, analyzing the necessary elements of driverless vehicle urban road test scene, and mastering the 3DMAX modeling function and steps. In this paper, the layout and proportion of the three-dimensional model of the road test area for driverless vehicles are designed by selecting the corresponding urban road and street environment, referring to the panoramic map, field data and road traffic standards. Using 3DMAX modeling lane, zebra crossing, safety island, traffic lights, vegetation, buildings and other scene elements, simulate the real urban road scene, and carry out panoramic synthesis for a typical scene, providing a reference for the simulation of driverless scene. Finally, it discusses the representativeness of the scene and the feasibility of the scene as an unmanned road test, points out the shortcomings of the scene, looks forward to the direction of in-depth study, and puts forward the development suggestions for the mutual adaptation of the unmanned vehicle and the urban traffic combined with the existing road traffic standards and the urban traffic environment.

请支付后下载全文，论文总字数：30040字