时空轨迹数据驱动的自动驾驶场景元建模方法

时空轨迹数据驱动的汽车自动驾驶场景建模,是当前汽车自动驾驶领域中驾驶场景建模、仿真所面临的关键问题,对于提高系统的安全性具有重要的研究意义.近年来,随着时空轨迹数据建模及应用研究的快速发展,时空轨迹数据应用于特定领域建模的研究引起人们的广泛关注.但是,由于时空轨迹数据所反映的现实世界的多元性和复杂性以及时空轨迹数据的海...     

基于VRML的虚拟汽车驾驶系统

介绍了基于VRML的汽车虚拟驾驶系统的设计，采用3DsMax建立城市建筑、街道、路灯、道路维护设备及标志等静态物体的三维几何模型，将这些几何模型装配成城市模型，将其转换成VRML格式，然后，采用ParallelGraphies的InternetSceneAssembler工具装配具有动态行为的几何模型并进行动态行为建模，如交通信号灯、汽车、行人。该工具使用基于VRML语言的可视化界面进行建模，使用很方便，效率高。动态行为使用VRML2．0结合JavaScript进行编程。开发的虚拟汽车驾驶系统可以在Internet上发布和游览。     

面向智能驾驶测试的仿真场景构建技术综述

随着汽车智能化程度的不断提高,智能汽车通过环境传感器与周边行驶环境的信息交互与互联更为密切,需应对的行驶环境状况也越来越复杂,包括行驶道路、周边交通和气象条件等诸多因素,具有较强的不确定性、难以重复、不可预测和不可穷尽。限于研发周期和成本、工况复杂多样性,特别是安全因素的考虑,传统的开放道路测试试验或基于封闭试验场的测试难以满足智能驾驶系统可靠性与鲁棒性的测试要求。因此,借助数字虚拟技术的仿真测试成为智能驾驶测试验证一种新的手段,仿真场景的构建作为模拟仿真的重要组成部分,是实现智能驾驶测试中大样本、极限边界小概率样本测试验证的关键技术,这对提升智能驾驶系统的压力和加速测评水平显得尤为重要。面向智能驾驶测试的仿真场景构建技术已成为当前汽车智能化新的研究课题和世界性的研究热点,作为一种新兴技术仍面临许多挑战。本文提出了面向智能驾驶测试的仿真场景构建方法,系统阐述了国内外研究工作的进展与现状,包括场景自动构建方法和交通仿真建模方法,重点分析一些值得深入研究的问题并围绕场景构建技术的发展趋势进行了讨论分析,最后介绍了团队相关研究在2020中国智能驾驶挑战赛仿真赛和世界智能驾驶挑战赛的仿真场景应用情况。     

基于图象的实时绘制技术

随着计算机技术的发展，对虚拟场景的真实感程度提出了越来越高的要求．基于 图象实时绘制技术是实现对虚拟场景建模和实时绘制的新的有效方法,可以克服传统的基于 几何绘制方式中建立真实模型的困难和实时绘制计算量大的缺陷．本文对目前出现的各种基 于图象的实时绘制方法进行分类介绍和综述．     

基于单目视觉的高速公路车道保持与距离测量

汽车智能辅助操作系统是智能交通系统的重要组成部分，也是目前国际上的热点问题．基于计算机视觉技术的车道保持，即视觉导航是辅助操作系统的核心问题之一．该文提出一种基于单目视觉的公路分道线二维重建算法和距离测量算法．其基本思想是根据分道线灰度与几何特征建立约束方程，进而跟踪提取分道线，并用二维模型重建，在此基础上，首先获得车道保持的必要道路参数，然后在二维重建的基础上，进一步实现了基于分道线几何约束的单目测距算法．在四川省与重庆市的高速公路上的实验证明，车道保持算法的实时性与精确性已基本满足实用要求；单目测距算法的精确度也能满足安全行车防止碰撞的需要．     

后视镜可视化在汽车驾驶仿真系统中的实现

在汽车驾驶仿真系统设计问题的研究中,针对汽车驾驶仿真系统的视景显示与驾驶操纵滞后脱节的现象,提出一种采用纹理映射技术的后视镜实现方法。由车辆运动模型得到车辆在任意时刻下的位置坐标。采用四边形条连接各个顶点形成后视镜面。根据后视镜与车身相对位置的不变性,通过矩阵变换和坐标变换计算构成后视镜面的顶点数组。定义一个指向后视镜的指针并封装为R函数,通过实时替换纹理提高了渲染场景的绘制效率和仿真系统的实时性。采用Vega Prime建立汽车模拟驾驶视景仿真并进行虚拟驾驶试验。模拟驾驶仪的仿真结果表明,后视镜具有良好的视觉效果,并能达到视景仿真的要求。     

基于Creator的汽车试验场的建立

基于虚拟现实技术开发的汽车仿真试验场具有真实、廉价、实用性强等特点.文中详细介绍了使用Creator软件构建汽车试验场的过程和方法,论述了高质量场景的建模技术,对场景的地形构造、纹理制作及映射、道路模型的构造等建模技术进行了探讨.提出了场景数据的优化显示技术和场景模型简化方法,提高了场景模型在仿真运行过程中的实时效率.最后给出了汽车仿真试验场建模的实例,并探讨建模过程中应注意的问题.     

三维场景编辑器的设计与实现

汽车驾驶训练模拟器是一种能正确模拟汽车驾驶操作，并能在主要性能上获得与实车同样感觉的汽车驾驶训练仿真教学设备。三维场景编辑器是三维模拟驾驶软件中的一个子系统。本文描述了三维场景编辑系统Ｍａｐ Ｅｄｉｔｏｒ的设计与实现。Ｍａｐ Ｅｄｉｔｏｒ与主系统的接口一定格式的地景文件，是提供给用户进行地景计算机辅助设计的工具。它的实现给出了一种快速实现图形编辑的方法。实践表明，这是一种有效的方法。     

虚拟驾驶系统中智能汽车创建方法的研究

虚拟驾驶系统中智能自主汽车的驾驶水平对整个仿真系统的逼真性以及测试的可信性起着决定作用。本文通过对智能交通环境的研究，阐述了在虚拟驾驶仿真系统中智能汽车的建立方法．提出了利用数据库储存虚拟交通环境的道路信息，实现智能汽车对道路的识别。本文还提出了基于碰撞检测原理的视觉信息获取方法以及基于视觉信息的实时决策算法。实验结果表明，按照以上的方法，智能汽车能较真实的模拟人的视觉，做出人性化的驾驶行为，增强了智能汽车在复杂交通网络中行驶的正确性以及系统的快速性、稳定性,效果令人满意。     

基于UE4的在线驾驶仿真系统设计与实现

驾驶仿真系统在交通安全研究中发挥着重要作用,但目前仍存在三维道路建模受限、仿真场景还原度不高等问题;研究采用虚幻引擎开发仿真功能模块,根据道路几何线形与交通流原始资料,能高效搭建出驾驶仿真场景;采用像素流送技术将视频帧发送到前端,设置硬件设备的轴属性映射,支持模拟驾驶外部信号输入,并实时将驾驶行为数据存入数据库,设计了一种轻量化、自动化、智能化的在线驾驶仿真信息管理与可视化系统;系统主要实现了道路自动化建模、人-车-路-环境信息管理、车道级真实交通流场景还原、在线模拟驾驶与数据可视化等功能;经测试,系统部署后能在Web端稳定运行,弥补了现有驾驶仿真系统的不足,为实验人员提供轻量高效的驾驶仿真服务。     

基于空间六点不变量的道路立体目标识别方法

城市道路中常设置具有3D效果的平面路障或标志物,其具有高度的立体性和真实性,导致行人和辅助驾驶系统误判而造成严重事故,因此需要对道路立体目标进行识别,以获得真实路面情况。常见的射影不变量如交比是基于共面五点计算的,存在局限性,论文提出一种基于空间点元素的几何不变量计算方法,把空间元素的共点和共线用具有物理意义的量来表示,通过合理搭建把空间元素巧妙转化为共面关系。该几何不变量只依赖点的空间坐标,与投影视角、摄像机参数等无关。在三维造型软件SolidWorks上进行模拟实验,并用该计算方法对真实道路上的真假3D物体进行了测验,最终实测实验验证依据空间六点计算得到的几何量可用于道路立体目标的识别。     

Receding horizon maneuver generation for automated highway driving

This paper focuses on the problem of decision-making and control in an autonomous driving application for highways. By considering the decision-making and control problem as an obstacle avoidance path planning problem, the paper proposes a novel approach to path planning, which exploits the structured environment of one-way roads. As such, the obstacle avoidance path planning problem is formulated as a convex optimization problem within a receding horizon control framework, as the minimization of the deviation from a desired velocity and lane, subject to a set of constraints introduced to avoid collision with surrounding vehicles, stay within the road boundaries, and abide the physical limitations of the vehicle dynamics. The ability of the proposed approach to generate appropriate traffic dependent maneuvers is demonstrated in simulations concerning traffic scenarios on a two-lane, one-way road with one and two surrounding vehicles.     

On line background modeling for moving object segmentation in dynamic scenes

Fast and accurate moving object segmentation in dynamic scenes is the first step in many computer vision applications. In this paper, we propose a new background modeling method for moving object segmentation based on dynamic matrix and spatio-temporal analyses of scenes. Our method copes with some challenges related to this field. A new algorithm is proposed to detect and remove cast shadow. A comparative study by quantitative evaluations shows that the proposed approach can detect foreground robustly and accurately from videos recorded by a static camera and which include several constraints. A Highway Control and Management System called RoadGuard is proposed to show the robustness of our method. In fact, our system has the ability to control highway by detecting strange events that can happen like vehicles suddenly stopped in roads, parked vehicles in emergency zones or even illegal conduct such as going out from the road. Moreover, RoadGuard is capable of managing highways by saving information about the date and time of overloaded roads.     

车路协同路侧感知融合方法的研究

随着智能网联汽车技术和产业的不断发展,智能网联汽车逐渐成为人们交通出行的选择之一。但受智能网联汽车自身环境感知系统对特定道路交通场景信息处理的局限,无法实现在所有行驶工况下安全高效的运行,其需车路协同路侧感知技术的辅助方能更安全高效的运行。海量的车路协同感知数据是城市道路和高速公路车路协同、运行分析和科学管理的宝藏,理解和分析这些数据是车路协同路侧感知融合的关键。面对车路协同路侧多传感器的不同数据,如何高效准确地挖掘和提取雷达、视频在不同时间、不同空间维度的数据,实现对重点交通场景(如视野盲区、急转弯道、隧道、桥梁)和交通事件、环境、设施安全等的雷达、视频数据进行快速融合检测、识别与检索,通过蜂窝车联网C-V2X网络在一定时延范围内有效地将路侧感知融合结果数据发送给智能网联汽车,确保其安全高效的行驶,是面向智能网联汽车辅助驾驶的车路协同路侧感知融合的关键问题。基于智能网联汽车其自身环境感知能力,对道路智能基础设施感知网络中的多传感器融合方法进行研究分析,提出了基于误差方差的多传感器融合算法,与非智能道路相比,其效率更高,更加智能化,可有效解决道路交通运行环境中存在的常见问题,为人们提供更加安全、高效、优质的交通出行服务。     

结合张量投票和Snakes模型的SAR图像道路提取

目的 Snakes模型对曲线轮廓具有良好的拟合能力,被广泛应用于遥感图像的道路提取。但SAR图像受乘性斑点噪声影响严重,因此利用Snakes模型从SAR图像提取道路时,传统的以图像灰度负梯度为外部能量的方法难以取得理想结果。针对这一问题,利用计算机视觉中的张量投票算法可以从噪声掩盖的图像中提取显著结构特征的特点,将张量投票与Snakes模型结合从SAR图像提取道路。方法 首先利用模糊C均值分割法从SAR图像中分割出道路类,然后对道路类进行张量投票获得每点的曲线显著性值,最后以该曲线显著性值的负值作为Snakes模型外部能量从SAR图像提取道路。在Snakes模型能量最小化阶段,提出了一种优化的拟合策略,一边内插节点一边最小化Snakes模型能量。结果 利用机载和星载不同场景的SAR图像进行实验,与同类的基于Snakes模型的半自动方法相比,本文方法对曲率较大的道路仅需较少控制点即可取得较好的拟合效果;与基于MRF模型的自动方法相比,本文方法对道路提取的完整率、正确率、检测质量都优于基于MRF模型的方法,并且提取的时间远远快于基于MRF模型的方法,对于大范围的道路网提取将更为实用。结论 本文方法充分考虑到道路的几何形态特征,利用张量投票算法对该特征进行量化,并利用优化的拟合策略来最小化Snakes模型能量来提取道路。基于机载和星载SAR图像的实验表明本文方法可以较好地提取不同场景中的主要道路目标和道路网。     

一种高分辨率遥感影像道路提取方法

针对空间异质性导致的道路几何纹理特征突出性下降问题, 提出一种高分辨率遥感影像道路提取方法. 首先设定跟踪模型, 依据人工输入点, 自适应提取道路中心点和道路宽度, 设计迭代内插、双向迭代两种跟踪方式以及矩形跟踪模板; 然后提出多描述子道路匹配模型, 针对道路几何纹理特征突出性不足问题, 基于道路区域地物边缘与道路方向一致的语义关系, 通过线段峰值约束的思想, 提出一种多尺度线段方向直方图(Multi-scale line segment orientation histogram, MSLSOH)描述子, 以此对跟踪方向进行预测; 针对道路几何纹理特征均质性下降问题, 从道路区域与道路非道路混合区域纹理差异性出发, 组合三角形构成扇形描述子, 突出道路影像纹理特征, 以此不仅可对预测跟踪点进行验证, 而且也可在结构信息缺失的情况下对道路进行跟踪; 最后选取不同类型、不同分辨率、不同场景的高分辨率遥感影像, 通过与其他方法的实验对比, 表明该方法能够解决道路提取过程中几何纹理特征突出性下降问题, 具有准确率高和自动化程度高的优势.     

Data Dependent Thin Plate Energy and its use in Interactive Surface Modeling

When modeling spline surfaces of complex shape, one has to deal with an overwhelming number of control points. Modeling by direct manipulation of the control points is a tedious task. In particular, it is very difficult to maintain a generally pleasant looking surface shape. It becomes therefore increasingly important to build tools that allow the designer to specify only a few geometric constraints while automatically determining the explicit representation of the surface. The basic concept of such a tool is simple. In a first step one has to somehow measure the “fairness” (=quality) of a surface. Once this is achieved, an optimization process selects the one surface with optimal fairness from all surfaces satisfying the user specified geometric constraints. To measure the fairness, thin plate energy functionals are a good choice. However, for interactive use these functionals are far too complex. W e will present appropriate approximations to these functionals that allow an optimization nearly in real time. Thefunctionals are obtained by introducing reference surfaces thus leading to data dependent, quadratic approximations to the exact thin plate energy functionals. We apply the method to interactive surface manipulations based on energy constraints.     

Dynamic PDE-based surface design using geometric and physical constraints

PDE surfaces, which are defined as solutions of partial differential equations (PDEs), offer many modeling advantages in surface blending, free-form surface modeling, and specifying surface’s aesthetic or functional requirements. Despite the earlier advances of PDE surfaces, previous PDE-based techniques exhibit certain difficulties such as lack of interactive sculpting capabilities and restrained topological structure of modeled objects. This paper presents an integrated approach that can incorporate PDE surfaces into the powerful physics-based modeling framework, to realize the full potential of PDE methodology. We have developed a prototype system that allows interactive design of flexible topological surfaces as PDE surfaces and displacements using generalized boundary conditions as well as a variety of geometric and physical constraints, hence supporting various interactive techniques beyond the conventional boundary control. The system offers a set of sculpting toolkits that allow users to interactively modify arbitrary points, curve spans, and/or regions of interest across the entire PDE surfaces and displacements in an intuitive and physically meaningful way. To achieve real-time performance, we employ several simple, yet efficient numerical techniques, including the finite-difference discretization, the multigrid-like subdivision, and the mass-spring approximation of elastic PDE surfaces and displacements. In addition, we present the standard bivariant B-spline finite element approximations of dynamic PDEs, which can subsequently be sculpted and deformed directly in real-time subject to the intrinsic PDE constraints. Our experiments demonstrate many attractive advantages of the physics-based PDE formulation such as intuitive control, real-time feedback, and usability to both professional and common users.     

遥感道路的场景感知与分类检测

基于人类视觉的感知组织和分类融合的基本原理,对遥感图像道路目标的感知模型进行了改进,改进模型分为像素、基元、结构和目标4个处理层次,比传统模型增加了道路场景的自动分类和多类型道路的整合连通2个子过程.基于该模型,提出了一种遥感图像道路检测算法,根据不同的道路场景定义了块状和线状2种不同的道路基元,并分别利用不同方法加以提取、度量和连接.2种基元的连接都采用了显著度驱动的层次性搜索策略.所有连接的路段最终都以中心主线的形式统一描述,并被整合为平滑的全局道路曲线.实验结果表明,该算法能从实际的卫星图像中检测出多种类型的道路,并具有较好的适用能力和有效性.     

Road-edge delineation in rural areas: effects on driving behaviour

When driving on lower-category Dutch rural roads without any delineation, drivers are likely to drift off the road with their right-side wheels, thus incurring damage to the pavement edge or even leading to accidents. In two experiments, two types of road-edge delineation, with continuous or dashed edge lines, were compared with two control roads without lines or with only a dashed line on the road axis. The first experiment consisted of non-obtrusive video recordings of passing traffic. Vehicle position on the experimental roads was more to the road's centre than on the control roads. The second experiment was a driving test with an instrumented vehicle, during daytime lighting and during darkness. Again, vehicle lateral position was more central on the experimental roads, especially during darkness. Subjects could safely pass oncoming vehicles. Driving speed increased on the experimental roads compared with the unlined control road, but not beyond speeds found on the axis-lined control road. Driver's mental effort while driving over the experimental roads did not differ from the effort while driving over the control roads. Subjectively rated effort was higher for the unlined control road than for the three other roads. Subjects preferred the edge-lined roads to the unlined control road, but not more than the axis-lined control road. It was concluded that edge-lines may provide a simple and effective way of inducing a more favourable lateral position on rural roads without having negative effects on subjective appraisal, driving performance or mental workload.