Speed and Accuracy Tradeoff for LiDAR Data Based Road Boundary Detection

Road boundary detection is essential for autonomous vehicle localization and decision-making, especially under GPS signal loss and lane discontinuities. For road boundary detection in structural environments, obstacle occlusions and large road curvature are two significant challenges. However, an effective and fast solution for these problems has remained elusive. To solve these problems, a speed and accuracy tradeoff method for LiDAR-based road boundary detection in structured environments is proposed. The proposed method consists of three main stages: 1) a multi-feature based method is applied to extract feature points; 2) a road-segmentation-line-based method is proposed for classifying left and right feature points; 3) an iterative Gaussian Process Regression (GPR) is employed for filtering out false points and extracting boundary points. To demonstrate the effectiveness of the proposed method, KITTI datasets is used for comprehensive experiments, and the performance of our approach is tested under different road conditions. Comprehensive experiments show the road-segmentation-line-based method can classify left, and right feature points on structured curved roads, and the proposed iterative Gaussian Process Regression can extract road boundary points on varied road shapes and traffic conditions. Meanwhile, the proposed road boundary detection method can achieve real-time performance with an average of 70.5 ms per frame.      

自动驾驶在拥堵路段的道路几何信息估计

道路几何信息是自动驾驶系统中重要的信息来源,也是后续路径规划的关键参考信息之一。 该研究针对城市内车道线遮挡及多路径效应导致的全球定位系统失效等问题,提出了一种基于前车信 息的道路几何估计方法。通过对当前车辆、前车以及道路之间关系的建模,获得了系统的运动模型和 观测模型。采用无损卡尔曼滤波框架对观测到的前车相对位置、相对速度、相对角度和本车角速度进 行滤波处理,估计出当前车道的曲率参数。在仿真软件 Car learning to Act(Carla)上的实验结果表明, 相比地图匹配方法,在无法获取车道线目标及精确定位信息的情况下,该方法道路几何精度得到了显 著提升。     

融合道路图像关键信息的车距视觉测量方法

提出了一种融合道路图像关键信息的纵向车距视觉测量方法,在道路成像平面内建立了基于车道平面约束的视觉测距模型,运用边界约束Hough变换采集两侧道路标识线的斜率、聚点坐标以及车道宽度信息,自动求解视觉传感器的高度及俯仰角等测距参数。选取双通道Gabor滤波器提取目标车辆的5尺度8方向特征样本,联合AdaBoost分类器与级联Cascade筛选有效特征,快速精确提取目标特征点的坐标参数。实验结果表明,视觉测量值与实测值的绝对误差平均值为1.37 m,相对误差平均值为2.38%,测距平均耗时32 ms,与传统测距方法相比较,测量精度和实时性均得到了提高,适合于汽车主动防撞安全系统中应用。     

Ego-Lane Analysis System (ELAS): Dataset and algorithms

Decreasing costs of vision sensors and advances in embedded hardware boosted lane related research – detection, estimation, tracking, etc. – in the past two decades. The interest in this topic has increased even more with the demand for advanced driver assistance systems (ADAS) and self-driving cars. Although extensively studied independently, there is still need for studies that propose a combined solution for the multiple problems related to the ego-lane, such as lane departure warning (LDW), lane change detection, lane marking type (LMT) classification, road markings detection and classification, and detection of adjacent lanes (i.e., immediate left and right lanes) presence. In this paper, we propose a real-time Ego-Lane Analysis System (ELAS) capable of estimating ego-lane position, classifying LMTs and road markings, performing LDW and detecting lane change events. The proposed vision-based system works on a temporal sequence of images. Lane marking features are extracted in perspective and Inverse Perspective Mapping (IPM) images that are combined to increase robustness. The final estimated lane is modeled as a spline using a combination of methods (Hough lines with Kalman filter and spline with particle filter). Based on the estimated lane, all other events are detected. To validate ELAS and cover the lack of lane datasets in the literature, a new dataset with more than 20 different scenes (in more than 15,000 frames) and considering a variety of scenarios (urban road, highways, traffic, shadows, etc.) was created. The dataset was manually annotated and made publicly available to enable evaluation of several events that are of interest for the research community (i.e., lane estimation, change, and centering; road markings; intersections; LMTs; crosswalks and adjacent lanes). Moreover, the system was also validated quantitatively and qualitatively on other public datasets. ELAS achieved high detection rates in all real-world events and proved to be ready for real-time applications.     

Trajectory planning and yaw rate tracking control for lane changing of intelligent vehicle on curved road

The lateral control for lane changing of intelligent vehicle on curved road in automatic highway systems was studied. Based on trapezoidal acceleration profile, considering the curvature difference between starting lane and target lane, a new virtual trajectory planning method for lane changing on curved road was presented, and the calculating formulas for ideal states of vehicle in the inertial coordinate system during a lane changing maneuver were established. Applying the predetermined trajectory, the re...     

无人驾驶车路径跟踪控制研究

研究被控对象无人驾驶车,基于预瞄控制思想,设计一种无人驾驶车路径跟踪控制器,将控制器分为预瞄控制和补偿控制两部分,预瞄控制模拟驾驶员在驾驶车辆过程中对前方的道路环境信息进行预瞄,根据道路曲率程度决定方向盘转向,补偿控制是对车辆遇到干扰偏离原车道的纠正。仿真实验结果表明,该控制器能够保证无人驾驶车准确跟踪各种参考路径,且具有较好的鲁棒性。     

A Fuzzy-Logic-Based Approach for Mobile Robot Path Tracking

One important problem in autonomous robot navigation is the effective following of an unknown path traced in the environment in compliance with the kinematic limits of the vehicle, i.e., bounded linear and angular velocities and accelerations. In this case, the motion planning must be implemented in real-time and must be robust with respect to the geometric characteristics of the unknown path, namely curvature and sharpness. To achieve good tracking capability, this paper proposes a path following approach based on a fuzzy-logic set of rules which emulates the human driving behavior. The input to the fuzzy system is represented by approximate information concerning the next bend ahead the vehicle; the corresponding output is the cruise velocity that the vehicle needs to attain in order to safely drive on the path. To validate the proposed algorithm two completely different experiments have been run: in the first experiment, the vehicle has to perform a lane-following task acquiring lane information in real-time using an onboard camera; in the second, the motion of the vehicle is obtained assigning in real-time a given time law. The obtained results show the effectiveness of the proposed method     

高速路的车道检测与车辆跟踪

基于智能交通的快速发展,研究了基于高速路的车道检测和车辆跟踪技术.对于多车道检测,根据路面与分道线灰度级相差较大的特点来实现车道路面的分割,接着结合直线方程和Catmull-Rom Spline插值算法来拟合分道线.对于单车道检测,首先基于HSV颜色空间和Sobel边缘提取方法对其进行有效分割,接着在透视变换空间中提取分道线坐标点并用二次多项式拟合分道线.针对车辆检测,使用Hog+Gentle-Adaboost分类算法实现无人车前方路面车辆的检测,接着基于车底阴影的特征对车底阴影进行检测以验证学习算法检测到的车辆区域的真伪性.针对车辆跟踪,采用动态二阶自回归模型的方法预测车辆的状态.其中,对于粒子滤波固有的粒子退化问题,引入Thompson_Taylor算法改善了粒子退化和低多样性的缺陷.本文的车道检测和车辆跟踪算法能较容易地移植在嵌入式平台,可靠性和准确性较高,且有助于进一步实现车道偏离报警和前向防撞系统.     

Lane-keeping assistance control algorithm using differential braking to prevent unintended lane departures

This paper describes a hierarchical lane keeping assistance control algorithm for a vehicle. The proposed control strategy consists of a supervisor, an upper-level controller and a lower-level controller. The supervisor determines whether lane departure is intended or not, and whether the proposed algorithm is activated or not. To detect driver′s lane change intention, the steering behavior index has been developed incorporating vehicle speed and road curvature. To validate the detection performance on the lane change intention, full-scale simulator tests on a virtual test track (VTT) are conducted under various driving situations. The upper-level controller is designed to compute the desired yaw rate for the lane departure prevention, and for the guidance with ride comfort. The lower-level controller is designed to compute the desired yaw moment in order to track the desired yaw rate, and to distribute it into each tire′s braking force in order to track the desired yaw moment. The control allocation method is adopted to distribute braking forces under the actuator’s control input limitation. The proposed lane keeping assistance control algorithm is evaluated with human driver model-in-the-loop simulation and experiments on a real vehicle.     

高速公路弯道边界的识别与重建

分析了智能车辆安全辅助驾驶系统中弯道分道线的检测提取方法,提出一种基于道路区域分割的弯道检测新算法,包含道路区域分割和弯道边界检测。在分割出道路区域和天空区域并划定弯道检测的感兴趣区域后,提取分道线候选点,并对候选点进行校正,最终拟合并重建出弯道分道线,且准确判断了车道线弯曲方向。实验证明,该算法的实时性和准确性均高于在整幅图像中提取车道线的传统方法。     

Road‐Centered Map‐Aided Localization for Driverless Cars Using Single‐Frequency GNSS Receivers

Accurate localization with high availability is a key requirement for autonomous vehicles. It remains a major challenge when using automotive sensors such as single‐frequency Global Navigation Satellite System (GNSS) receivers, a lane detection camera, and proprioceptive sensors. This paper describes a method that enables the estimation of stand‐alone single‐frequency GNSS errors by integrating the measurements from a forward‐looking camera matched with lane markings stored in a digital map. It includes a parameter identification method for a shaping model, which is evaluated using experimental data. An algebraic observability study is then conducted to prove that the proposed state vector is fully observable in a road‐oriented frame. This observability property is the basis to develop a road‐centered Extended Kalman filter (EKF) that can maintain the observability of every component of the state vector on any road, whatever its orientation. To accomplish this, the filter needs to handle road changes, which it does using bijective transformations. The filter was implemented and tested intensely on an experimental vehicle for driverless valet parking services. Field results have shown that the performance of the estimation process is better than solutions based on EKF implemented in a fixed working frame. The proposed filter guarantees that the drift along the road direction remains bounded. This is very important when the vehicle navigates autonomously. Furthermore, the road‐centered modeling improves the accuracy, consistency, and robustness of the localization solver.     

基于机器视觉的轻量级驾驶辅助系统

为了解决计算机视觉应用中数据量大、算法复杂的问题,根据道路结构特征和车辆行为特征,采用单个摄像头作为传感器,实现了一种轻量级的安全辅助驾驶系统。首先采用改进的边缘提取算法和车道线检测算法对摄像机内外参数进行离线标定;接着根据标定结果在二维平面图像上采用标识出实际空间距离的多窗口划分方法,并按不同的车间距将不同窗口划分为不同安全系数的区域,以赋予道路视觉检测的几何先验知识;当区域中出现障碍物时发出相应警示信息进行安全驾驶辅助,能为智能辅助驾驶提供轻量级的视觉检测平台。以便携式计算机和固定在车内的摄像头作为实验装置,在城市道路上进行车载实验。系统在车载实验中能够快速地提取车辆两侧的车道线,并利用离线标定的结果快速生成不同安全系数的警示区域,其中车辆在车道内正常行驶时的误检率和漏检率很小,可以忽略不计。与传统的驾驶辅助系统相比,本系统计算量大大降低,检测流程得到简化,可实现轻量级的车道和车辆检测,为系统在嵌入式系统上的实现奠定基础。     

非结构道路环境下的智能汽车质心侧偏角估计

车辆质心侧偏角是描述车辆侧向运动状态的重要参量之一,其估计的精度直接影响车辆的安全控制,传统的质心侧偏角估计方法不能满足非结构道路环境下的智能汽车质心侧偏角估计的要求。通过建立3自由度智能汽车动力学模型,采用CarSim和MATLAB构建智能汽车整车参数化模型;基于扩展kalman滤波(EKF)算法,设计非结构道路环境下的状态观测器对智能汽车质心侧偏角进行估计。在高、低附着系数路面双移线工况和蛇形工况下,对状态观测器的估计效果进行联合仿真验证。仿真结果表明:该方法能较精确地估计出非结构道路环境下智能汽车的质心侧偏角。     

A model-driven approach for real-time road recognition

This article describes a method designed to detect and track road edges starting from images provided by an on-board monocular monochromic camera. Its implementation on specific hardware is also presented in the framework of the VELAC project. The method is based on four modules: (1) detection of the road edges in the image by a model-driven algorithm, which uses a statistical model of the lane sides which manages the occlusions or imperfections of the road marking – this model is initialized by an off-line training step; (2) localization of the vehicle in the lane in which it is travelling; (3) tracking to define a new search space of road edges for the next image; and (4) management of the lane numbers to determine the lane in which the vehicle is travelling. The algorithm is implemented in order to validate the method in a real-time context. Results obtained on marked and unmarked road images show the robustness and precision of the method. Received: 18 November 2000 / Accepted: 7 May 2001     

一种快速的公交专用车道检测方法

为了提高车道线检测的准确性和实时性,提出了一种快速准确的车道线检测方法。首先根据道路的纹理特征求出道路的消失点,再采用改进的Hough变换检测出车道线,结合车道线的一些特征以及摄像头的参数,在不影响测量结果的情况下缩小检测空间,快速准确地检测道路的车道线,并结合BRT车道(快速公交车道)的一些特征识别车辆所在车道是否为BRT车道,从而实现对BRT车道内前方车辆的监督。将代码移植到DM6437平台,实验结果表明,该方法具备较好的实时性和鲁棒性。     

引入辅助损失的多场景车道线检测

目的 为解决实时车辆驾驶中因物体遮挡、光照变化和阴影干扰等多场景环境影响造成的车道线检测实时性和准确性不佳的问题,提出一种引入辅助损失的车道线检测模型。方法 该模型改进了有效的残差分解网络（effcient residual factorized network,ERFNet）,在ERFNet的编码器之后加入车道预测分支和辅助训练分支,使得解码阶段与车道预测分支、辅助训练分支并列,并且在辅助训练分支的卷积层之后,利用双线性插值来匹配输入图像的分辨率,从而对4条车道线和图像背景进行分类。通过计算辅助损失,将辅助损失以一定的权重协同语义分割损失、车道预测损失进行反向传播,较好地解决了梯度消失问题。语义分割得到每条车道线的概率分布图,分别在每条车道线的概率分布图上按行找出概率大于特定阈值的最大点的坐标,并按一定规则选取相应的坐标点,形成拟合的车道线。结果 经过在CULane公共数据集上实验测试,模型在正常场景的F1指标为91.85%,与空间卷积神经网络（spatial convolutional neural network,SCNN）模型相比,提高了1.25%,比其他场景分别提高了1%~7%;9种场景的F1平均值为73.76%,比目前最好的残差网络——101-自注意力蒸馏（ResNet-101-self attention distillation,R-101-SAD）模型（71.80%）高出1.96%。在单个GPU上测试,每幅图像的平均运行时间缩短至原来的1/13,模型的参数量减少至原来的1/10。与平均运行时间最短的车道线检测模型ENet——自注意力蒸馏（ENet-self attention distillation,ENet-SAD）相比,单幅图像的平均运行时间减短了2.3 ms。结论 在物体遮挡、光照变化、阴影干扰等多种复杂场景下,对于实时驾驶车辆而言,本文模型具有准确性高和实时性好等特点。     

面向复杂城市交通场景的一种实时车道线检测方法

车道线检测是智能交通监控及自动驾驶的基础步骤,为提高其鲁棒性和实时性,针对复杂城市交通场景中自动驾驶需要检测车道线的需求,提出了一种实时车道线检测算法,首先运用改进灰度化变换突显车道线的特征,并通过改进的Gabor滤波算法增强车道线的边缘信息;最后采用多约束霍夫变换筛选得到平行车道线从而实现实时车道线检测。实验表明,该方法在三种不同真实的交通道路场景下,提高了车道线检测精度及处理速度,可应用于实时车道线检测系统。     

面向交通信息采集的无线传感器网络节点

智能交通系统的关键技术环节之一是能够准确地获取实时交通参数,包括交通流量、车速、车道占有率等,无线传感器网络在智能交通方面有潜在的广泛应用前景.设计实现了面向交通信息采集的无线传感器网络节点,提出了一系列相关交通信息采集专用算法,包括基于数字滤波和匹配滤波的交通流量监测算法、车速测量算法和车辆识别算法,在道路上进行了实测验证并对节点功耗进行了分析.实测结果表明,交通信息采集节点能以较高精度得到交通流量、车速、车道占有率等信息,并能较准确地对机动车和自行车进行识别.     

基于Haar-like和MB-LBP特征分区域多分类器车辆检测*

高级辅助驾驶系统中的预碰撞系统不仅需要检测前向车辆,预防追尾碰撞,同时需要检测临近车道斜侧向车辆,实现对其潜在换道、合流行为的预测,提供实时的预警功能.文中提出分区域融合多种特征的车辆检测方法,解决前向车辆特征与斜侧向车辆特征存在明显差异的问题.同时,文中方法分别检测远近不同的车辆,进行不同程度的图像降采样,提高检测系统的实时性.多种交通场景的实地测试表明,文中方法可以实时、稳定、准确地检测到前向车辆和斜侧向车辆,在行车环境良好的情况下,具有较高的召回率和准确率.     

基于车道映射矩阵的多车道车辆计数系统

针对城市交通路口的多车道车辆数量统计问题,提出基于车道映射矩阵的车道划分方案。根据车道映射矩阵分布情况判断车道中心线位置,利用相邻车道中心线间的距离对车道进行划分,与车辆计数算法结合,实现多车道车辆计数系统。该方案能够减少环境因素对车道划分的影响,在白天与夜晚情况下均取得较好效果。实验结果表明,该方案准确率为93.6%,能够适应复杂道路,多车道实时车辆计数系统准确率能够达到91%以上,在保证系统实时性的基础上,提供了较高的系统准确性和鲁棒性。