驾驶员接管自动驾驶车辆研究进展

为全面认识自动驾驶接管行为特征,分析驾驶员接管行为,结合自动驾驶接管的技术和现实背景,从自动驾驶接管绩效的影响因素,包括场景、技术、心理和生理等因素,梳理国内外驾驶员接管自动驾驶车辆的研究成果,总结主要研究内容和方法,并展望其未来研究趋势.通过归纳和分析揭示了影响驾驶员接管自动车辆的因素,主要包括接管场景和接管请求方式,同时非驾驶相关任务和年龄等因素也会影响驾驶员的接管行为和表现.针对驾驶员接管的驾驶特征及行为研究,内容精确丰富,方法科学完善;而对驾驶员接管绩效评价和干预研究,重点关注在统计学的基础上,建立完善的数据指标评价体系,用于评价和干预驾驶员接管绩效.未来针对自动驾驶接管的研究,一方面寻求自动车辆技术突破;另一方面综合心理学和统计学理论基础,建立驾驶员接管能力培训体系.      

驾驶模拟器在自动驾驶系统中的应用研究

自动驾驶车辆在实际道路上行驶之前的测试阶段是一个至关重要的环节。一个低成本、高效率以及高精度测量的自动驾驶车辆的测试方式,对于自动驾驶车辆的开发具有重要意义。将驾驶模拟器运用到研究自动驾驶车辆测试已是近年来的一个研究热点。基于虚拟驾驶场景的自动驾驶车辆的检测,通过组合虚拟驾驶场景的背景车辆、行人、交通灯、建筑、指示标牌等元素,研究将驾驶模拟器与虚拟驾驶场景的联合应用来测试自动驾驶车辆。设计了典型的交通场景,通过自动驾驶车辆和背景车辆的实时交互,研究自动驾驶车辆的各项性能指标。研究结果表明:该驾驶模拟器可以高度拟合人类驾驶体验,驾驶员通过驾驶模拟器控制背景车辆能够很好的模拟现实中的驾驶行为,对自动驾驶车辆的仿真测试起到了促进作用。     

城市道路环境下自动驾驶车辆接管绩效分析（双语出版）

为了分析城市道路环境下高度自动驾驶中非驾驶相关任务和接管紧迫度对接管绩效的影响,基于驾驶模拟器设计了自动驾驶紧急接管场景并开展驾驶模拟试验,接管请求时间分别设定为3,4,5 s,非驾驶相关任务为读新闻、看视频、玩游戏,自动驾驶车速为50 km·h^-1,试验中共招募了49名被试（男性30名,女性19名）,被试的平均年龄为31.06岁（标准差为7.1岁）,驾驶人在自动驾驶阶段始终执行非驾驶相关任务,听到接管请求提示后需要接管车辆的控制权,并实施紧急避让操作。研究结果表明：在紧急接管情况下,接管紧迫度对合成加速度和最小TTC有影响,而对接管时间无影响,与5 s的接管请求时间条件相比,3,4 s的接管请求时间条件下的合成加速度明显增加,而最小TTC则随接管请求时间的减少而降低;非驾驶相关任务对接管时间和最小TTC有影响,而对合成加速度无影响,与无非驾驶相关任务相比,非驾驶相关任务会显著增加接管时间和降低最小TTC;碰撞几乎都发生在3 s和4 s的接管请求时间下,5 s的接管请求时间能够基本保证接管的安全性。     

基于整车在环仿真的自动驾驶汽车室内快速测试平台

整车在环仿真测试方法可以安全、高效地验证复杂环境和极端工况等场景下自动驾驶汽车性能的有效性,基于此研发一种基于整车在环仿真的自动驾驶汽车室内快速测试平台,该平台由前轴可旋转式转鼓试验台、试验台测控子系统、虚拟场景自动生成子系统、虚拟传感器模拟子系统、驾驶模拟器、自动驾驶汽车和测试结果自动分析评价子系统组成。通过在试验台滚筒上独立加载转矩模拟车辆行驶阻力,可动态模拟不同的路面附着系数,同时利用坡度、侧倾和转向随动机构可模拟车辆俯仰角、侧倾角和航向角3个自由度;采用虚拟现实技术柔性集成车辆动力学模型、传感器仿真、复杂道路交通环境及测试用例仿真,模拟多种道路交通场景,并通过传感器仿真及数据融合等技术快速测试自动驾驶汽车智能感知与行为决策等性能指标。将自动驾驶汽车、虚拟仿真场景和试验台耦合构建一个闭环系统,完成了多项关键技术研发,包括：多自由度高动态试验台结构设计、虚拟测试场景自动重构方法和传感器数据模拟及注入方法,可满足在各种场景下测试自动驾驶汽车整车性能的需求。此外,为验证快速测试平台的有效性,以U-turn轨迹跟踪控制为研究实例,基于简化的车辆运动学模型和模型预测控制算法,在平台上搭建U-turn场景并对自动驾驶汽车的轨迹跟踪控制算法性能进行大量测试。结果表明：自动驾驶汽车室内快速测试平台可以真实地模拟汽车在道路上的运行工况,自动驾驶汽车在虚拟场景中的轨迹跟踪效果良好,与参考轨迹的偏差小于8%,证明了该测试平台检测方法的有效性。     

自动驾驶汽车在城市道路上面临的主要场景挑战

基于人类的自然驾驶数据、交通事故数据等信息提出的自动驾驶汽车技术要求和测试场景难以反映自动驾驶汽车特有的局限性和安全风险。文章从车辆视角出发,利用搭载高度自动驾驶功能的汽车开展了实际道路测试,并通过分析测试过程中遇到的危险场景和边缘场景,总结了城市道路环境下自动驾驶汽车在感知、定位、决策规划、控制执行和网联通信等方面存在的主要场景挑战。该研究成果弥补了基于人类驾驶数据开展相关研究的不足,能够为自动驾驶汽车的研发和基于场景的测试验证提供参考。     

驾驶次任务沉浸等级对接管行为的影响分析

为探究自动驾驶中驾驶次任务沉浸等级对接管行为的影响,基于驾驶模拟器搭建自动驾驶接管行为测试平台,设计事故接管场景,基于驾驶次任务(娱乐任务和工作任务)和接管请求时间(5 s和10 s)因素组合开发4个事故接管情景,招募被试参与驾驶模拟实验并采集驾驶人的接管行为数据,选择速度、横向偏移、接管反应时间和接管正确时间4个指标...     

不同情景下自动驾驶接管行为的影响特征

为探究不同情景下自动驾驶接管行为的影响特征,面向驾驶人、自动驾驶车辆、交通环境等内容提出自动驾驶测试研究框架。基于驾驶模拟技术开发自动驾驶测试平台,通过案例验证其有效性,为自动驾驶相关技术的测试评估提供有力支撑。研究以接管场景、接管请求时间、驾驶次任务、交通流为要素设计18个高速公路接管情景,邀请被试开展驾驶模拟试验测试。从主观维度探究驾驶人对自动驾驶的适应性差异,从客观维度构建广义线性混合效应模型,研究驾驶人属性因素(性别、年龄、驾龄)和接管情景因素(接管场景、接管请求时间、驾驶次任务)的主效应及其交互作用对接管行为的影响。统计分析结果表明：①性别因素对自动驾驶的信任度和状态感知度有统计学差异,男性对自动驾驶的适应性高于女性;②驾驶人的年龄和驾龄因素对试验前和试验后的技术接受度具有显著影响,对技术信任度和状态感知度具有统计学差异,中年人和老年人、中驾龄和高驾龄人群的适应性相对较高;③不同因素水平对应的接管成功率、正确率和第一操纵行为不同。广义线性混合效应模型结果表明：①接管情景因素及其交互作用对接管行为指标具有显著影响;②模型中引入驾驶人属性因素,发现与接管情景因素存在交互效用。研究基于驾驶模拟技术开发自动驾驶测试平台的方法具有一定的推广性,研究结果可为深度挖掘自动驾驶接管行为影响因素及其作用机理奠定基础。     

世界汽车组织自动驾驶汽车"三支柱"认证方法提议及对我们的启示

自动驾驶技术发展背景下,软件在定义汽车,车辆功能日趋复杂、软件数量也越来越多.与传统车辆相比,影响自动驾驶汽车的安全因素、场景变量日益增长,通过传统的场地测试和台架测试,难以全部评估上述影响安全的变量.因此,为了车辆上道路运行,需要提出新的认证思路,支持车辆型式认证或自我认证.世界汽车组织在分析自动驾驶汽车技术特性和安...     

预期接管场景下接管绩效及接管风险研究

为了解条件自动驾驶预期接管场景下,预提醒(有预提醒和无预提醒)、速度差(30 km·h-1速度差和无速度差)和接管预留时距(5 s预留时距和7 s预留时距)对驾驶人接管绩效的影响,基于驾驶模拟器,设计了高速公路出口前的预期接管场景,进行接管过程的模拟驾驶试验.试验共招募39名被试(26名男性,13名女性),被试平均年龄...     

基于人工智能算法的换道超车功能开发

随着汽车电气化、智能化的不断发展,汽车行驶的场景越来越趋于多样化和复杂化,从而促使汽车从辅助驾驶向智能驾驶不断创新。随着人工智能的引入,汽车智能驾驶功能越来越趋于实用,正在逐步实现向解放驾驶员双手、向车载高级驾驶辅助系统代替人脑进行复杂驾驶场景实时响应的阶段发展;高阶复杂场景智能驾驶功能则在辅助驾驶功能实现的基础上,针对驾驶员实际驾驶感受并结合人工智能算法实现向车辆复杂场景下的自动驾驶操作的方向发展。介绍了基于人工智能算法的换道超车功能开发,即通过换道条件的智能选择,使车辆以最佳方式自动完成换道超车过程。     

高速公路行车安全驾驶技术控制探讨

市场经济迅速发展以及人们生活水平的提高,汽车已经成为了人们日常出行必不可少的交通工具之一.随着汽车保有量的持续增加,社会各界对高速公路行车安全的关注度也随之越来越高.所以,加强高速公路行车安全驾驶技术分析研究的力度,制定切实可行的安全驾驶技术控制措施,才能从根本上促进高速公路行车安全水平的有效提升.文章主要是就高速公路...     

Vehicle sensor data-based analysis on the driving style differences between operating indoor simulator and on-road instrumented vehicle

Abstract

Indoor simulator and on-road instrumented vehicle are the most popular ways to analyze driving behaviors by using collected Vehicle Sensor Data (VSD). However, for a same driver, the driving performance could be different in the real world and in the simulated world. Even though many studies have been conducted to discover the differences of driving behaviors in these two circumstances, little research has focused on analyzing the differences in driving style, which can provide more integrated knowledge of a driver from the natural structure, stimulus–response mechanism, of driving behaviors. Therefore, in this paper, the driving styles in both the real world and the simulated world are extracted by implementing the nonnegative matrix factorization method on the collected VSD data. Through this analysis, the driving style differences can be quantitatively described and discussed in detail. It is found that the drivers tend to be more unstable and sometimes aggressive when driving the simulator and the deviation in the perception of temporal gap in two circumstances is also discovered. The research findings are particularly valuable to calibrate the driving simulator and construct more reliable driving behavior models.     

Profile-speed data-based models to estimate operating speeds for urban residential streets with a 30 km/h speed limit

A speed limit of 30 kilometres per hour (km/h) has been widely introduced for urban residential streets to ensure traffic safety and allow these streets to fulfil other intended functions. However, excessive speeds on these roads are very common, causing traffic safety problems and threatening the liveability of neighbourhoods. An effective and active way to deal with speeding is the application of a performance-based design approach, as mentioned in previous research. In a performance-based design approach, street geometrics and roadside elements are selected based on their influence on the desired driving speeds. The relationship between driving speeds and street features therefore needs to be determined. Although several studies have developed operating speed models for urban streets, all of these models were calibrated based on data for streets with speed limits of more than 30 km/h. The present research is designed to investigate the influence of various roadway and roadside characteristics on operating speeds on urban tangent street sections with a 30 km/h speed limit using profile-speed data. A simultaneous equation regression with a three-stage-least-square (3SLS) estimator was used for the modelling effort. The driving speed models developed in this study incorporate several street design factors, which provide helpful information for urban planners and street designers to cope with speeding issues on residential streets.     

T��l��phoner au volant : impact sur la conduite et risque d��accident, une revue de la litt��rature

This paper gives a literature review from experimental, epidemiological and naturalistic studies, having tried to measure on the one hand, the impact of mobile phone use on driving and, on the other hand, the risk of accident associated with this practice. The overview of these studies allowed first of all to show that the driving task is altered during the period of mobile phone use, and to quantify the disturbance (experimental studies), then to verify that this disturbance can be at the origin of accidents and quantify this risk (epidemiological studies). The naturalistic studies, consisting in following-up fleet vehicles, give some more explanations and tend to confirm the results. Advantages and limits of each approach are also discussed.     

Assessing the effect of inattention-related error and anger in driving on road accidents among Iranian heavy vehicle drivers

Extensive studies have examined the effect of variables such as demographic characteristics, insomnia, and working conditions of drivers individually on inattention-related error as well as expression of anger in driving. Nevertheless, so far no study has tested the concurrent effect of these factors on crashes. This study has dealt with indirect investigation of the effect of variables including demographic characteristics, insomnia, and working conditions of drivers on inattention-related error and expression of anger in driving (as mediation model). Next, the effect of these two variables on the probability of incidence of road crashes has been assessed among truck drivers. For this purpose, 780 Iranian truck drivers were interviewed by validated questionnaires including insomnia severity index (ISI), attention related driving errors scale (ARDES), and driving anger expression (DAX). To confirm the validity of these questionnaires, confirmatory factor analysis (CFA) was used. Next, structural equation modeling (SEM) was employed for investigating the relationships between these latent variables and truck drivers' demographic characteristics, working conditions and their crash involvement. SEM results indicated that as the severity of insomnia among drivers increased, they committed more inattention-related errors, and also expressed more anger during driving. Generally, the results of this study indicated that with increase in the extent of experience and safety knowledge of drivers and improvement in their working conditions and their sleep status, it is possible to reduce inattention-related error and expression of anger during driving among heavy vehicle drivers.     

Driving segment simulation for determination of the most effective driving features for HEV intelligent control

This paper presents a methodological approach for determination of the most effective driving features for hybrid electric vehicle intelligent control, using the driving segment simulation. In this approach, driving data gathering is first performed in real traffic conditions using Advanced Vehicle Locator systems. The vehicle's speed time series are then divided into small segments. Subsequently, 19 driving features are defined for each driving segment, and the influence of the driving features on the vehicle's fuel consumption (FC) and exhaust emissions is investigated, using driving the driving segment simulation. The simulation approach is also verified by experimental test. Finally, the driving features are ranked by a new approach based on the definition of an effectiveness index and a correlation analysis. The results demonstrate that the velocity-dependent driving features such as ‘energy’, ‘mean of velocity’, ‘displacement’ and ‘maximum velocity’ are more effective on vehicle's FC and exhaust emissions. However, because of high dependency between these features, this study suggests independent driving features among the most effective driving features.     

Design of a headway distance control system for ACC

A headway distance control system for adaptive cruise control (ACC) must provide the driver's desired response at all times while maintaining stability. Therefore, a two- degree-of-freedom model was adopted for the headway distance control unit to allow response and stability to be designed as independent parameters. The desired response was decided based on the results of various driving tests and driving simulations. Using these results as the reference model has achieved the smooth response and stability required for ACC.     

Car drivers' road safety performance: A benchmark across 32 countries

The road safety performance of a country and the success of policy measures can be measured and monitored in different ways. In addition to the traditional road safety indicators based on the number of fatalities or injured people in road traffic crashes, complementary road safety performance indicators can be used in relation to vehicles, infrastructure, or road users' behaviour. The last-mentioned can be based on data from roadside surveys or from questionnaire surveys. However, results of such surveys are seldom comparable across countries due to differences in aims, scope, or methodology.This paper is based on the second edition of the E-Survey of Road Users' Attitudes (ESRA), an online survey carried out in 2018, and includes data from more than 35,000 road users across 32 countries. The objective is to present the main results of the ESRA survey regarding the four most important risky driving behaviours in traffic: driving under the influence (alcohol/drugs), speeding, mobile phone use while driving, and fatigued driving. The paper explores several aspects related to these behaviours as car driver, such as the self-declared behaviours, acceptability and risk perception, support for policy measures, and opinions on traffic rules and penalties.Results show that despite the high perception of risk and low acceptability of all the risky driving behaviours analysed, there is still a high percentage of car drivers who engage in risky behaviours in traffic in all the regions analysed. Speeding and the use of a mobile phone while driving were the most frequent self-declared behaviours. On the other hand, driving under the influence of alcohol or drugs was the least declared behaviour. Most respondents support policy measures to restrict risky behaviour in traffic and believe that traffic rules are not being checked regularly enough, and should be stricter.The ESRA survey proved to be a valuable source of information to understand the causes underlying road traffic crashes. It offers a unique database and provides policy makers and researchers with valuable insights into public perception of road safety.     

A similarity-based neuro-fuzzy modeling for driving behavior recognition applying fusion of smartphone sensors

Drivers’ behavior evaluation is one of the most important problems in intelligent transportation systems and driver assistant systems. It has a great influence on driving safety and fuel consumption. One of the challenges in this regard is the modeling perspective to treat with uncertainty in judgments about driving behaviors. Really, assessing a single maneuver with a rigid threshold leads to a weak judgment for driving evaluation. To fill this gap, a novel neuro-fuzzy system is proposed to classify the driving behaviors based on their similarities to fuzzy patterns when all of the various maneuvers are stated with some fuzzy numbers. These patterns are also fuzzy numbers and they are extracted from statistical analysis on the smartphone sensors data. Our driving evaluation system consists of three processes. Firstly, it detects the type of all of the maneuvers through the driving period, by using a multi-layer perceptron neural network. Secondly, it extracts a new feature based on the acceleration and assigns three fuzzy numbers to driver’s lane change, turn and U-turn maneuvers. Thirdly, it determines the similarity between these three fuzzy numbers and the fuzzy patterns to evaluate the safe and the aggressive driving scores. To validate this model, Driver’s Angry Score (DAS) questionnaires are used. Results show that the fusion of Inertial Measurement Unit (IMU) sensors of smartphones is enough for the proposed driving evaluation system. Accuracy of this system is 87% without using GPS and GIS data and this system is independent of smartphones and vehicles types.     

信号交叉口行车风险场建立及车辆通行行为优化

随着汽车逐步向智能化、网联化发展,智能网联车辆逐步进入实际应用阶段。进行智能网联车辆的通行行为优化,对提升驾驶安全性和行车效率,避免事故发生和交通拥堵至关重要。车辆在通过交叉口时将受到很多环境及运动因素的影响,而现有的通行优化模型难以准确表达各类因素共同作用下的行驶环境。为此,基于风险场理论建立由环境场和运动场组成的信号交叉口行车风险场,表征信号交叉口中每点的实时行车风险程度,从而引导车辆驶向风险值低点,并提供下一步长的位移及速度指引,实现车辆的动态轨迹优化及速度控制。典型场景下的仿真结果表明：在优化模型的控制下单车的信号交叉口通行效率明显提升,其中直行方向车辆单车平均通行效率提升最高,平均提升6.35%,通过对交叉口面积内所有车辆进行通行行为优化,交叉口通行效率提升了9.3%,这表明所建模型可以准确表达交叉口行车环境并优化车辆通行行为。研究结论可应用于自动驾驶车辆的交叉口通行控制,并为网联环境下的行车环境表达和安全驾驶控制提供模型基础。