考虑个体行为的改进CA模型人员疏散研究

为更好模拟行人疏散过程中微观个体行为,考虑行人身材半径及在疏散过程中行人步行速度随运动状态变化,将社会力模型运行规则引入元胞自动机模型,建立了一种社会力模型计算步行速度、空间离散化程度和步行速度较高的疏散模型,用于模拟紧急情况下的行人疏散过程。在该模型中空间划分为更小网格,每个行人占用一到多个单元格,行人的身材半径不再不变,每个行人移动的距离由其速度决定,根据基于速度的出口选择方法和行人运动规律,通过数值模拟分析,研究了疏散过程中的动态性。研究表明基于速度的网格移动数量、行人数量、期望速度、行人身材半径、松弛时间等参数影响疏散效率,结合连续模型的优点能够更加客观真实刻画疏散过程,有助于离散模型描述行人疏散微观行为特征。     

基于云网格集成调度的防拥堵车辆路径规划算法

在道路交通路网中,车辆拥堵问题是流量与路网结构之间相互作用的一个复杂动态过程,通过车辆路径规划,实现对路网网格集成调度,从而提高路网通行吞吐量。传统方法采用并行微观交通动态负载平衡预测算法实现车辆拥堵调度和车辆路径规划,不能准确判断路面上的车辆密度,路径规划效益不好。提出一种基于云网格集成调度的防拥堵车辆路径规划算法,即构建基于Small-World模型的云网格路网模型,采用RFID标签信息进行路况信息采集,实现交通网络拥堵评估信息特征的提取,采用固有模态函数加权平均求得各车道的车辆拥塞状态函数,对所有车道内车辆密度取统计平均可获得簇内的车辆密度。设计交通路网拥堵检测算法来对当前个体道路信息进行一维邻域搜索,从而实现车辆路径规划控制目标函数最佳寻优。通过动态博弈的方式求得车辆防拥堵路径的近似最优轨迹,实现路径规划算法的改进。仿真结果表明,该算法能准确规划车辆路径,实现最优路径控制,从而提高严重拥堵路段的车流速度和路网吞吐性能,性能优越。     

基于多智能体的信号交叉口微观仿真

交通系统是一个典型的多智能体系统（Multi-Agent System,MAS）。应用多智能体技术建立了信号交叉口微观仿真模型,给出了机动车Agent的三层结构,详细分析了机动车在信号交叉口中的基于期望间隙和期望速度的运动决策逻辑,探讨了机动车加速度模型。针对国内交通仿真软件缺少行人与非机动车模拟的不足,以元胞自动机模型为基础引入了行人及非机动车实体,使得仿真系统更具实用性。给出了该仿真系统的实现结果和结论。     

基于VISSIM微观交通仿真软件的导流岛机非冲突元胞自动机模型

针对复杂的导流岛冲突区机非冲突行为,应用传统解析法得到的右转车道通行能力往往与实际运行状况偏差较大。为此,提出了基于VISSIM微观交通仿真软件的导流岛机非冲突元胞自动机模型。该模型利用VISSIM的组件对象模型编程,依据提出的元胞自动机规则集,通过设置一系列模拟元胞的检测器,来控制右转机动车的车速变化,从而模拟右转机动车面对非机动车或行人冲突时的截流效应,并同时利用VISSIM仿真软件的让行设置来控制非机动车或行人的过街行为。仿真结果表明,利用该模型得到的右转车道通行能力值与实际观测值的平均相对误差为5.45%,优于传统的解析法,能够较好地反映导流岛冲突区的实际运行状况,从而为混合交通条件下导流岛渠化形式的规划、设计、交通管理与组织提供理论依据。     

Agent-CA的体育场馆人群疏散模型

根据体育场馆人群疏散的特点与规律,提出一种基于多智能体和元胞自动机相融合的大型体育场馆人群疏散模型（Agent-CA）。将元胞空间中被虚拟人个体占据的元胞视为一个独立的智能体,将元胞及其状态进行封装,扩展为具有自主性的智能体,通过设计各种人群疏散行为策略做为演化规则,实现个体的差异性以体现个人个性、体力、心理等对疏散行为的影响,对体育场馆的人群疏散进行仿真实验。结果表明,Agent-CA综合了多智能体和元胞自动机的优点,充分考虑了个体内在因素,更接近现实大型体育场馆的人群疏散情形,缩短了疏散时间。     

基于智能体模型的路径规划与行为模拟仿真研究

智能体模型是当前在模拟仿真领域内能够实现的最为接近人类的方式。是实现复杂行为模拟仿真的重要方法,智能体能够通过逻辑驱动或自学习模式进行决策。多智能体系统在复杂战场环境的模拟仿真方面有着较大的优势,其分布式交互原理能够模拟不同兵种或者不同个体的行为与路径规划决策。文章针对当前智能体模型的研究现状以及路径规划领域的各种方法进行了简要介绍,并指出优点和不足,最后对其发展趋势进行了展望。     

基于模糊逻辑的路口交通信号控制

城市交通控制(UTC)是智能交通系统中非常重要的组成部分。在一些复杂的交通情况下,使用传统的控制方法进行交通信号控制的效果并不好。由于模糊逻辑能够较好地描述复杂系统的定性模型,因此它非常适合对路口交通信号灯进行控制。该文给出了一种采用模糊逻辑进行路口信号灯控制的方法。通过从路口检测器获取车辆信息,模糊规则对信号灯进行优化控制。通过仿真模拟,给出了实验结果。     

人群应急疏散中一种多智能体情绪感染仿真模型

为了从情绪的视角分析紧急情境下人群的疏散行为,梳理了现有情绪感染的研究工作,总结了人群紧急状况下行为特点.采用智能体描述人群个体,提出一种多智能体情绪感染模型.其主体框架分为感知层、情绪层、感染层、行为层和行动层.归纳了产生情绪感染现象的3个条件及情绪感染的3个规则,提出了情绪感染的算法,考虑个体的个性和个体间距离因素,采用情绪强度和人群紧密度来计算个体疏散速度.用C#语言编制了仿真实验,采用真实的地震疏散案例,验证了仿真疏散时间和实际观测的基本一致.通过与以往基于传染病思路的情绪感染模型对比,所提出的模型可以更好地描述情绪感染从局部到整体的过程.实验结果表明,所提出的模型可以推演情绪驱动下的群体聚集行为,有望为制定应急疏散预案提供一种可视化分析方法.     

无信号人行横道交通流特性的研究

考虑到行人穿越人行横道特点和心理因素,构建人行横道处机动车和行人相互干扰行为的元胞自动机模型。模型中重新制定行人和机动车的冲突干扰规则,引入临界安全间隙和临界决策间隙概念计算行人的通过概率,根据行人过街等待时间阈值定义了行人的冒险概率。模拟结果显示,临界决策间隙、行人忍耐时间阈值和行人的临界跟随间隙对机动车与行人流量均有不同程度的影响,其中行人的临界跟随间隙对交通流影响最大。     

TR-light：基于多信号灯强化学习的交通组织方案优化算法

针对多变环境条件下的交通堵塞问题,将强化学习、神经网络、多智能体和交通仿真技术结合起来,提出了用于优化多路口条件下交通状况的trajectory reward light(TR-light)模型。该方法具有几个显著特点：基于红绿灯拟定交通组织方案;将多智能体强化学习用于红绿灯控制;通过红绿灯的协同达到区域级的交通组织优化;在智能体每次行为执行结束后实施轨迹重构,在OD对不改变的情况下改变车辆行驶路径,根据方案和重构轨迹来计算智能体的最终回报。通过SUMO进行交通仿真实验和交通指标对比,验证了该模型在多交叉口中能够提高路网畅通率,改善交通状态。实验表明该模型可行,可有效缓解交通拥堵。     

A modeling framework for urban traffic systems microscopic simulation

In this paper a modeling framework for urban traffic systems (UTS) is presented. The model, used for agent based micro-simulation, describes both the traffic network and dynamic entities, namely vehicles, traffic lights, and pedestrians. The framework allows defining systematically the necessary components and their behavior of a model oriented to event driven simulation, which can be executed in a distributed way. In the model, the vehicles are conceived as mobile agents with decision making capabilities that interact with the environment and other entities within the traffic network, performing diverse activities according to numerous situations arisen during the simulation. A multi-level Petri net based formalism, named n-LNS is used for describing the structure of the UTS and the components behavior. The first level describes the traffic network; the second level models the behavior of diverse road network users considered as agents, and the third level specifies detailed procedures performed by the agents, namely travel plans, tasks, etc.     

基于道路环境上下文的行人跟踪方法

针对目前城市交通中人车混行场景中行人跟踪效果不佳的问题,提出了一种基于道路环境上下文的行人跟踪方法。首先通过对道路环境上下文进行分析,建立道路模型;其次在道路模型的约束下建立行人与环境的交互运动模型;最后利用该模型进行行人的跟踪。在真实场景中的实验表明使用了道路上下文信息的跟踪方法与连续离散连续能量最小化的多行人跟踪方法相比,多目标跟踪准确度从47.6%提升至63.2%,多目标跟踪精度从68.8%提升至74.3%。数值结果表明道路上下文信息对于提高人车混行场景中行人跟踪效果的有效性。     

Right of way

Pedestrian models typically represent interactions between agents in a symmetric fashion. In general, these symmetric relationships are valid for a large number of crowd simulation scenarios. However, there are many cases in which symmetric responses between agents are inappropriate, leading to unrealistic behavior or undesirable simulation artifacts. We present a novel formulation, called right of way, which provides a well-disciplined mechanism for modeling asymmetric relationships between pedestrians. Right of way is a general principle, which can be applied to different types of pedestrian models. We illustrate this by applying right of way to three different pedestrian models (two based on social forces and one based on velocity obstacles) and show its impact in multiple scenarios. Particularly, we show how it enables simulation of the complex relationships exhibited by pilgrims performing the Islamic religious ritual, the Tawaf.     

Modeling the impact of interaction on pedestrian group motion

Mobile social robots aimed at interacting with and assisting humans in pedestrian areas need to understand the dynamics of pedestrian social interaction. In this work, we investigate the effect of interaction on pedestrian group motion by defining three motion models to represent (1) interpersonal-distance, (2) relative orientation and (3) absolute difference of velocities; and model them using a dataset of 12000+ pedestrian trajectories recorded in uncontrolled settings. Our contributions include: (i) Demonstrating that interaction has a prominent effect on the empirical distributions of the proposed joint motion attributes, where increasing levels of interaction lead to more regular behavior (ii) Developing analytic motion models of such distributions and reflect the effect of interaction on model parameters (iii) Detecting the social groups in a crowd with almost perfect accuracy utilizing the proposed models, despite the constant flow direction in the environment which causes unrelated pedestrians to move in a correlated way, and thus makes group recognition more difficult (iv) Estimating the level of intensity with considerable rates utilizing the proposed models     

双目机器视觉及RetinaNet模型的路侧行人感知定位

目的 行人感知是自动驾驶中必不可少的一项内容,是行车安全的保障。传统激光雷达和单目视觉组合的行人感知模式,设备硬件成本高且多源数据匹配易导致误差产生。对此,本文结合双目机器视觉技术与深度学习图像识别技术,实现对公共路权环境下路侧行人的自动感知与精准定位。方法 利用双目道路智能感知系统采集道路前景图像构建4种交通环境下的行人识别模型训练库;采用RetinaNet深度学习模型进行目标行人自动识别;通过半全局块匹配（semi-global block matching,SGBM）算法实现行人道路前景图像对的视差值计算;通过计算得出的视差图分别统计U-V方向的视差值,提出结合行人识别模型和U-V视差的测距算法,实现目标行人的坐标定位。结果 实验统计2.5 km连续测试路段的行人识别结果,对比人工统计结果,本文算法的召回率为96.27%。与YOLOv3（you only look once）和Tiny-YOLOv3方法在4种交通路况下进行比较,平均F值为96.42%,比YOLOv3和Tiny-YOLOv3分别提高0.9%和3.03%;同时,实验利用标定块在室内分别拍摄3 m、4 m和5 m不同距离的20对双目图像,验证测距算法,计算标准偏差皆小于0.01。结论 本文提出的结合RetinaNet目标识别模型与改进U-V视差算法能够实现对道路行人的检测,可以为自动驾驶的安全保障提供技术支持,具有一定的应用价值。     

Modeling pedestrians’ interest in locations: A concept to improve simulations of pedestrian destination choice

Large environments that are designed for travel, leisure, and for everyday life – such as transport hubs, amusement parks, and shopping centers – feature different locations that are frequently visited by pedestrians. Each visit is evoked by one’s motivation to engage in some kind of activity at a certain location. By means of modeling the pedestrians’ interests in locations with the aid of computer simulations, it is possible to forecast the occupancy at locations by utilizing sophisticated pedestrian destination choice models. In the field of pedestrian dynamics research, location preference modeling is not common, but it is all the more rare to include a psychological grounding into such choice models. Here we show that our psychologically inspired and mathematically defined model to describe pedestrians’ interests in locations is able to improve the exactness of pedestrian destination choice models. The interest function model is based on the psychological concept of goal-related memory accessibility and on fundamental coherences found in pedestrian-related data that is measurable at locations. We validated the interest function model and our results provide evidence that our approach improves the simulation fidelity regarding occupancy forecasting. Because the interest concept is designed as a framework that can be coupled to existing microscopic pedestrian simulators, it can be used in most pedestrian destination choice models to describe pedestrian visiting preferences. Consequently, the reliability of the occupancy predictions of pedestrian simulations can be enhanced by integrating the interest function model into choices models.     

基于视野域机制的行人轨迹预测

为提高行人在复杂交通场景中交互的安全性,提出一种基于social-GAN（social-generative adversarial network）的行人轨迹预测算法SAN-GAN（social angle norm-GAN）。该算法首先以行人历史位置信息与头部信息为输入,通过轨迹生成器LSTM网络（long short term memory networks）获取行人隐藏特征信息,并基于行人视野域模块捕捉行人视野域动态变化,对所有行人建立扇形视野域并筛选有效信息,从而驱动神经网络模型预测行人未来轨迹变化。将SAN-GAN与LSTM、social-LSTM（social-long short term memory networks）、social-GAN等轨迹预测算法进行对比实验,结果表明SAN-GAN算法相较于其他算法,在预测3.2 s的行人轨迹时,ADE分别平均降低65.8%、51.2%、10.7%,FDE分别平均降低73.6%、60.9%、10.4%。SAN-GAN能够有效地预测行人在复杂交通环境中进行交互的未来轨迹。     

Modeling pedestrian mobility in disaster areas

Realistic mobility modeling is necessary for testing disaster management strategies as well as performance of disaster–resilient networks. Evacuation of the people from a disaster area depends on the environment and type of the hazard which cause certain changes in the pedestrian flows. Although most models focus on the building evacuations or city-scale evacuation planning, there is a need for a mobility model that captures the pedestrians’ movement behavior during evacuation from large and crowded disaster areas such as theme parks.In this paper, we propose a mobility model of the pedestrians in disaster areas. In our application scenario of theme parks, the main mission of the operators is the evacuation of the visitors and providing access to transportation vehicles such as ambulances. We use real maps to generate theme park models with obstacles, roads, and disaster events. We incorporate macro and micro mobility decisions of the visitors, considering their local knowledge and the social interactions among the visitors. We analyze the outcomes of the simulation of our theme park disaster (TP-D) mobility model with simulations of currently used models and real-world GPS traces. Moreover, using the proposed model as a baseline, we analyze the performance of an opportunistic network application.     

自动驾驶场景的尺度感知实时行人检测

目的 行人检测是目标检测中的一个基准问题,在自动驾驶等场景有着较大的实用价值,在路径规划和智能避障方面发挥着重要作用。受限于现实的算法功耗和运行效率,在自动驾驶场景下行人检测存在检测速度不佳、遮挡行人检测精度不足和小尺度行人漏检率高等问题,在保证实时性的前提下设计一种适合行人检测的算法,是一项挑战性的工作。方法 本文旨在解决自动驾驶场景中耗时长、行人遮挡和小尺度行人检测结果精度低的问题,提出了一种尺度注意力并行检测算法（scale-aware and efficient object detection,Scale-aware EfficientDet）：在特征提取与检测中使用了EfficientDet的主干网络,保证算法效率和功耗的平衡;在行人遮挡方面,为了提高模型对遮挡现象的检测精度,引入了可以增强行人与其他物体之间特征差异的损失函数;在提高小目标行人检测精度方面,采用scale-aware双路网络算法来增加对小目标行人的检测精度。结果 本文选择Caltech行人数据集作为对比数据集,选取YOLO（you only look once）、YOLOv3、SA-FastRCNN（scale-aware fast region-based convolutional neural network）等算法进行对比,在运行效率方面,本文算法在连续输入单帧图像的情况下达到了35帧/s,多图像输入时达到了70帧/s的工作效率;在模型精度测试中,本文算法也略胜一筹。本文算法应用于2020年中国智能汽车大赛中,在安全避障环节皆获得满分。结论 本文设计的尺度感知的行人检测算法,在EfficientDet高性能检测器的基础上,通过结合损失函数、scale-aware双路子网络的改进,进一步提升了本文检测器的鲁棒性。