Development of driver-vehicle-road closed-loop semi-physical simulation system

As hybrid vehicles introduced the motor, the vehicle structure has a significant change in the power matching. A driver-vehicle-road closed-loop semi-physical simulation system, which makes real driving parts together with the simulation car, will bring convenience to the new car design. We used the computer software to simulate the road with a slope, curve and some other features based on the actual road condition, and analyzed the whole road scene in addition to geometry and physical characteristics. Analyzing and constructing the vehicle dynamics basic template, appropriate changes to the template can obtain the desired vehicle dynamics model with an external device to control the model vehicle. It combined the physical operation system with visual display, which gave us real driving feelings and increased the vehicle design predictive accuracy.

Efficient Vision Transformers for Autonomous Off-Road Perception Systems

The development of autonomous vehicles has become one of the greatest research endeavors in recent years. These vehicles rely on many complex systems working in tandem to make decisions. For practical use and safety reasons, these systems must not only be accurate, but also quickly detect changes in the surrounding environment. In autonomous vehicle research, the environment perception system is one of the key components of development. Environment perception systems allow the vehicle to understand its surroundings. This is done by using cameras, light detection and ranging (LiDAR), with other sensor systems and modalities. Deep learning computer vision algorithms have been shown to be the strongest tool for translating camera data into accurate and safe traversability decisions regarding the environment surrounding a vehicle. In order for a vehicle to safely traverse an area in real time, these computer vision algorithms must be accurate and have low latency. While much research has studied autonomous driving for traversing well-structured urban environments, limited research exists evaluating perception system improvements in off-road settings. This research aims to investigate the adaptability of several existing deep-learning architectures for semantic segmentation in off-road environments. Previous studies of two Convolutional Neural Network (CNN) architectures are included for comparison with new evaluation of Vision Transformer (ViT) architectures for semantic segmentation. Our results demonstrate viability of ViT architectures for off-road perception systems, having a strong segmentation accuracy, lower inference speed and memory footprint compared to previous results with CNN architectures.

基于UC-win/Road的城市高架路建模及应用

随着城市的飞速发展和道路基础设施的完善,城市高架路作为主要城市快速路之一,已成为国内外城市解决交通问题的重要手段之一。利用UC-win/Road软件及其模拟器可以实现城市高架路的模拟仿真。以南京市宁溧高架为例,设计资料为平、纵、横参数,在UC-win/Road平台上进行建模并实现模拟驾驶,最后对其应用前景作出展望。

The Belt and Road Initiative from Three Theoretical Perspectives

Efforts have generally been made from policy considerations to interpret reasons for the proposition of the Belt and Road Initiative and its global layout.Though they are of some significance,such a grand initiative certainly has its own internal theoretical basis.This paper is in the opinion that the spillover effect of oversized economy,the extension of the value chain and its climbing effect,as well as channel suitability and location orientation constitute the endogenous driving forces of the Belt and Road Initiative.Each of the three theoretical perspectives has its own internal realistic basis,all of which may offer some reference that is conducive to understanding the practice and future development of the Belt and Road Initiative.

Progressive LiDAR Adaptation for Road Detection

Despite rapid developments in visual image-based road detection, robustly identifying road areas in visual images remains challenging due to issues like illumination changes and blurry images. To this end, LiDAR sensor data can be incorporated to improve the visual image-based road detection,because LiDAR data is less susceptible to visual noises. However,the main difficulty in introducing LiDAR information into visual image-based road detection is that LiDAR data and its extracted features do not share the same space with the visual data and visual features. Such gaps in spaces may limit the benefits of LiDAR information for road detection. To overcome this issue, we introduce a novel Progressive LiDAR adaptation-aided road detection(PLARD) approach to adapt LiDAR information into visual image-based road detection and improve detection performance. In PLARD, progressive LiDAR adaptation consists of two subsequent modules: 1) data space adaptation, which transforms the LiDAR data to the visual data space to align with the perspective view by applying altitude difference-based transformation; and 2) feature space adaptation, which adapts LiDAR features to visual features through a cascaded fusion structure. Comprehensive empirical studies on the well-known KITTI road detection benchmark demonstrate that PLARD takes advantage of both the visual and LiDAR information, achieving much more robust road detection even in challenging urban scenes. In particular, PLARD outperforms other state-of-theart road detection models and is currently top of the publicly accessible benchmark leader-board.

An Investigation of Drivers’ Comprehension of Distracted Driving Slogans

Distracted driving occurs when a driver diverts the primary attention from driving to another task. Using mobile devices such as a cellphone for texting, calls, or other manipulation while driving has the highest potential for distraction because it combines both forms of distractions, manual, visual, and cognitive. Some states in the US have posted slogans including “W8 2 TXT, it’s a law”, “Don’t Drive inTEXTicated”, “PLS dnt txt n drv”, “Don’t tempt F8 that txt can w8”, and “DNT TXT & DRV” along highways to convey the dangers and laws regarding distracted driving to minimize incidences of distracted-related crashes This study surveyed 347 people using the five distraction slogans in a college town. The results showed that younger drivers have a higher level of comprehension compared to older drivers. Further, the results showed that drivers with university education or more years of driving experience have a higher comprehension level of distraction signs compared to their counterparts.

Three-year caregiver’s reports on driving performance: A divination or diagnostic tool for fitness to drive in Parkinson Disease patients

Objectives: This study aims to examine whether information provided by spouse or relatives can be employed to identify PD patients with deteriorated driving performance, using three-year caregiver’s reports on their driving ability as the outcome measure. Methods: Fifty-three idiopathic PD subjects were assessed on open roads. Prior to the driving assessment, participants were examined by a geriatrician with various clinical assessments. The caregivers filled out a questionnaire, the scores of which is a reflection of their concern on driving performanceof their PD relatives. The same measurements were collected for the subsequent two years. Hierarchical Poisson regression analysis, adjusting for gender, age and driving exposure (hours of driving per week), was then undertaken to determine whether the measures of driving assessment were associated with the score of the questionnaire. Results: During the three-year period, all PD participants were rated at least 3 questions positive in the caregiver’s questionnaire;the worst participant was rated positive eight times. Except the assessment criteria to gauge thetraffic rulesandregulations compliance,all other measures of the driving assessment were found to be significantly associated with the information provided by the caregivers. Conclusions: This study demonstrated that the information provided by caregivers was useful to identify PD patients with deteriorated driving performance. If adopted as part of the off-road driving assessment for PD patients, the questionnaire can provide reliable information to clinicians.

考虑道路识别的四驱电动汽车再生制动策略

为了充分利用路面附着条件分配再生制动力,提高制动能量回收率,根据双电动机四驱结构电动汽车复合制动系统的特点,对其制动能量回收控制策略进行了研究.分析了驱动轮与地面附着特性,设计了道路识别器来计算每个轮胎与当前路面之间的峰值附着系数,并使用模糊控制策略确定每个车轮与8种道路的滑移率和附着利用率的相似输入.根据双电动机外部特性的制动力分配关系,提出了使更多制动能量回馈给电池的策略.结果表明:制动强度为0.3时,能量回收率为65.55%,具有较高的回收效率.

基于驾驶风格的山区公路穿村镇段行车风险场灵敏度分析

针对山区公路穿村镇段交通事故多发的现实问题,本文提出一种考虑驾驶风格的行车风险识别方法,通过灵敏度分析揭示驾驶风格在风险评估中的影响,并以云南省典型山区公路穿村镇段为实例进行验证。首先,利用无人机采集车辆轨迹数据,建立轨迹分析数据库;其次,基于行车风险场理论并通过事故数据的相关性分析,刻画穿村镇段行车环境对车辆行驶的风险影响,进一步引入驾驶风格因子,建立考虑驾驶风格的行车风险场,实现多要素的综合考虑与行车风险量化;最后,基于Sobol全局灵敏度分析方法,分析考虑驾驶风格前后模型关键参数的全局灵敏度,并实现区域风险的可视化。研究结果表明:当选择速度均值、加速度均值和冲击度方差进行驾驶风格划分,K值为4时聚类效果最佳;Sobol法有效评估了参数灵敏度,当考虑驾驶人行为场后,总灵敏度分布更加均匀,模型考虑因素更为广泛,其中速度和冲击度标准差为最显著参数,总灵敏度分别为0.41和0.34;对多车共同作用产生的风险可视化显示,势场范围随道路线形变化而变化,在穿村镇交叉口风险值最显著,整体行驶速度降比为17.39%,将进一步发挥风险场理论在微观交通风险评估中的作用。

基于行驶特征区间划分的重型载货汽车实际道路能耗评价方法

提出一种基于行驶特征区间划分的能耗评价方法。通过对中国货车行驶工况的工况点分布特征进行分析,确定了速度和加速度区间的数量与边界的划分方案。考虑道路坡度的影响,进行实际道路试验数据筛选,并用于计算不同行驶特征区间内的燃油消耗量平均值。结合中国货车行驶工况在不同行驶特征区间内的累计行驶里程,计算能耗评价结果。为了验证该方法的有效性,开展了3次底盘测功机试验和2次实际道路行驶试验,结果表明,所提出的能耗评价方法能提高重型载货汽车实际道路能耗评价结果的复现性。

草原道路驾驶员个性心理特征对疲劳的影响

为科学预防疲劳驾驶,采用模拟实验进行草原道路长时程行车,运用MP150多通道生理信号采集仪,采集行车过程中驾驶员的脑电(EEG)信号,定量分析急躁型、谨慎型和常规型驾驶员在不同行车时段的疲劳特性。结果表明:α波、β波这两项指标对反应急躁型驾驶员疲劳状态最为敏感;θ波在评价谨慎型驾驶员疲劳状态时最敏感;在30 min~40 min驾驶员疲劳反应初次显现,120 min~140 min为自我调整临界点,此后自我调节效果不明显;在驾驶任务中,急躁型驾驶员最先出现疲劳,随着时间的累积,谨慎型驾驶员的疲劳程度远高于其他两类驾驶员,说明这类驾驶员缺少持久性,不宜长时程驾驶。

自动-人工驾驶车辆混行下路网可靠性研究

随着自动驾驶技术的快速发展,自动和人工驾驶车辆混行情况将不可避免,对于自动-人工驾驶混行环境下的交通特性亟待研究。交通路网可靠性可以更全面地反映交通流运行状态,而现有交通路网评价模型未能充分考虑突发事件的干扰,难以准确评估自动-人工混行条件下路网的可靠性。为此,在现有模型基础上,考虑路网受到干扰下的恢复属性,引入道路韧性指数,改进现有路网可靠性模型。基于北京市永乐店镇实测路网数据,利用Vissim构建自动驾驶条件下混合交通流仿真模型,对不同自动驾驶渗透率下的路网可靠性进行研究。结果表明:考虑道路韧性指数的可靠性评价模型可以有效降低最小误差平方和,显著提高路网可行性评价准确性;在自由流状态下,随着渗透率增加,路段可靠性显著提升,全自动驾驶情形下路段可靠性可提升18%;在重度阻塞流状态下,当渗透率超过40%后,路段可靠性呈上升趋势,全自动驾驶情形下路段可靠性可提升27.8%;当加入较小比例的自动驾驶车辆时,路网可靠性改善并不明显,但当渗透率超过50%时,该路网可靠性逐渐提升,全自动驾驶情形下路网的可靠性可提升23%。

基于设计运行条件的实际道路测试场景复杂度计算模型

自动驾驶车辆在实际道路行驶的安全性与外部交通要素、驾驶员状态以及本车状态密切相关。自动驾驶面临的一大难题是实际交通环境具有道路形态、自然环境、交通参与者及事件等在时间和空间维度的随机性特征,测试场景复杂度存在的差异造成了自动驾驶测试过程的不可复现性和测试结果的不可对比性,导致自动驾驶评价缺乏统一量化的测试环境基准。本文提出了一种基于设计运行条件的实际道路测试场景复杂度计算模型,系统考虑了网联功能、驾驶员感知能力、本车执行能力对自动驾驶车辆在实际道路中面临相关场景时的复杂度影响,建立了以道路等级、交通设施、临时交通变化、交通参与者、自然环境、网联信息、驾驶员状态和本车状态等8个要素大类为基础的自动驾驶实际道路测试场景复杂度计算模型要素库,建立了基于设计运行条件和层次分析法的自动驾驶功能实际道路测试场景复杂度计算模型。采用基于智能化网联化的技术逻辑影响传递机制计算场景要素权重系数,并在实际道路测试中验证了本文方法的可行性和合理性。

城市生态道路混合交通流节能驾驶策略优化

针对生态道路自动网联车在混合交通流条件下的驾驶节能问题,提炼对于生态道路节能驾驶问题影响较大的野生动物通道场景,构建面向车联网的野生动物通道应用框架;并构建生态道路车联网环境下的车辆驾驶模型,运用动态规划进行离散化分析和状态划分,优化建立混合交通流车辆节能驾驶模型;通过强化学习Q-learning算法,对单辆汽车的节能驾驶模型进行优化求解;基于上海城市生态道路,建构考虑野生动物通道动物穿行风险的仿真场景,开展对车联网环境下的混合交通流节能驾驶策略仿真验证。结果表明该节能策略在车联网环境下能够使生态道路上车辆节省油耗量在6%~11%之间,并且节能效果将随着混合车流密度的增加而更优,验证了模型的合理性,以及算法求解的有效性。

基于改进DEMATEL-ISM模型的道路拥堵下行车风险评估方法

为准确识别城市道路拥堵条件下的行车风险系统中各影响因素之间的内在联系以及该系统中的关键因素,本文提出一种改进DEMATEL-ISM的评估方法。首先,分析城市道路拥堵条件下的行车风险,选取18个风险因子;其次,运用多群体问询的调查问卷方法,并结合最大类间方差法对标准的集成决策试验和评价试验室(DEMATEL)以及解释结构模型(ISM)改进,建立城市道路拥堵的行车风险评估模型。最后,实例验证结果表明:基于改进DEMATEL-ISM模型的道路拥堵条件下行车风险评估体系能准确划分层次解释结构图的层级,与其他方法相比更加符合实际;交通密度、车头时距、停车次数、换道频次等关键链接因素对其他因素有较强联动影响,并且会增加发生交通事故或二次交通拥堵的概率。

自然驾驶状态下城市地下环道驾驶人心理负荷研究

为探究城市地下环道驾驶人心理负荷特性,组织驾驶人自然驾驶试验,使用生理仪获取自然驾驶状态下驾驶人心电信号数据。以驾驶人心率、心率增长率和心率变异性指标为关键变量,基于因子分析法构建心理负荷量化模型,揭示地下环道平曲线半径、曲线类型、出入口以及熟练程度对驾驶人心理负荷的影响。研究表明:心率指标和心率变异性指标分别与地下环道平曲线半径呈负相关和正相关关系;驾驶人心理负荷与平曲线半径呈负相关,在相同半径条件下,连续曲线场景下的心理负荷显著高于单个曲线场景,且连续曲线间负荷累积作用超过单一曲线半径对驾驶人心理负荷的影响;入口段驾驶人心理负荷比出口段高21.39%;非熟练驾驶人心理负荷比熟练驾驶人高30.75%;建议在城市地下环道出入口和曲率变化区段增设视线诱导设施,设计车速为20~40 km·h^(-1)时,平曲线半径不宜小于30 m。

点云空间与反射强度融合的非结构化道路可行驶区域检测

可行驶区域检测旨在检测和提取智能车辆在道路上可行进的区域,目前主流的检测方法主要基于三维激光雷达的空间特征实现,难以处理路面边缘无清晰空间特征的非结构化道路.为此,本文提出了一种基于点云空间和反射强度融合的非结构化道路可行驶区域检测方法.首先,通过融合反射强度因子改进了基于空间特征的柱坐标系检测模型;然后,使用强度和降维空间检测对检测精度较低的环形检测模型进行优化,并将其与柱坐标系检测模型联合使用以提高方法检测准确率;最后,在自录实际道路数据集上进行对比实验.实验结果表明本文方法显著提高了非结构化道路可行驶区域检测的成功率与精确率,在结构化道路上也具有良好效果.

分布式驱动电动汽车状态与路面附着系数估计

【目的】准确获取车辆行驶状态参数和路面附着系数对于维持分布式驱动电动汽车的安全性和稳定性十分重要,针对于此,研究了一种能够准确估计其参数的算法,既能降低量产车辆的成本,又能达到车辆安全稳定的目的。【方法】首先,建立三自由度车辆模型,应用基于奇异值分解方法(SVD)的容积卡尔曼滤波算法搭建参数估计器。然后,搭建Carsim/Simulink联合仿真实验平台,采用高、低附着和对接路面双移线工况进行仿真验证。【结果】仿真实验结果表明,估计器能够较准确地估计出参数的真实值。基于仿真的良好估计效果,搭建实车实验平台,进一步进行验证得出,实车实验结果与仿真结果趋势一致,该算法具有较好的估计准确性和时效性。

公路缓和曲线长度的合理取值分析

为给车辆由直线(圆曲线)驶入圆曲线(直线)提供合理的路径长度,确保驾驶员能以安全、舒适、自然的状态驶入目标曲线单元,通过对现行路线规范中缓和曲线长度取值规定的合理性分析,发现回旋线参数取值规定易存在曲线单元间长度分配均匀、取值长度过长、不易与地形地物相协调等问题。结合美国《绿皮书》对缓和曲线取值的相关规定,从汽车行驶轨迹、缓和曲线偏移值特征及超高过渡等方面对缓和曲线长度取值进行理论分析研究。结果表明:在设计速度一定的情况下,随着圆曲线半径的增加,缓和曲线长度取值应随之减小;随着设计速度的增加,缓和曲线长度取值也应随之增加,但增加值较小,可忽略不计;在设计速度和圆曲线半径均相同的情况下,随着缓和曲线长度的增加,驾驶员判断前方平曲线的曲率半径偏小,易驶入临近车道,不利于交通安全。另外,根据缓和曲线偏移值特征及超高过渡所需长度研究发现,通过控制缓和曲线偏移值1 m、缓和曲线长度200 m范围,圆曲线半径与缓和曲线之间的变化关系较为稳定,符合车辆行驶轨迹,基本满足超高过渡段设置的要求。经计算得出当圆曲线半径大于1666 m时,缓和曲线长度宜随圆曲半径的增大而减短。

四轮独立电驱动越野车驱动力优化控制

为了充分利用四轮独立电驱动越野车各轮转矩独立可控的优势,提高越野车的牵引力和越野能力,提出了一种驱动转矩协调控制策略.根据越野车前后轴载荷预分配驱动电动机转矩,利用基于车轮滑转率-路面附着系数的极值寻求算法实现路面最优滑转率估计.采用比例-积分-微分(proportional-integral-differential, PID)控制-滑模控制(sliding mode control, SMC)和基于状态的驱动力再分配方法实现各轮驱动转矩的协调分配,抑制越野车在恶劣路面条件下的车轮打滑.通过CarSim/MATLAB联合仿真以及硬件在环(hardware-in-the-loop, HIL)测试,进行了爬陡坡、单一附着路面、对开路面及连续起伏路面工况的试验验证.结果表明:提出的控制策略能根据实际工况分配驱动轮的转矩,降低驱动轮滑转率,实现整车驱动力的优化.