Road Topology Research Articles

Traffic demand prediction based on spatial-temporal guided multi graph Sandwich-Transformer

The ability of spatial-temporal traffic demand prediction is crucial for urban computing, traffic management and future autonomous driving. In this paper, a novel Spatial-Temporal Guided Multi-graph Sandwich-Transformer (STGMT) is suggested to address the ubiquitous spatial-temporal heterogeneity in traffic demand forecasting. Compared to the original Transformer, we employ Time to Vector (Time2Vec) and Node to Vector (Node2Vec) in the embedding layer to obtain universal representations for temporal nodes and spatial nodes, respectively, which are then combined to form Spatial-Temporal Embedding (STE) blocks. The STE guides the attention mechanism, maintaining a unique parameter space for spatial-temporal nodes and enabling the learning of node-specific patterns. In STGMT, we develop Multi-head Temporal Attention (MTA) and Multi-head Temporal Interactive Attention (MTIA) for extracting temporal features, while Multi-head Spatial Attention (MSA) is employed for extracting spatial features. Furthermore, MSA incorporates both the accessibility graph determined by road topology and the similarity graph determined by specific traffic events to characterize the pairwise relationships among spatial nodes. Various attentions and feed-forward layers are rearranged and combined to form the Sandwich-Transformer. Extensive experiments are conducted on public datasets of node-level tasks of two different types (highway and urban) and indicate that the STGMT outperforms state-of-the-art models. The proposed STGMT effectively addresses the ubiquitous spatial-temporal heterogeneity challenge in traffic demand forecasting, thereby enhancing the accuracy of traffic demand prediction and offering valuable guidance for traffic planning and operations. Our code and data are open source at https://github.com/YanJieWen/STGMT-Tensorflow-implementation.

FitFun: A modelling framework for successfully capturing the functional form and noise of observed traffic flow–density–speed relationships

Measurements of the average properties of vehicular traffic are inherently noisy. The distributions of flow and speed measurements at any particular density are non-Gaussian with density-dependent variance, skewness, and kurtosis. Previous studies have failed to properly account for these complicated noise properties. In remediation, we present FitFun, a general framework for modelling any observed flow–density–speed relationship. Models specified within FitFun incorporate components for both the functional form and the noise. We define three flexible noise model components and we fit 200 different models to a high-quality sample of 10,150 observed urban flow-occupancy relationships. We compare the fits using information criteria and assess fit quality through analysis of the residuals. We find that the non-parametric Sun model for the functional form component combined with a Skew Exponential Power Type III noise component significantly outperforms all of the other models. Interestingly, we find that the city, country, road topology, and detector location have virtually no impact on model performance and fit quality, which is very convenient for model selection. The only factor of relevance from those that we studied is the effective occupancy coverage of the data. We conclude that certain models specified judiciously within FitFun can successfully capture the functional form and noise of observed flow–density–speed relationships without the need to discard data taken during non-stationary conditions. This is particularly advantageous for urban data where stationary traffic conditions are rarely observed.

Analysis and Forecast of Traffic Flow between Urban Functional Areas Based on Ride-Hailing Trajectories

Urban planning and function layout have important implications for the journeys of a large percentage of commuters, which often make up the majority of daily traffic in many cities. Therefore, the analysis and forecast of traffic flow among urban functional areas are of great significance for detecting urban traffic flow directions and traffic congestion causes, as well as helping commuters plan routes in advance. Existing methods based on ride-hailing trajectories are relatively effective solution schemes, but they often lack in-depth analyses on time and space. In the paper, to explore the rules and trends of traffic flow among functional areas, a new spatiotemporal characteristics analysis and forecast method of traffic flow among functional areas based on urban ride-hailing trajectories is proposed. Firstly, a city is divided into areas based on the actual urban road topology, and all functional areas are generated by using areas of interest (AOI); then, according to the proximity and periodicity of inter-area traffic flow data, the periodic sequence and the adjacent sequence are established, and the topological structure is learned through graph convolutional neural (GCN) networks to extract the spatial correlation of traffic flow among functional areas. Furthermore, we propose an attention-based gated graph convolutional network (AG-GCN) forecast method, which is used to extract the temporal features of traffic flow among functional areas and make predictions. In the experiment, the proposed method is verified by using real urban traffic flow data. The results show that the method can not only mine the traffic flow characteristics among functional areas under different time periods, directions, and distances, but also forecast the spatiotemporal change trend of traffic flow among functional areas in a multi-step manner, and the accuracy of the forecasting results is higher than that of common benchmark methods, reaching 96.82%.

Limi-TFP: Citywide Traffic Flow Prediction With Limited Road Status Information

Citywide traffic flow prediction is a basic but challenging task in the field of Intelligent Transportation Systems (ITS). Most existing traffic flow prediction methods require historical traffic data of all the target roads in both training and application stages. However, it is not easy, and usually costly, to collect real-time traffic data across the entire road network. Therefore, many existing methods have to narrow down the scope of their prediction on the parts of road networks. In this paper, we propose a deep-learning model, named Limi-TFP, which has the ability to identify a limited number of monitored roads, and achieves citywide traffic flow prediction by using the historical traffic data of these selected roads. Specifically, an embedding module is proposed to capture the spatial context (road topology structure) and the attributes (e.g., road type, length, lanes, etc.) of each road by embedding all road segments into vectors. A road ranking method is then developed for selecting a limited number of roads to be monitored. A group of multi-head attention mechanisms is exploited for capturing the dynamic correlations with each monitored road. Meanwhile, the fusion with external factors, including the point of interests (POIs) and weather data, is also taken into account to further improve the prediction accuracy. Extensive experiments on the real-world datasets demonstrate that the proposed method achieves the superior performance and is robust with the presence of noise compared to the state-of-the-art baselines, with only 5% roads being monitored.

Road Topology Extraction From Satellite Imagery by Joint Learning of Nodes and Their Connectivity

Road topology extraction from satellite images, which has long been of interest, is an essential task in remote sensing. The graph representation of road networks is one of the most challenging aspects of road topology extraction. Most existing approaches cast road extraction as binary segmentation and then use post-processing, such as skeletonization, to infer networks from pixel-wise prediction. In our work, we believe that a road network can be represented by an undirected graph denoted as G=(V,E), where V and E represent the set of road nodes and the set of edges between nodes, respectively. Thus, to construct the road topology, we propose NodeConnect, a new method of extracting nodes for a road network and inferring the connectivity between nodes. A convolutional neural network is jointly trained to predict the nodes and connectivity map for nodes, and the edges between nodes are inferred from the connectivity map. We compare our approach with several segmentation methods on the DeepGlobe and RoadTracer datasets. The experiments show that our approach achieves state-of-the-art performance in terms of pixel-based metrics and topological precision and recall.

Segmentation Is Not the End of Road Extraction: An All-Visible Denoising Autoencoder for Connected and Smooth Road Reconstruction

With a plethora of remote sensing (RS) images, deep neural network-based semantic segmentation models (SegModels) achieve commendable road extraction performance. However, the occlusions caused by vehicles, roadside objects and shadows cannot be directly identified as road pixels, especially on high-resolution RS images. Therefore, relying only on a single SegModel to guarantee road connectivity and boundary smoothness in road extraction tasks is extremely difficult. To address this issue, this paper puts forward a “Segmentation-with-Reconstruction” framework, which comprises a SegModel to generate the binary road labels from RS images, and a reconstruction model to refine the road labels. Specifically, the former can be compatible with arbitrary existing SegModels, while the latter is built by our proposed model named as all-visible denoising auto-encoder (AV-DAE). The AV-DAE is designed to be an encoder-decoder architecture that takes topology-corruption road labels as inputs and true road labels as outputs. To better train the AV-DAE, we further present three noise-adding strategies to corrupt road labels for diverse patterns, and train the AV-DAE to reconstruct them. Being RS-image-agnostic, the AV-DAE pays more attention to the spatial features rather than the spectral features, which enables it to recover the road topology through improving the connectivity and boundary smoothness. Finally, elaborate simulation results demonstrate that the proposed framework can significantly improve the connectivity and boundary smoothness of the extracted roads, while achieving a competitive road extraction performance and high generalization ability, as compared to the benchmarks.

Design of a Switching Nonlinear MPC for Emission Aware Ecodriving

Advanced automation and information systems are shaping the future of transportation by enabling advanced approaches, such as ecodriving control strategies, which improve safety and energy efficiency of personal transport. This paper contributes to the field of ecodriving control by developing a novel approach capable to deal with curvy roads, combining road grade and curvature look ahead to keep the driving comfortable, while also considering pollutant emissions. The formulation and solution of emission aware ecodriving as an optimal control problem is presented and, more specifically, a novel switching Nonlinear Model Predictive Control (NMPC) is proposed. Its feasibility and potential are shown by means of simulation case studies based on real world test cases, a validated vehicle model, and measured road topology. Hence, it is shown that emission aware ecodriving is a viable option and that developers of future ecodriving systems should expand their focus from solely energy efficiency towards additional targets.

Landscape and population drivers of ungulate-vehicle collisions in Portugal

Ungulate-vehicle collisions (UVC) have been steadily increasing throughout Europe, posing a risk to human safety, representing an extra factor of animals death. Here, we used information of a 2-year roadkill (2019–2020) monitoring of wild ungulates (wild boar, red deer and roe deer) to assess the drivers that influence the likelihood of UVC in Portugal. Our results support the contention that the variables promoting the collisions varied markedly between species, including the importance of land use and cover (wild boar and red deer), water availability and movement corridors (wild boar), ungulate population density (wild boar) and road topology (wild boar, red deer and roe deer) in the likelihood of UVC. The significant and positive interaction between urbanized and forest areas shows that regions sharing characteristics of both urban and natural environments represent a high-risk for the occurrence of collisions. The application of mitigation actions in these regions are welcome and should focus on reducing its attractiveness and/or the abundance of ungulate populations. Awareness campaigns and species-specific signalling may also contribute to a general decrease of casualties. We advocate that continued efforts of data collection are pivotal to assess the local relevance of fine-scale predictors amenable to manage.

Black Carbon Personal Exposure during Commuting in the Metropolis of Karachi

Black carbon (BC) exposure and inhalation dose of a commuter using four traffic modes (car, bus, auto-rickshaw, and motorbike) were monitored in Karachi, Pakistan. The real-time exposure concentrations in office-peak and off-peak hours were recorded during the winter season using microAeth® AE51 BC monitors. Exposure concentrations were higher in peak hours and were reduced to half in the off-peak time. The inclination levels of the inhaled dose were similar, and this trend was observed with all four modes of commute. The motorbike was found to be the most exposed mode of transportation, followed by auto-rickshaws, cars, and buses, respectively. However, the order was reversed when accounting for inhaled doses, e.g., the inhalation dose for auto rickshaws was highest, followed by the bus, motorbike, and car, respectively. Spatiotemporal analysis reveals that driving roads with lower traffic intensity and fewer intersections resulted in lower exposures. Therefore, traffic intensity, road topology, the timing of the trip, and the degree of urbanization were found to be the major influences for in-vehicle BC exposure.

SPRNN: A spatial–temporal recurrent neural network for crowd flow prediction

The capability of predicting the future trends of crowds has rendered crowd flow prediction more critical in building intelligent transportation systems, and attracted substantial research efforts. The trend of crowd flows is closely related to time and the urban topography. Therefore, extracting and leveraging both spatial features and temporal features are key gradients for effectively predicting crowd flows. Many previous works extract spatial features from crowd-flow data in an iteration way. As a result, models suffer from a heavy computation cost while ignoring details of road topology and structure information. Meanwhile, temporal features, including short-term features and long-term features, are separately extracted. The fusion of all features at the last stage before accomplishing the prediction also neglects the underlying associativity between various features. To address the limitations, we leverage spatial features by extracting structural information of road structures, such as road connection, road density, road width, etc. Rather than extracting spatial features from crowd-flow data, we capture them from images of city maps by adopting convolutional neural networks. Moreover, we implement a new sequence feature fusion mechanism to merge both spatial features and temporal features from various time scales so as to predict crowd flows. We conduct extensive experiments to evaluate our model on three benchmark datasets. The experimental results demonstrate that the model outperforms 15 state-of-the-art methods. The source code is available at: https://github.com/CVisionProcessing/SPRNN.

Dual Mode Localization Assisted Beamforming for mmWave V2V Communication

This study is motivated by the fact that localization in millimeter wave(mmWave) Vehicle-to-Vehicle communication becomes a more critical problem because both the terminals of the communication link are in motion. The positional awareness merely based on GPS or local sensors has an error margin of around 10 meters, which can worsen in uncertain real-time conditions such as road topology and highway traffic. The paper analyses the relation between vehicle dynamics, beamforming, and beam alignment for highly directive antennas. When the subsystem models presented in this paper are taken into consideration, the joint vehicle dynamics-beamforming approach will optimize the link connectivity and SNR. The vehicle dynamics model is designed to be more realistic considering the non-linear acceleration based on the throttle-brake jerks due to induced noises. The Unscented Kalman Filter based prediction subsystem predicts the alignment values even during non-linear acceleration and jerks. The feasibility of antenna design for the beamforming strategies is supported by the requirements of phase shifters and the number of elements.

Optimal Planning of Electric Vehicle Battery Centralized Charging Station Based on EV Load Forecasting

This article studies the planning of swapping electric vehicle (SEV) battery centralized charging station (BCCS) based on EV spatial-temporal load forecasting. First, according to the road topology of the planning area, the road topology model is built. The driving energy consumption process and endurance mileage model of EVs are developed. Then, we analyze the travel demand of people, the driving segments of SEVs (e.g., swapping private EVs and swapping taxis) are analyzed. Next, the spatial distribution model of SEVs is built. Thus, the spatial-temporal load model of SEVs is established. The sizing and locating planning model based on operation analysis is established for BCCSs. In addition, the photovoltaic (PV) configuration model in the urban grid is also established. Finally, considering the coupling relationship between urban distribution system and power line pipes, a grid planning method considering power line pipes is proposed. The planning model of BCCS aimed at minimizing annual comprehensive cost considering distribution system with PV is constructed. The MATLAB/YALMIP toolbox is applied to optimize the planning model in this article, and the feasibility and effectiveness of the proposed method are verified.

SP-TRUST: a trust management model for speed trust in vehicular networks

ABSTRACT Information security mechanisms are crucial for Vehicular Ad-hoc NETworks (VANET) applications in order to preserve their robustness. The efficiency of these applications relies on the reliability of the used information, especially the vehicle-related information such as location and speed. Trust management models are crucial to evaluate the quality of the used information. Accordingly, we focus in this paper on the assessment of speed information to detect the malicious vehicles using a fuzzy-based model (SP-TRUST). The proposed model relies on two fuzzy inference systems. The first one evaluates the speed trust based on traffic rules (inter-vehicle distance) while considering the road topology (angle deviation) and the traffic state (density). The correlation between these parameters and the speed value is assessed. The second inference system assesses the speed trust based on the behavior of neighbor vehicles using the median speed of vehicles. Simulations were carried out to evaluate the robustness and the scalability of SP-TRUST model. The results of the experimental studies proved that the model performs well in detecting different behaviors of malicious vehicles in different scenarios, especially when the percentage of malicious vehicles is lower than 50%.

Learning Based Energy Efficient Task Offloading for Vehicular Collaborative Edge Computing

Extensive delay-sensitive and computation-intensive tasks are involved in emerging vehicular applications. These tasks can hardly be all processed by the resource constrained vehicle alone, nor fully offloaded to edge facilities (like road side units) due to their incomplete coverage. To this end, we refer to the new paradigm of vehicular collaborative edge computing (VCEC) and make the best use of vehicles’ idle and redundant resources for energy consumption reduction within the VCEC system. To realize this target, we are faced with several nontrivial challenges, including short-term decision making coupled with long-term queue delay constraints, information uncertainty, and task offloading conflicts. Accordingly, we apply Lyapunov optimization to decouple the original problem into three sub-problems and then tackle them one by one: the first sub-problem is resolved by Lagrange multiplier method; the second is handled by UCB learning-matching approach; the third is addressed by a carefully designed greedy method. Scenarios without volatility and real-world road topology with realistic vehicular traffics are utilized to evaluate the proposed solution. Results from extensive numerical simulations demonstrate that our solution can achieve superior performances compared with the benchmark methods, in terms of energy consumption, learning regret, task backlog, and end-to-end delay.

Probabilistic Traffic Motion Labeling for Multi-Modal Vehicle Route Prediction.

The prediction of the motion of traffic participants is a crucial aspect for the research and development of Automated Driving Systems (ADSs). Recent approaches are based on multi-modal motion prediction, which requires the assignment of a probability score to each of the multiple predicted motion hypotheses. However, there is a lack of ground truth for this probability score in the existing datasets. This implies that current Machine Learning (ML) models evaluate the multiple predictions by comparing them with the single real trajectory labeled in the dataset. In this work, a novel data-based method named Probabilistic Traffic Motion Labeling (PROMOTING) is introduced in order to (a) generate probable future routes and (b) estimate their probabilities. PROMOTING is presented with the focus on urban intersections. The generation of probable future routes is (a) based on a real traffic dataset and consists of two steps: first, a clustering of intersections with similar road topology, and second, a clustering of similar routes that are driven in each cluster from the first step. The estimation of the route probabilities is (b) based on a frequentist approach that considers how traffic participants will move in the future given their motion history. PROMOTING is evaluated with the publicly available Lyft database. The results show that PROMOTING is an appropriate approach to estimate the probabilities of the future motion of traffic participants in urban intersections. In this regard, PROMOTING can be used as a labeling approach for the generation of a labeled dataset that provides a probability score for probable future routes. Such a labeled dataset currently does not exist and would be highly valuable for ML approaches with the task of multi-modal motion prediction. The code is made open source.

Enhanced UAV-aided vehicular delay tolerant network (VDTN) routing for urban environment using a bio-inspired approach

The evolution of VDTN routing came to solve the problem of routing in sparse Vehicular Ad-hoc Networks (VANETs). Recently, the implication of Unmanned Aerial Vehicles (UAVs) in VANET communication has gathered quicker information dissemination than Roadside Units (RSUs) in urban VANETs, where ground vehicles’ mobility is restricted by the topology of city roads. Gradually, UAVs are emerging for VDTNs: UAV-aided VDTN (UVDTN) routing field has been introduced for sparse areas in order to reduce the network gaps between Vehicle-to-Vehicle (V2V) connections. Meanwhile, the conventional UVDTN routing protocols expose the short contact opportunity time in the case of Vehicle-to-UAV (V2U) and UAV-to-Vehicle (U2V) connections. This constraint entails more congested bundles’ flooding rates and a higher number of lost bundle copies compared to V2V-based forwarding. In this paper, a bio-inspired UVDTN routing protocol is conceived to improve the bundles’ forwarding between ground vehicles and UAVs in terms of flooding optimization and Quality-of-Service (QoS) performances for urban VANET environments. For this purpose, a swarm-inspired approach called the Social Spider Optimization (SSO) is proposed. SSO seeks to find a better selection of next Store-Carry-and-Forward (SCF) relay nodes specifically for V2U and U2V connections. In order to control the bundle replication rate and enhance bundles’ delivery ratio and delay, the proposed bio-inspired UVDTN protocol adjusts the SCF selection decision. The latter is based on predefined historical knowledge-based forwarding parameters and urban geographic topology information. The proposed bio-inspired routing protocol is simulated using the Opportunistic Network Environment (ONE) simulator and compared to a few VDTN and UVDTN routing protocols.

A trajectory map matching algorithm via exploring the average directional features of continuous windows

Many location-based services require a pre-processing step of map matching. Due to the error of the original position data and the complexity of the road network, the matching algorithm will have matching errors when the complex road network is implemented, which is therefore challenging. Aiming at the problems of low matching accuracy and low efficiency of existing algorithms at Y-shaped intersections and roundabouts, this paper proposes a space-time-based continuous window average direction feature trajectory map matching algorithm (STDA-matching). Specifically, the algorithm not only adaptively generates road network topology data, but also obtains more accurate road network relationships. Based on this, the transition probability is calculated by using the average direction feature of the continuous window of the track points to improve the matching accuracy of the algorithm. Secondly, the algorithm simplifies the trajectory by clustering the GPS trajectory data aggregation points to improve the matching efficiency of the algorithm. Finally, we use a real and effective data set to compare the algorithm with the two existing algorithms. Experimental results show that our algorithm is effective.

Optimal Electric Vehicles Route Planning with Traffic Flow Prediction and Real-Time Traffic Incidents

Electric Vehicles (EVs) are regarded to be among the most environmentally and economically efficient transportation solutions. However, barriers and range limitations hinder this technology’s progress and deployment. In this paper, we examine EV route planning to derive optimal routes considering energy consumption by analyzing historical trajectory data. More specifically, we propose a novel approach for EV route planning that considers real-time traffic incidents, road topology, charging station locations during battery failure, and finally, traffic flow prediction extracted from historical trajectory data to generate energy maps. Our approach consists of four phases: the off-line phase which aims to build the energy graph, the application of the A* algorithm to deliver the optimal EV path, the NEAT trajectory clustering which aims to produce dense trajectory clusters for a given period of the day, and finally, the on-line phase based on our algorithm to plan an optimal EV path based on real traffic incidents, dense trajectory clusters, road topology information, vehicle characteristics, and charging station locations. We set up experiments on real cases to establish the optimal route for electric cars, demonstrating the effectiveness and efficiency of our proposed algorithm.

Multisensor Management Method for Ground Moving Target Tracking Based on Doppler Blind Zone Information

Tracking ground moving target with sensors proves to be a challenge due to the uncertainty of target motion area, the existence of Doppler blind zone (DBZ), and the complex terrain. In this paper, a multisensor management method based on DBZ information is presented, in which the available sensors are selected to obtain the best operational revenues for ground moving target tracking. First, the ground target motion model is established considering the off-road/on-road state based on road topology information. Second, the sensor measurement model is given combined with the DBZ information, and a decorrelation method of measurement noise is proposed. Third, a target state estimation algorithm is derived using particle filter, in which the DBZ information is regarded as prior information. Then, combined with the variable structure interacting multiple model method, an estimation algorithm for tracking maneuvering target is proposed. Furthermore, an optimization model of nonmyopic sensor management is constructed to obtain the best sensor management scheme. Finally, the advancement and effectiveness of the proposed management method are verified in the simulations.

Unraveling Gap Selection Process During Discretionary Lane Changing by Vehicle Class

This paper studies and compares the gap selection process of multiple vehicle classes (passenger cars, delivery vans, and trucks) within their discretionary lane changing activities. Given a trajectory or a sequence of gap selection decisions, we aim to predict whether a vehicle will change or keep a lane. For this purpose, we use a large trajectory dataset, collected for the Netherlands, consisting of 3,647 trajectories of passenger car drivers, 1,080 trajectories of delivery van drivers, and 2,226 trajectories of truck drivers. We apply gated recurrent unit neural networks to separately model their gap selection processes. These three models can not only handle class imbalance but also capture long-term interdependencies. The models can predict gap selection of three vehicle classes with geometric mean accuracies of 84% or higher. To obtain insights into their gap selection processes, we apply a gradient-based technique to analyze what neural networks have learned. Our results suggest that there exist significant differences between vehicle classes in terms of the importance of historical information and features. Trucks seem to value a relatively long period, recent 6 seconds, of driving experience to select gaps compared to passenger cars and delivery vans. In addition, the perception of road topology seems to be a significant factor for delivery vans and trucks, contrary to passenger cars which highly value their kinematic features and interactions with surrounding vehicles, to select gaps. These insights offer a novel contribution towards better understanding and modeling of the driving behavior of multiple vehicle classes.