Traffic flow prediction of urban road network based on LSTM-RF model

Traffic flow prediction,as the basis of signal coordination and travel time prediction,has become a research point in the field of transportation.For traffic flow prediction,researchers have proposed a variety of methods,but most of these methods only use the time domain information of traffic flow data to predict the traffic flow,ignoring the impact of spatial correlation on the prediction of target road segment flow,which leads to poor prediction accuracy.In this paper,a traffic flow prediction model called as long short time memory and random forest(LSTM-RF)was proposed based on the combination model.In the process of traffic flow prediction,the long short time memory(LSTM)model was used to extract the time sequence features of the predicted target road segment.Then,the predicted value of LSTM and the collected information of adjacent upstream and downstream sections were simultaneously used as the input features of the random forest model to analyze the spatial-temporal correlation of traffic flow,so as to obtain the final prediction results.The traffic flow data of 132 urban road sections collected by the license plate recognition system in Guiyang City were tested and verified.The results show that the method is better than the single model in prediction accuracy,and the prediction error is obviously reduced compared with the single model.

Research on a non-linear chaotic prediction model for urban traffic flow

In order to solve serious urban transport problems, according to the proved chaotic characteristic of traffic flow, a non linear chaotic model to analyze the time series of traffic flow is proposed. This model reconstructs the time series of traffic flow in the phase space firstly, and the correlative information in the traffic flow is extracted richly, on the basis of it, a predicted equation for the reconstructed information is established by using chaotic theory, and for the purpose of obtaining the optimal predicted results, recognition and optimization to the model parameters are done by using genetic algorithm. Practical prediction research of urban traffic flow shows that this model has famous predicted precision, and it can provide exact reference for urban traffic programming and control.

Traffic chaos and its prediction based on a nonlinear car-following model

This paper discusses the dynamic behavior and its predictions for a simulated traffic flow based on the nonlinear response of a vehicle to the leading car＇s movement in a single lane. Traffic chaos is a promising field, and chaos theory has been applied to identify and predict its chaotic movement. A simulated traffic flow is generated using a car-following model（ GM model）, and the distance between two cars is investigated for its dynamic properties. A positive Lyapunov exponent confirms the existence of chaotic behavior in the GM model. A new algorithm using a RBF NN （radial basis function neural network） is proposed to predict this traffic chaos. The experiment shows that the chaotic degree and predictable degree are determined by the first Lyapunov exponent. The algorithm proposed in this paper can be generalized to recognize and predict the chaos of short-time traffic flow series

Control Chaos of a Unidirectionally Traffic Coupled Map Lattice Model

Spatiotemporal chaos is studied by using the unidirectional traffic coupled lattice model with hyperbolic tangent local map. The coupled map lattice (CML) model can simulate the complex traffic flow phenomenon which is similar to the traditional traffic flow model. The nonlinear feedback method is used to study the control of the chaotic system of the unidirectionally traffic coupled map lattice model. The stability of spatiotemporal chaos in the coupled map lattice is realized. The results of numerical simulation show that there is a relationship between control results and control parameters when controlling spatiotemporal chaos to a uniform stable state in a certain phase space compression parameter region.

Network traffic prediction by a wavelet-based combined model

Network traffic prediction models can be grouped into two types, single models and combined ones. Combined models integrate several single models and thus can improve prediction accuracy. Based on wavelet transform, grey theory, and chaos theory, this paper proposes a novel combined model, wavelet-grey-chaos （WGC）, for network traffic prediction. In the WGC model, we develop a time series decomposition method without the boundary problem by modifying the standard à trous algorithm, decompose the network traffic into two parts, the residual part and the burst part to alleviate the accumulated error problem, and employ the grey model GM（1,1） and chaos model to predict the residual part and the burst part respectively. Simulation results on real network traffic show that the WGC model does improve prediction accuracy.

Inatorial forecasting method considering macro and micro characteristics of chaotic traffic flow

Traffic flow prediction is an effective strategy to assess traffic conditions and alleviate traffic congestion. Influenced by external non-stationary factors and road network structure, traffic flow sequences have macro spatiotemporal characteristics and micro chaotic characteristics. The key to improving the model prediction accuracy is to fully extract the macro and micro characteristics of traffic flow time sequences. However, traditional prediction model by only considers time features of traffic data, ignoring spatial characteristics and nonlinear characteristics of the data itself, resulting in poor model prediction performance. In view of this, this research proposes an intelligent combination prediction model taking into account the macro and micro features of chaotic traffic data. Firstly, to address the problem of time-consuming and inefficient multivariate phase space reconstruction by iterating nodes one by one, an improved multivariate phase space reconstruction method is proposed by filtering global representative nodes to effectively realize the high-dimensional mapping of chaotic traffic flow. Secondly, to address the problem that the traditional combinatorial model is difficult to adequately learn the macro and micro characteristics of chaotic traffic data, a combination of convolutional neural network(CNN) and convolutional long short-term memory(ConvLSTM) is utilized for capturing nonlinear features of traffic flow more comprehensively. Finally,to overcome the challenge that the combined model performance degrades due to subjective empirical determined network parameters, an improved lightweight particle swarm is proposed for improving prediction accuracy by optimizing model hyperparameters. In this paper, two highway datasets collected by the Caltrans Performance Measurement System(PeMS)are taken as the research objects, and the experimental results from multiple perspectives show that the comprehensive performance of the method proposed in this research is superior to those of the prevalent methods.

Intelligent Slime Mould Optimization with Deep Learning Enabled Traffic Prediction in Smart Cities

Intelligent Transportation System(ITS)is one of the revolutionary technologies in smart cities that helps in reducing traffic congestion and enhancing traffic quality.With the help of big data and communication technologies,ITS offers real-time investigation and highly-effective traffic management.Traffic Flow Prediction(TFP)is a vital element in smart city management and is used to forecast the upcoming traffic conditions on transportation network based on past data.Neural Network(NN)and Machine Learning(ML)models are widely utilized in resolving real-time issues since these methods are capable of dealing with adaptive data over a period of time.Deep Learning(DL)is a kind of ML technique which yields effective performance on data classification and prediction tasks.With this motivation,the current study introduces a novel Slime Mould Optimization(SMO)model with Bidirectional Gated Recurrent Unit(BiGRU)model for Traffic Prediction(SMOBGRU-TP)in smart cities.Initially,data preprocessing is performed to normalize the input data in the range of[0,1]using minmax normalization approach.Besides,BiGRUmodel is employed for effective forecasting of traffic in smart cities.Moreover,the novelty of the work lies in using SMO algorithm to effectively adjust the hyperparameters of BiGRU method.The proposed SMOBGRU-TP model was experimentally validated and the simulation results established the model’s superior performance in terms of prediction compared to existing techniques.

Towards explainable traffic flow prediction with large language models

Traffic forecasting is crucial for intelligent transportation systems.It has experienced significant advancements thanks to the power of deep learning in capturing latent patterns of traffic data.However,recent deep-learning architectures require intricate model designs and lack an intuitive understanding of the mapping from input data to predicted results.Achieving both accuracy and explainability in traffic prediction models remains a challenge due to the complexity of traffic data and the inherent opacity of deep learning models.To tackle these challenges,we propose a traffic flow prediction model based on large language models(LLMs)to generate explainable traffic predictions,named xTP-LLM.By transferring multi-modal traffic data into natural language descriptions,xTP-LLM captures complex time-series patterns and external factors from comprehensive traffic data.The LLM framework is fine-tuned using language-based instructions to align with spatial-temporal traffic flow data.Empirically,xTP-LLM shows competitive accuracy compared with deep learning baselines,while providing an intuitive and reliable explanation for predictions.This study contributes to advancing explainable traffic prediction models and lays a foundation for future exploration of LLM applications in transportation.

A Survey of Model Predictive Control Methods for Traffic Signal Control

Enhancing traffic efficiency and alleviating(even circumventing) traffic congestion with advanced traffic signal control(TSC) strategies are always the main issues to be addressed in urban transportation systems. Since model predictive control(MPC) has a lot of advantages in modeling complex dynamic systems, it has been widely studied in traffic signal control over the past 20 years. There is a need for an in-depth understanding of MPC-based TSC methods for traffic networks. Therefore, this paper presents the motivation of using MPC for TSC and how MPC-based TSC approaches are implemented to manage and control the dynamics of traffic flows both in urban road networks and freeway networks. Meanwhile, typical performance evaluation metrics, solution methods, examples of simulations,and applications related to MPC-based TSC approaches are reported. More importantly, this paper summarizes the recent developments and the research trends in coordination and control of traffic networks with MPC-based TSC approaches. Remaining challenges and open issues are discussed towards the end of this paper to discover potential future research directions.

大型IP网络流量矩阵分析预测的探讨研究

高效、准确的网际协议(internet protocol,IP)网络流量流向分析预测是网络规划建设的基础。通过部署流量采集分析系统,运营商可轻松获取网络总流量、节点流量、节点分方向流量等较完备的历史基础数据,为流量分析预测提供关键的输入。IP网络流量分析预测方法主要包括两类:传统统计模型和神经网络模型,近年提出的NeuralProphet模型因结合两者优点而得到广泛关注和应用。首次基于NeuralProphet模型对大型运营级IP网络源节点到目的节点的流量流向进行直接预测,并采用改进的损失函数优化模型训练,预测结果表明NeuralProphet模型能够更科学、准确地预测IP网络流量矩阵,整体预测精度提升了8.7%,同时模型扩展性和鲁棒性也具有更佳的表现,可以更好地满足IP网络规划建设和运行维护的实际需求。

基于GWO-HMM的空中交通网络流系统态势预测研究

针对空中交通流量管理部门如何更高效地实施流量管理的问题,本文将态势感知理论应用于空中交通网络流系统(ATNFS,air traffic network flow system),建立空中交通网络流系统的运行态势预测模型。首先,给出了空中交通网络流系统的态势感知过程,从节点和航线的角度筛选出航线饱和度、不正常航班率、节点饱和度、节点延误架次比、节点航班取消率5个态势要素,使用态势值作为态势理解的指标;其次,分析隐马尔可夫模型(HMM,hidden Markov model)的优势与不足,建立了基于灰狼优化(GWO,grey wolf optimization)算法和改进隐马尔可夫模型的态势预测模型;最后,使用某空中交通网络流系统的实际运行数据进行算例验证。结果表明,改进后的预测模型相较于原本的隐马尔可夫预测模型精度更高,预测结果更准确。

领域知识与数据驱动的混合交通流车辆轨迹预测

自动驾驶车辆需具备预测周围车辆轨迹的能力。诸多发展中国家普遍存在弱规则、强交互的混合交通流道路,高密度混合交通流的车辆轨迹预测是极具挑战性的任务。为了兼顾混合交通流道路环境下轨迹预测的高精度和可解释性,设计一个融合领域知识和经验的深度学习模型(DKConv-LSTM)实现车辆的长、短时轨迹预测。该模型采用卷积结构(Conv)提取交互特征,并将融合车辆历史信息的特征向量送入长短时记忆网络(LSTM)模型实现轨迹预测。知识经验通过嵌入损失函数的方式引导深度学习模型的训练。与基础的LSTM相比,仅添加卷积层结构的Conv-LSTM模型可提升终点轨迹误差(FDE)约30.46%,提升平均轨迹误差(ADE)约34.78%;而DK-Conv-LSTM模型可分别提升FDE 46.81%和ADE 49.08%;同时DK-Conv-LSTM模型可还原多前车跟驰、超车行为的驾驶轨迹。

基于误差补偿的多模态协同交通流预测模型

交通流量因受周期性特征、突发状况等多重因素影响,现有模型的预测精度无法满足实际要求.对此,本文提出了基于误差补偿的多模态协同交通流预测模型(Multimodal Collaborative traffic flow prediction model based on Error Compensation,MCEC).针对传统预测模型不能兼顾时间序列和协变量的问题,提出基于小波分析的特征拓展方法,该方法引入聚类算法得到节假日标签特征,将拥堵指数、交通事故图、天气信息作为拓展特征,对特征进行多尺度分解.在训练阶段,为达到充分学习各部分数据、最优匹配模型的效果,采用差分整合移动平均自回归模型(Autoreg Ressive Integrated Moving Average Model,ARIMA)、长短期记忆神经网络(Long Short-Term Memory network,LSTM)、限制动态时间规整技术(Dynamic Time Warping,DTW)以及自注意力机制(Self-Attention),设计了多模态协同模型训练.在误差补偿阶段,将得到的相应过程值输入基于支持向量机回归(Support Vector Regression,SVR)的误差补偿模块,对各分量的误差进行学习、补偿,并重构得到预测结果.使用公开的高速公路数据集对MCEC进行验证,在多个时间间隔下对比实验结果表明,MCEC在交通流量预测中的平均绝对百分比误差(Mean Absolute Percentage Error,MAPE)达到17.02%,比LSTM-SVR、ConvLSTM(Convolutional Long Short-Term Memory network)、ST-GCN(Spatial Temporal Graph Convolutional Networks)、MFFB(Multi-stream Feature Fusion Block)、Transformer等预测模型具有更高的预测精度,MCEC模型具有较好的有效性与合理性.

基于图提示微调的交通流量预测

交通流量预测是建设智慧城市重要的基础功能,对城市的交通管理和用户出行规划具有重要意义.由于时间维度和空间维度的扩展,交通流量的数据具有规模大、增长快速、实时更新等特征,传统的训练模型通常需要将大量的历史数据进行训练预测,导致较长的计算时间和较高的算力成本,因此,如何使用低计算成本的预测模型来满足广泛的流量预测需求是重要的技术挑战.近年来兴起的提示微调范式在自然语言处理的下游任务推广中取得了较好的效果,受其启发,提出利用少量的实时数据来微调优化大规模历史数据预训练的模型,为交通流量模型预测的优化应用提出了一种新的思路.通过引入图提示微调的交通流量预测(traffic flow prediction based on graph prompt-finetuning,TPGPF)模型的泛化能力,在时空多维度下的交通流量图预测模型中,基于历史数据集进行预测模型的预训练,并引入可学习的提示向量,在预训练模型固化的情况下指导预训练的自监督学习模型,以适应新的数据预测任务,提升交通流量预测模型的通用性和有效性.通过在5个公开数据集上进行了大量的实验,证明了TPGPF的有效性.

结合变种残差模型和Transformer的城市公路短时交通流预测

城市公路交通流的预测受到历史交通流量和相邻车道交通流量的影响,蕴含了复杂的时空特征。针对传统交通流预测模型卷积长短时记忆(ConvLSTM)网络进行交通流预测时,未将时空特征分开提取而造成的特征提取不充分、特征信息混淆和特征信息缺失等问题,对ConvLSTM模型作出改进。首先,提取每个采样时刻的交通流数据的短期时间特征和空间特征,并在特定的维度下将交通流的短期时空特征融合;其次,进行残差映射;最后,将映射后的短期时空特征交由Transformer模型捕捉交通流数据长期的时空特征,并根据所捕捉的长期特征对未来时刻每个采样点交通流进行预测。使用加州城市快速路数据对模型进行验证,以平均绝对误差(MAE)作为模型评价指标时,所提模型相较于Conv-Transformer模型,预测精度提高了18%,验证了所提模型的有效性。

自适应多视图融合图神经网络地铁客流预测模型

针对传统方法对地铁车站的多视角空间交互建模不足的问题,本文提出自适应多视图融合图神经网络模型(Adaptive Multi-view Fusion Graph Neural Network Model, AMFGNN)进行地铁车站短时客流预测。在空间维度,模型包括了物理拓扑图、线路可达性图、空间距离图等多个局部视图,并使用图注意力网络(Graph Attention Networks, GAT)学习每个视图内车站间的动态空间交互;以单视图车站为中心节点,结合其他视图中该车站作为邻居节点构建融合视图,并使用GAT学习多视图间动态交互;在时间维度,使用门控循环单元神经网络学习车站客流的时变特征。以重庆市地铁网络为例,全网出站客流的预测实验结果表明:相较于基线中的物理虚拟结合图网络模型(PVCGN),AMFGNN的平均绝对误差和均方根误差分别降低3.06%和2.49%。多视图内节点间注意力分数可视化结果表明,基于GAT的多视图建模思路能够自适应地融合不同视图中提取到的车站空间信息。此外,AMFGNN模型性能影响因素分析结果表明,以物理拓扑、线路可达性等结构稳定的视图作为中心节点构建融合视图能够获得更准确、稳定的预测模型。

基于DACO-Bi-LSTM的交通流量预测

针对交通流量预测任务存在预测精度低、泛化性不足且对深度学习模型调参不全面等问题,提出了一种基于改进蚁群优化算法的双向LSTM交通流量预测模型,利用改进蚁群算法的全局寻优能力对Bi-LSTM网络的层数、神经元个数、批次大小、训练次数进行优化调参。在英国高速公路和深圳政府开放平台发布的宝安区日车流量两个公开数据集上进行实验,以RMSE、MAE为评估指标,结果表明:DACO-Bi-LSTM模型具有较强的寻优能力,同时表现出更好的预测性能。

面向高速公路连续瓶颈的协同可变限速控制

高速公路瓶颈处极易因通行能力受限而诱发交通拥堵,进而导致通勤时间延长、燃料消耗增加、驾驶舒适性降低,尤其是在连续多瓶颈处,更容易对道路网络造成连锁反应而大范围增加事故风险。既有拥堵控制策略聚焦于孤立瓶颈处拥堵消除,通过对多个瓶颈路段实施多个孤立瓶颈控制策略,由于控制策略间交通状态与控制参数缺乏协同,致使控制效果受限,且极易因策略协同失衡导致更为复杂的交通拥堵。本文提出一种基于模型预测控制方法的连续多瓶颈拥堵消除的协同可变限速控制策略:首先,针对连续多瓶颈主动管控场景改进元胞传输模型,以模拟高速公路多瓶颈生成、通行能力下降、随机交通流波动和可变限速控制状态与效果;其次,基于拉格朗日坐标系改进LWR模型,以精准预测交通拥堵场景下的交通流随机状态;再次,构建以多瓶颈路段总体交通运行状态作为输入的可变限速控制策略,并采用反馈式控制框架协同多瓶颈可变限速控制策略;最后,通过构建仿真实验测试了所提出控制策略消除连续瓶颈处拥堵的控制效率,并与其他控制策略进行了效果对比。实验结果表明,本文所提出的协同可变限速策略可有效消除连续瓶颈的多处拥堵,且控制效果显著优于其他控制策略,车辆总行程时间和总延误时间分别降低21.3%和70.6%。

基于双通道注意力机制的AE-BIGRU交通流预测模型

交通流预测是智能交通系统的关键。针对目前交通流数据复杂的时空关联性以及自身的不确定性,为准确预测高速公路交通流并缓解交通拥堵问题,提出以自编码器网络(AE)和双向门控循环单元(BIGRU)相结合的深度学习组合预测模型(AE-BIGRU),并在此基础上引入双通道注意力机制进行模型训练。将预处理后的数据采用滑动窗口的方式作为参数输入模型,通过AE提取交通流的空间特征,得到输入信息特征的最优抽象表示;利用BIGRU从前向和后向传播中获取信息,充分提取交通流的时间相关特征,更全面地捕捉时间演变规律;最后结合双通道注意力机制,增强预测模型的特征提取能力,最大限度地保留特征信息,提升模型的预测精度,从而得到最终短时流量的预测目标值。为验证模型的适用性,采用多组短时交通流数据进行仿真实验,与其他基准模型对比发现:该交通流预测模型能够有效捕获交通流的动态时空特征,加强关键信息的提取,所预测的流量更加接近真实值,具有良好的泛化能力。其中测试集的均方根误差值下降了约0.061~0.604,平均绝对误差值下降了约0.025~0.512,相关系数值R2提高了约0.007~0.062。研究结果表明,随着预测步长的增加,该实验模型在交通流数据的时间特性上仍能表现出稳定的预测性能,所建的组合预测模型在预测精度和鲁棒性方面表现出更高水平。