High-resolution Trajectory Data Research Articles

Near-crash risk identification and evaluation for takeout delivery motorcycles using roadside LiDAR

The proliferation of motorcycles in urban areas has raised concerns regarding traffic safety. However, traditional sensors struggle to obtain precise high-resolution trajectory data, which hinder the accurate identification and quantification of near-crash risks for takeout delivery motorcycles. To fill this gap, this study presents a novel approach utilizing roadside light detection and ranging (LiDAR) to identify and evaluate the risk of near crashes of takeout delivery motorcycles. First, a trajectory amendment method incorporating speed and steering angle was introduced to enhance the accuracy and continuity of the trajectory prediction. Second, a trajectory prediction method combining the steering intention and a repulsive force model was proposed for near-crash risk prediction. Subsequently, a near-crash identification method was developed that relied on the closest distance and risk radius. Finally, near-crash risk fields were constructed to quantify risk levels by leveraging velocity, position, and weight. The experimental results demonstrated 92.10 % accuracy in intention prediction, with mean absolute error (MAE) and root mean square error (RMSE) values of 0.53 m and 0.45 m, respectively. In addition to its higher accuracy, the proposed method makes it easier to quantify near-crash risk and supports a proactive approach for visualizing and analyzing traffic safety.

Trajectory-based vehicle emission evaluation for signalized intersection using roadside LiDAR data

Signalized intersections are the bottleneck of urban traffic network and often lead to traffic congestion and increased traffic emissions. Studying and analyzing the spatiotemporal emission patterns at intersections is the prerequisite for traffic emission reduction. This study develops a novel framework for vehicle emission estimation using high-resolution trajectory data based on the roadside light detection and ranging sensor. A meshing method was proposed to divide the area and the trajectory repair model was developed by considering multiple influenced factors to obtain refined LiDAR trajectory data. Then, the Virginia Tech microscopic model was used to perform traffic emission spatiotemporal analysis. In the experiments, trajectory repair results indicate that Root Mean Square Error and Mean Absolute Error are 1.1299m and 0.7816m within 3s, which has excellent prediction accuracy and provides a reliable data basis for accurately estimating emissions. Emission results show that vehicle speed and acceleration are positively correlated with emissions, with the highest emissions generated during acceleration, of which CO could reach 0.0037 g/s. From a temporal perspective, emissions gradually decrease during red light phases and increase when vehicles accelerate during green light phases, suggesting recommendations for signal optimization for areas where emissions exceed standards to reduce vehicle's start and stop. From a spatial perspective, emission rates are highest in the downstream area, with CO reaching more than 0.0043 g/s. It can be suggested that relevant authorities could install speed limit signs or plant trees in this area. Overall, these findings have the potential to alleviate emission pressures at intersections.

Insights into vehicle conflicts based on traffic flow dynamics

The utilization of traffic conflict indicators is crucial for assessing traffic safety, especially when the crash data is unavailable. To identify traffic conflicts based on traffic flow characteristics across various traffic states, we propose a framework that utilizes unsupervised learning to automatically establish surrogate safety measures (SSM) thresholds. Different traffic states and corresponding transitions are identified with the three-phase traffic theory using high-resolution trajectory data. Meanwhile, the SSMs are mapped to the corresponding traffic states from the perspectives of time, space, and deceleration. Three models, including k-means, GMM, and Mclust, are investigated and compared to optimize the identification of traffic conflicts. It is observed that Mclust outperforms the others based on the evaluation metrics. According to the results, there is a variation in the distribution of traffic conflicts among different traffic states, wide moving jam (phase J) has the highest conflict risk, followed by synchronous flow (phase S), and free flow (phase F). Meanwhile, the thresholds of traffic conflicts cannot be fully represented by the same value through different traffic states. It reveals that the heterogeneity of thresholds is exhibited across traffic state transitions, which justifies the necessity of dynamic thresholds for traffic conflict analysis.

A constrained spectral clustering method for lane identification using trajectory data

The rapid development of information and communications technologies acts as an enabler for the successful implementation of vehicle automation and advanced traffic management applications. In particular, the appearance of new sources of high-resolution trajectory data, such as videos obtained from drones, provides an opportunity to build accurate maps and enrich applications in traffic research at an unprecedented resolution. However, existing methods cannot handle certain features, such as, among others, accurate of road lane identification.This paper proposes a constrained spectral clustering method to identify lane information from high-resolution trajectory data. Contrary to state-of-the-art methods, such as the Gaussian mixture model, the proposed method is directly applicable to two-dimensional trajectory data, without assuming a constant number of lanes characterised by the same lane width. The trajectory data is clustered by taking into account the neighbourhood distances and prior knowledge via defining so-called must-link and cannot-link constraints, which significantly improve the clustering results, especially in cases where the number of lanes or the lane width changes. The proposed method has been evaluated through numerical experiments using data obtained from drone videos, and the results indicate that the method performs well on complex road segments, even in the presence of a varying number of lanes or lane-changing manoeuvres.

High-resolution reconstruction of non-motorized trajectory in shared space: A new approach integrating the social force model and particle filtering

High-resolution trajectory data from the intersection shared space (ISS) is an ideal resource for testing autonomous driving and studying microscopic traffic flow theory in urban road traffic environments. It is common to use computer vision (CV) technologies to extract trajectory data from camera recordings at the ISS. However, it is challenging to accurately extract non-motorized (NONM) trajectories due to vehicle occlusion and mutual interference, which cause trajectory noise and missing. This study proposed a novel NONM trajectory reconstruction method, which integrates the social force model with the particle filter to reduce trajectory noise. The approach was tested on trajectories extracted from two advanced CV algorithms (ie, Yolov7 and MASK RCNN), and its performance was compared with seven SOTA methods. The results show that the proposed method achieved a much better RMSE (0.6 meters) than both the baseline (unprocessed CV trajectories, the RMSE is 1.16 meters) and selected SOTA methods (the best RMSE is 0.78 meters). The contribution of the study is to enhance the classical noise reduction algorithm by introducing the driving interaction model; ablation experiments and sensitivity analysis further demonstrate the importance of the driving interaction model in reducing trajectory noise and the stability of the method. The approach also includes an LSTM-Hungarian part to make up the missing trajectory. The proposed method is expected to serve as a useful post-processing tool for current CV algorithms.

Low-Rank Hankel Tensor Completion for Traffic Speed Estimation

This paper studies the traffic state estimation (TSE) problem using sparse observations from mobile sensors. Most existing TSE methods either rely on well-defined physical traffic flow models or require large amounts of simulation data as input to train learning algorithms. Different from previous studies, in this paper we propose a purely data-driven and model-free solution. We consider TSE as a spatiotemporal matrix completion/interpolation problem and apply spatiotemporal delay embedding to transform the original incomplete matrix into a fourth-order Hankel structured tensor. By imposing a low-rank assumption on this tensor structure, we can approximate and characterize both global patterns and local patterns in a data-driven manner. We use a truncated nuclear norm of a balanced spatiotemporal unfolding to approximate the tensor rank and develop an efficient algorithm based on the Alternating Direction Method of Multipliers (ADMM) to solve the problem. The proposed framework only involves two hyperparameters, spatial and temporal window lengths, which are easy to set given the degree of data sparsity. To validate the effectiveness of our proposed method, we conducted numerical experiments on real-world high-resolution trajectory data, which demonstrated its superiority in some challenging scenarios. The proposed method shows great potential for solving the TSE problem using sparse observations from mobile sensors and can be applied in various traffic applications.

Analysing freeway diverging risks using high-resolution trajectory data based on conflict prediction models

Abstract This study aims to develop a reliable safety evaluation model for diverging vehicles and investigates the impact of the surrounding traffic environment on freeway diverging risks. High-resolution trajectory data from three sites in the Netherlands (Delft, Ter-Heide and Zonzeel) were employed for the risk analysis. Linear regression (LR), support vector machine (SVM), random forest (RF), extreme randomize trees (ET), adaptive boosting (Adaboost), extreme gradient boosting (XGboost) and multilayer perceptron (MLP), were developed for safety evaluation. The result showed that MLP outperforms the other models for diverging risk prediction over all the indicators, conflict thresholds and locations. Pairwise matrix, shapely addictive explanation (SHAP), and LR algorithms were further adopted to interpret the influence of the surrounding environment. It indicates that an increase in traffic density, subject vehicle (SV) lateral speed, the distance of SV from ramp nose and SV length would increase the diverging risk. At the same time, an increase in leading vehicle speed and space headway would decrease diverging risk. Finally, spatial analysis was also conducted to explore the stability of identified traffic features regarding the impact on the diverging risk across the sites.

Estimation of spatiotemporal travel speed based on probe vehicles in mixed traffic flow

<abstract> <p>Conventional fixed traffic detectors are limited to their installed locations and are unable to collect general traffic information or monitor microscopic traffic flows. Mobile detectors overcome spatial constraints by allowing the vehicle to act as a detector and can observe microscopic traffic flows by collecting high-resolution trajectory data from individual vehicles. The objective of this study is to estimate spatiotemporal traffic information based on the autonomous driving sensor headway distance and to calculate the appropriate spatiotemporal interval according to the sampling rate. First, individual vehicle trajectory data was collected, and a traffic information estimation was established. Travel speed was calculated based on generalized definitions, and its estimation and errors were analyzed. In addition, the appropriate spatiotemporal interval according to cell size, time interval, and sampling rate was analyzed. The analysis demonstrated that the estimation accuracy was improved by cell size, time interval, and sampling rate. Based on this, the appropriate time and space to minimize the error rate were calculated considering the sampling rate. When the sampling rate was 40% or more, the error rate was 5% or less in all time and space; however, error rate differences occurred in several cases at sampling rates below 40%. These results are anticipated for efficient management of collecting, processing and providing traffic information.</p> </abstract>

Machine Learning Framework for Real-Time Assessment of Traffic Safety Utilizing Connected Vehicle Data

Assessment of roadway safety in real-time is a necessary component for providing proactive safety countermeasures to ensure the continued safety and efficiency of roadways. A framework for utilizing data from connected vehicles and other probe sources is proposed in this study. Connected vehicles present an opportunity to provide live fingerprinting and activity monitoring on roadways. Taking advantage of high-resolution trajectory data streaming directly from connected vehicles, variables are extracted and the relationship with crashes are explored utilizing statistical and machine learning models. Hard acceleration events, in conjunction with segment miles are shown to have strong positive correlations with historical crash outcomes as proven by OLS, Poisson and Gradient Booster regression models. An XGBoost classification model is then trained to predict the real-time instances of crash outcomes at 5 min temporal bins with high levels of accuracy when trained with data including the real-time segment speed, reference speed, segment miles, a segment crash risk factor and other variables related to the difference in speeds between consecutive segments as well as the hour of the day. A weighted ensemble model achieved the best performance with an accuracy of 0.95. The results present evidence that the framework can capitalize on the richness of data available via connected vehicles and is implementable as a component in Advanced Traffic Management Systems for the analysis of safety critical situations in real-time.

A Systematic Solution of Human Driving Behavior Modeling and Simulation for Automated Vehicle Studies

Though automated vehicles (AVs) are believed to play a crucial role in future transport, human driving vehicles will share the road with automated vehicles for a relatively long period. So, we need to enable automated vehicles to run along with human drivers especially when they may have conflicts in the right of way. One key problem is how to appropriately model human driving behaviors and quickly simulate their actions when training/testing automated vehicles. Many existing models were originally built for traffic flow studies and may not be suitable for automated vehicles studies. In this paper, we propose a set of new principles of human driving behaviors modeling and simulations. Then, we propose a Data-Driven Simulator (D2Sim) model for human behavior learning, description, and vehicle interaction simulation. In contrast to conventional microscopic traffic flow models, the D2Sim is a trajectory generation model that accepts rich driving environment information (e.g., lane geometry, crosswalks, traffic signals, surrounding vehicles, etc.). Different from many empirical trajectory records replay models, we can arbitrarily set the long-term intentions of the simulated vehicles and intentionally design the corner cases that had not been observed in practice. In addition, the D2Sim adopts adversarial learning to comprehend complex yet stochastic human driving behaviors from empirical data. Testing results show that the proposed model can quickly generate high-resolution trajectory data for training and testing.

Expressway rear-end crash risk evolution mechanism analysis under different traffic states

The proportion of rear-end crashes is the highest for expressways. An effective ways to reduce the rear-end crash risk is Active Traffic Management (ATM), and knowing the mechanism of how contributing factors affect crash risk in space and time is the foundation of ATM. However, the existing studies are mainly based on highly aggregated traffic data. It is hard to capture the evolution mechanisms of crash risk. Meanwhile, crash risk mechanisms might be heterogeneous in smooth and congestion states. This study explored the crash mechanisms in different traffic states with high-resolution trajectory data. First, an ordered clustering method is used to divide a four km expressway section into several segments as the spatial unit. Second, the spatial–temporal ranges are decided by the spatial–temporal correlations analyses between crash risk and potential contributing factors. Thirdly, three types of time-series models are established to quantitatively obtain the impacts of contributing factors on the crash risks. The results showed that crash risk is mainly decided by the upstream contributing factors under smooth states, and determined by the downstream factors under congestion conditions.

An empirical study on fuel consumption of commercial automated vehicles

Increasing commercial vehicles are equipped with automated driving features. Adaptive cruise control, a critical longitudinal control system of commercial automated vehicles (AVs), may have significant impacts on fuel consumption. To investigate the impacts, this paper collected high-resolution trajectory data of commercial AVs with different operating scenarios, speed ranges, and headway settings on the highway system. The AVs’ fuel consumption was calculated by several state-of-the-art or classical vehicle fuel consumption models. From empirical analyses, we found that as the AV headway setting increases, the corresponding fuel consumption decreases. Also, we found that as the speed of AV traffic increases, the impacts of AV headway settings on fuel consumption decrease. Moreover, we compared the fuel consumption of AVs and human-driven vehicles (HVs). We found that for the same settings, the AVs always require less fuel consumption than the HVs. Following these findings, a set of managerial insights were provided into relevant stakeholders for future AV traffic.

Characterizing lane changing behavior and identifying extreme lane changing traits

ABSTRACT This study characterizes lane changing behavior of drivers under differing congestion levels and identifies extreme lane changing traits using high-resolution trajectory data. Total lane change frequency exhibited a reciprocal relationship with congestion level, but the distribution of lane change per vehicle remained unchanged as congestion increased. On average, the speed of trajectories increased by 5.4 ft/s after changing a lane. However, this gain significantly diminished as congestion worsened. Further, the average speed of lane changing vehicles was 3.9 ft/s higher than those that executed no lane changes. Two metrics were employed to identify extreme lane changing behavior: critical time-to-line-crossing (TLCc) and lane changes per unit distance. The lowest 1% TLCc varied between 0.71–1.57 seconds. The highest 1% of lane change rates for all lane changing vehicles was 2.5 lane changes per 1,000 ft traveled. Interestingly, no drivers in thisdataset had both excessive lane changes and lane changes with low TLCc.

Using traffic flow characteristics to predict real-time conflict risk: A novel method for trajectory data analysis

The real-time conflict prediction model using traffic flow characteristics is much less studied than the crash-based model. This study aims at exploring the relationship between conflicts and traffic flow features with the consideration of heterogeneity and developing predictive models to identify conflict-prone conditions in a real-time manner. The high-resolution trajectory data from the HighD dataset is used as empirical data. A novel method with the virtual detector approach for traffic feature extraction and a two-step framework is proposed for the trajectory data analysis. The framework consists of an exploratory study by random parameter logit model with heterogeneity in means and variances and a comparative study on several machine learning methods, including eXtreme Gradient Boosting (Boosting), Random Forest (Bagging), Support Vector Machine (Single-classifier), and Multilayer-Perceptron (Deep neural network). Results indicate that (1) traffic flow characteristics have significant impacts on the probability of conflict occurrence; (2) the statistical model considering mean heterogeneity outperforms the counterpart and lane differences variables are found to significantly impact the means of random parameters for both lane variables and lane differences variables; (3) eXtreme Gradient Boosting trained on an under-sampled dataset turns out to be the best model with the highest AUC of 0.871 and precision of 0.867, showing that re-sampling techniques can significantly improve the model performance. The proposed model is found to be sensitive to the conflict threshold. Sensitivity analysis on feature selection further confirms that the conflict risk prediction should consider both subject lane features and lane difference features, which verifies the consistency with exploratory analysis based on the statistical model. The consistency between statistical models and machine learning methods improves the interpretability of results for the latter one.

A high-resolution trajectory data driven method for real-time evaluation of traffic safety

Real-time safety evaluation is essential for developing proactive safety management strategy and improving the overall traffic safety. This paper proposes a method for real-time evaluation of road safety, in which traffic states and conflicts are combined to explore the internal relationship based on high-resolution trajectory data. In order to assess the real-time traffic safety at a lane level, the trajectory data of the HighD dataset from Germany are utilized to collect lane-based dataset. A surrogate safety measure, time-to-collision (TTC) index, is used for the conflict identification. A binary logistic regression model is employed to quantify the relationship between traffic states and conflicts. Moreover, machine learning methods, including support vector machine, decision tree, random forest, and gradient boosting decision tree, are applied for real-time evaluation. A total of 24 models are trained using the selected four classifier algorithms, and random forest achieves the best performance with 0.85 of the overall accuracy. The results show that the conflict risk can be well estimated by the proposed method. The findings of this study contribute to the high-precision evaluation of real-time traffic safety and the development of proactive safety management.

Constructing a fundamental diagram for traffic flow with automated vehicles: Methodology and demonstration

Increasingly, commercial vehicles are equipped with automated vehicle (AV) features such as adaptive cruise control systems. The AV feature can automatically control the headway between the current vehicle and the preceding vehicle in an adaptive manner. The automatic control may lead to significantly different car- following motions compared with those of human-driven vehicles, which challenges the applicability of classic traffic flow theory to emerging road traffic with AVs. To investigate the impacts of commercial AVs on traffic flow, this paper proposes a general methodology that combines both empirical experiments and theoretical models to construct a fundamental diagram (FD), i.e., the foundation for traffic flow theory for AV traffic. To demonstrate the empirical experiment settings, we collected high-resolution trajectory data with multiple commercial AVs following one another in a platoon with different headway settings. The field experiment results revealed that the traditional triangular FD structure remains applicable to describe the traffic flow characteristics of AV traffic. Further, by comparing the FDs between AVs and human-driven vehicles, it was found that although the shortest AV headway setting can significantly improve road capacity, other headway settings may decrease road capacity compared with existing human-driven-vehicle traffic. It was also found that headway settings may affect the stability of traffic flow, which has been revealed by theoretical studies but was first verified by empirical AV data. With these findings, mixed traffic flow FDs were derived by incorporating different headway settings and AV penetration rates. The method proposed in this paper, including experiment designs, data collection approaches, traffic flow characteristics analyses, and mixed traffic flow FD construction approaches, can serve as a methodological foundation for studying future mixed traffic flow features with uncertain and evolving AV technologies.

Empirical study on car-following characteristics of commercial automated vehicles with different headway settings

A recent study (Li, 2020) analytically predicted tradeoffs between automated vehicle (AV) following characteristics on safety, mobility, and stability using a parsimonious linear car-following model. This work aimed to verify the key theoretical findings in the above study with empirical experiments using commercial AVs, e.g., vehicles with adaptive cruise control (ACC) functions. We collect high-resolution trajectory data of multiple commercial AVs following one another in a platoon with different headway settings. Parsimonious linear AV-following models that capture the first-order parameters on safety, mobility, and stability aspects are estimated with the data. The estimation results of the key parameters validate several theoretical predictions predicted by Li (2020). Specifically, it was found that as the time lag setting increases, the corresponding safety buffer decreases, indicating that AV safety could be improved with less pursuit of AV mobility or, conversely, AV mobility improvement may come at a cost of more stringent safety requirements. Also, as the time lag setting increases, AV string stability increases, indicating that stop-and-go traffic potentially could be dampened by compromising AV mobility. With this, one possible explanation to the observed string instability of commercial AV following control (i.e., ACC function) is that automakers may prefer to ensure a relatively short headway (and thus better user experience on vehicle mobility) at a cost of compromising string stability. It was also found that as the time lag increases, the cycle period of traffic oscillations gets longer, and the oscillation amplification gets smaller, which supports the tradeoff between mobility and stability. On the other hand, field experiments revealed issues beyond the predictivity of a simple linear model. That is, vehicle control sensitivity factors vary across different speed and headway settings, and the model estimation results for key parameters are not consistent over different speed ranges. This opens future research needs for investigating nonlinearity and stochasticity in the AV following modeling.

Longitudinal-Scanline-Based Arterial Traffic Video Analytics with Coordinate Transformation Assisted by 3D Infrastructure Data

High-resolution vehicle trajectory data can be used to generate a wide range of performance measures and facilitate many smart mobility applications for traffic operations and management. In this paper, a Longitudinal Scanline LiDAR-Camera model is explored for trajectory extraction at urban arterial intersections. The proposed model can efficiently detect vehicle trajectories under the complex, noisy conditions (e.g., hanging cables, lane markings, crossing traffic) typical of an arterial intersection environment. Traces within video footage are then converted into trajectories in world coordinates by matching a video image with a 3D LiDAR (Light Detection and Ranging) model through key infrastructure points. Using 3D LiDAR data will significantly improve the camera calibration process for real-world trajectory extraction. The pan-tilt-zoom effects of the traffic camera can be handled automatically by a proposed motion estimation algorithm. The results demonstrate the potential of integrating longitudinal-scanline-based vehicle trajectory detection and the 3D LiDAR point cloud to provide lane-by-lane high-resolution trajectory data. The resulting system has the potential to become a low-cost but reliable measure for future smart mobility systems.

An optimization model for arterial coordination control based on sampled vehicle trajectories: The STREAM model

Traditional arterial coordination control models adopt aggregated demand from fixed detectors or manual counts as main input, which lack a self-feedback mechanism between the input and the optimization objective. With the population of connected vehicles and intelligent mobility, high-resolution trajectory data provide new possibilities for signal control evaluation and optimization. Therefore, the objective of this study is to propose a new arterial coordination control model for two-way arterial progression solely using sampled trajectories. Sampled trajectories are first used to extract prior arrival information and queuing states. Then, vehicle arrival time at the stop-line at each intersection is estimated as the function of signal timing parameters and prior arrival rates based on the shockwave theory, considering different sampled trajectory statuses. The delay of each sample vehicle is thus obtained as the difference between the actual arrival time and the projected arrival time. Sum of the delays of all the sampled trajectories ever travelling on the mainline of the arterial is selected to be the optimization objective, and the optimization problem is solved through a multi-sub-swarm Particle Swarm Optimization (PSO) algorithm. A simulation model is built to test the performance of the proposed model compared with the Synchro model and the MULTIBAND model under various demand scenarios. Results show that the performance of the proposed method is relatively satisfactory, which demonstrates that the optimization of fixed-time arterial coordination control solely using sample trajectories is feasible.

Biking islands in cities: An analysis combining bike trajectory and percolation theory

Bike trajectories generated by the dockless bike-sharing service provides a great opportunity to explore users' travel behavior within the shared mobility transportation ecosystem. This paper proposes a new concept, namely biking islands, defined as geographical areas of interest with a high concentration of bike usage. Leveraging high-resolution trajectory data, biking islands are identified via percolation theory because of its suitability in describing the formation of clusters and critical road segments that have significant influence on biking behavior in an urban context. We showcase our methodology using the bike trajectory dataset provided by a market-leading company and use Shanghai, China as the study area. Results reveal a hierarchical structure of biking islands. With the increase of threshold, the biking islands start to shrink and split into various smaller ones. Larger biking islands are usually located in the central urban area of Shanghai and the Huangpu River acts as a natural barrier that impedes biking continuity across the region. Besides, the formation of biking islands is highly influenced by the surrounding land uses. The proposed concept and methods are not only helpful to understand travel behavior of cyclists and urban structures used for cycling, but also has the potential to support relevant urban and transportation planning, such as identifying designated non-motorized areas for cycling and biking facilities and pinpointing critical road segments that can improve the cycling efficiency of the entire network. Biking islands could be designed as designated areas for cycling where sufficient bike facilities are provided, and/or motorized transport modes are restricted or even prohibited, so as to ensure the convenience and safety of cyclists and support the development of bike-friendly cities.