基于人-车交互的行人轨迹预测

针对行人轨迹预测具有复杂、拥挤的场景和社会交互问题,基于长短时记忆网络（Long Short-term Memory Network, LSTM）对行人与车辆、行人与其他行人的交互进行建模,提出一种基于人-车交互的行人轨迹预测模型（VP-LSTM）。该模型同时考虑了行人与行人的交互、行人与车辆的交互,更适用于复杂的交通场景。所构建的VP-LSTM包括3个输入,以行人的方向和速度作为历史轨迹序列输入,行人与行人的相对位置作为人-人交互信息输入,行人与车辆的相对位置作为人-车交互信息输入。该方法首先设计扇形人-人交互邻域和圆形人-车交互邻域来准确捕捉对被预测行人有相互作用的行人和车辆;其次建立3种不同的LSTM编码层来编码历史行人轨迹序列、人-人、人-车社交信息;然后定义人-人、人-车交互的防碰撞函数和方向注意力函数作为人-车、人-人社交信息的权重,进一步提高社会信息的精度;再将人-人、人-车交互信息输入到注意力模块中筛选出对行人影响大的社会信息;最后将筛选后的社会信息与行人历史轨迹序列一起输入到LSTM神经网络中进行行人轨迹预测,并在构建的DUT人-车交互数据集上验证提出的网络。研究结果表明：提出的方法能够准确地预测出交通场景中,人-车交互行人未来一段时间内的运动轨迹,有效提高了预测精度,提高了智能驾驶决策的准确性。     

A generic data assimilation framework for vehicle trajectory reconstruction on signalized urban arterials using particle filters

With trajectory data, a complete microscopic and macroscopic picture of traffic flow operations can be obtained. However, trajectory data are difficult to observe over large spatiotemporal regions—particularly in urban contexts—due to practical, technical and financial constraints. The next best thing is to estimate plausible trajectories from whatever data are available. This paper presents a generic data assimilation framework to reconstruct such plausible trajectories on signalized urban arterials using microscopic traffic flow models and data from loops (individual vehicle passages and thus vehicle counts); traffic control data; and (sparse) travel time measurements from whatever source available. The key problem we address is that loops suffer from miss- and over-counts, which result in unbounded errors in vehicle accumulations, rendering trajectory reconstruction highly problematic. Our framework solves this problem in two ways. First, we correct the systematic error in vehicle accumulation by fusing the counts with sparsely available travel times. Second, the proposed framework uses particle filtering and an innovative hierarchical resampling scheme, which effectively integrates over the remaining error distribution, resulting in plausible trajectories. The proposed data assimilation framework is tested and validated using simulated data. Experiments and an extensive sensitivity analysis show that the proposed method is robust to errors both in the model and in the measurements, and provides good estimations for vehicle accumulation and vehicle trajectories with moderate sensor quality. The framework does not impose restrictions on the type of microscopic models used and can be naturally extended to include and estimate additional trajectory attributes such as destination and path, given data are available for assimilation.     

Estimating traffic volumes for signalized intersections using connected vehicle data

Recently connected vehicle (CV) technology has received significant attention thanks to active pilot deployments supported by the US Department of Transportation (USDOT). At signalized intersections, CVs may serve as mobile sensors, providing opportunities of reducing dependencies on conventional vehicle detectors for signal operation. However, most of the existing studies mainly focus on scenarios that penetration rates of CVs reach certain level, e.g., 25%, which may not be feasible in the near future. How to utilize data from a small number of CVs to improve traffic signal operation remains an open question. In this work, we develop an approach to estimate traffic volume, a key input to many signal optimization algorithms, using GPS trajectory data from CV or navigation devices under low market penetration rates. To estimate traffic volumes, we model vehicle arrivals at signalized intersections as a time-dependent Poisson process, which can account for signal coordination. The estimation problem is formulated as a maximum likelihood problem given multiple observed trajectories from CVs approaching to the intersection. An expectation maximization (EM) procedure is derived to solve the estimation problem. Two case studies were conducted to validate our estimation algorithm. One uses the CV data from the Safety Pilot Model Deployment (SPMD) project, in which around 2800 CVs were deployed in the City of Ann Arbor, MI. The other uses vehicle trajectory data from users of a commercial navigation service in China. Mean absolute percentage error (MAPE) of the estimation is found to be 9–12%, based on benchmark data manually collected and data from loop detectors. Considering the existing scale of CV deployments, the proposed approach could be of significant help to traffic management agencies for evaluating and operating traffic signals, paving the way of using CVs for detector-free signal operation in the future.     

智能网联车环境下高速匝道汇入车流轨迹优化模型

为了解决传统匝道控制车流汇入时车辆需要减速至停止，从而造成延误时间过长的问题，提出了一种智能网联车环境下的高速匝道汇入车辆轨迹优化的两阶段优化模型，其中，第1 阶段优化车辆进入匝道口的时序；第2 阶段基于第1 阶段的最优时序，优化车辆轨迹. 根据所构建的模型设计了一种启发式算法优化车辆通过匝道冲突区域的时序，然后结合 GPOPS工具优化车辆的轨迹.为了验证所提出方法的有效性，将所提出的方法应用到20 min 随机到达的车流，进行仿真实验.实验结果表明，与先进先出的方法相比，本文所提出的方法能够使总延误减少59.7%，总油耗减少10.5%，说明该方法能够实现车辆以较高的速度通过匝道冲突区域，有效地减少了车辆汇入延误，同时也节约了油耗.     

Predicting market potential and environmental benefits of deploying electric taxis in Nanjing,China

This paper investigates the market potential and environmental benefits of replacing internal combustion engine (ICE) vehicles with battery electric vehicles (BEVs) in the taxi fleet in Nanjing, China. Vehicle trajectory data collected by onboard global positioning system (GPS) units are used to study the travel patterns of taxis. The impacts of charger power, charging infrastructure coverage, and taxi apps on the feasibility of electric taxis are quantified, considering taxi drivers’ recharging behavior and operating activities. It is found that (1) depending on the charger power and coverage, 19% (with AC Level 2 chargers and 20% charger network coverage) to 56% (with DC chargers and 100% charger network coverage) of the ICE vehicles can be replaced by electric taxis without driving pattern changes; (2) by using taxi apps to find nearby passengers and charging stations, drivers could utilize the empty cruising time to charge the battery, which may increase the acceptance of BEVs by up to 82.6% compared to the scenario without taxi apps; and (3) tailpipe emissions in urban areas could be significantly reduced with taxi electrification: a mixed taxi fleet with 46% compressed-natural-gas-powered (CNG) and 54% electricity-powered vehicles can reduce the tailpipe emissions by 48% in comparison with the fleet of 100% CNG taxis.     

基于移动导航数据的信号配时反推

在交通管理和评价时,信号配时对监测评价路口运行状态,评价路口配时方案至关重要.但是,大范围的实时信号配时方案的获取尚缺乏简明有效的途径.本文提出两种基于移动导航数据计算固定配时路口信号配时的方法.第一种方法是在不考虑驾驶员驾驶行为差异性时,得到路口红灯和车均延误的关系模型,从而计算某相位的红灯时长.另外一种方法是基于车辆通过停止线的时间,结合本文提出的上升梯度法,得到某阶段红灯时长.本文通过实际的路口案例计算,将预测结果和已知路口的信号配时比较,表明此方法计算得到的红绿灯时长准确度较高,为后续进行路口运行状态和通行能力研究提供了数据支持.     

Studying the benefits of carpooling in an urban area using automatic vehicle identification data

Carpooling has been considered a solution for alleviating traffic congestion and reducing air pollution in cities. However, the quantification of the benefits of large-scale carpooling in urban areas remains a challenge due to insufficient travel trajectory data. In this study, a trajectory reconstruction method is proposed to capture vehicle trajectories based on citywide license plate recognition (LPR) data. Then, the prospects of large-scale carpooling in an urban area under two scenarios, namely, all vehicle travel demands under real-time carpooling condition and commuter vehicle travel demands under long-term carpooling condition, are evaluated by solving an integer programming model based on an updated longest common subsequence (LCS) algorithm. A maximum weight non-bipartite matching algorithm is introduced to find the optimal solution for the proposed model. Finally, road network trip volume reduction and travel speed improvement are estimated to measure the traffic benefits attributed to carpooling. This study is applied to a dataset that contains millions of LPR data recorded in Langfang, China for 1 week. Results demonstrate that under the real-time carpooling condition, the total trip volumes for different carpooling comfort levels decrease by 32–49%, and the peak-hour travel speeds on most road segments increase by 5–40%. The long-term carpooling relationship among commuter vehicles can reduce commuter trips by an average of 30% and 24% in the morning and evening peak hours, respectively, during workdays. This study shows the application potential and promotes the development of this vehicle travel mode.     

Insights into vehicle trajectories at the handling limits: analysing open data from race car drivers

Race car drivers can offer insights into vehicle control during extreme manoeuvres; however, little data from race teams is publicly available for analysis. The Revs Program at Stanford has built a collection of vehicle dynamics data acquired from vintage race cars during live racing events with the intent of making this database publicly available for future analysis. This paper discusses the data acquisition, post-processing, and storage methods used to generate the database. An analysis of available data quantifies the repeatability of professional race car driver performance by examining the statistical dispersion of their driven paths. Certain map features, such as sections with high path curvature, consistently corresponded to local minima in path dispersion, quantifying the qualitative concept that drivers anchor their racing lines at specific locations around the track. A case study explores how two professional drivers employ distinct driving styles to achieve similar lap times, supporting the idea that driving at the limits allows a family of solutions in terms of paths and speed that can be adapted based on specific spatial, temporal, or other constraints and objectives.     

城市公交GPS轨迹路线图制作研究

以哈尔滨市104路公交为主要研究对象,详细介绍城市公交的GPS轨迹路线图制作时所需的软硬件,以及数据的采集方法,叙述数据的整合与分类过程。通过将主要公交站点周围兴趣点信息增添到轨迹点附近,构成包含有站点轨迹信息、站点影像信息的沿途轨迹信息。经过整合与分类,通过HOLUX ezTour软件对数据进行处理,阐述轨迹查询、地标的新增与删改和轨迹的添加与变换等功能。并使用Google Earth三维图像对轨迹路线全面整合,最终制作成城市公交GPS轨迹路线图。     

����ʱ�չ켣���ٵĽ�ͨ�����Բ���������

基于视频的交通流检测在智能交通系统中具有重要意义。本文针对广泛采用的低位摄像机,提出了一种交通流特性参数的检测分析方法。首先基于三级虚拟检测线和自适应更新率局部背景建模来快速提取车辆特征点并消除活动阴影对提取精度的影响;然后基于Adaboost（Adaptive Boosting,自适应增强）分类器实现特征点按车分组,并在跟踪过程中根据运动特征相关度消除分组误差,获取高精度的车辆轨迹;进而自动生成多车道轨迹时空图并提取各车道交通流的多种特性参数。实验结果验证了算法的高效性;同时,自动生成的多车道轨迹时空图也为更多的交通信息获取和更深入的交通流特性分析提供了有力支持。