Microscopic Traffic Data Research Articles

A Microscopic Traffic Flow Data Generation Method Based on an Improved DCGAN

Microscopic traffic flow data, an important input to virtual test scenarios for autonomous driving, are often difficult to obtain in large quantities to allow for batch testing. In this paper, a neural network for generating microscopic traffic flow scene fragments is proposed, which is improved by adding Gate Recurrent Units (GRU) to the discriminator of the Deep Convolutional Generative Adversarial Network (DCGAN) to enable it to better discriminate continuous data. Subsequently, this paper compares individual sample motion trajectories of the generated data using Grey Relational Analysis (GRA) and Dynamic Time Warping algorithm (DTW) at the microscopic scale, and evaluates the overall scenes generated using averaged statistics at the macroscopic scale. The results show that the method proposed in this paper can generate realistic microscopic traffic flow data very well and that the neural network proposed in this paper can generate better near-realistic microscopic traffic flow data than the original DCGAN under the evaluation metrics used in this paper.

A revise on using surrogate safety measures for rear-end crashes

Traffic safety is among the most serious topics of transportation engineering studies. Different methods have been proposed for safety evaluations. Applying surrogate safety measures (SSM) to predict the crash likelihood and severity is a relatively new approach, which needs microscopic traffic data instead of crash data. Numerous SSMs have been introduced in the literature. Now, it is somehow hard to decide about the selection of suitable SSMs for safety assessments based on microscopic traffic data. In this paper, first a review on SSMs relating to rear-end conflicts introduced in the high-quality papers is done. Then, with the help of microscopic traffic data of NGSIM in I-80 freeway, SSMs are calculated for each car-following scenario. Because of the diversity of SSMs and contradicting results it was tried to propose a method for combining the results of these indicators to report one number as the rear-end collision risk. The efficiency of the new indicator is evaluated in comparison to individual SSMs. The proposed methodology can be beneficial to improve the efficiency of In-vehicle collision avoidance warning system (IVCAWS) and enhancing highway safety evaluations based on SSMs.

Understanding the mechanism of lane changing process and dynamics using microscopic traffic data

The lane changing manoeuvre is a fundamental driver behaviour that determines vehicle distribution across lanes. Lane-changing behaviour has a significant effect on traffic flow and may cause traffic oscillations, relaxation, moving bottleneck, and capacity drop/breakdown. This study aims to identify the lane change window and classify the lane change behaviour of vehicles using NGSIM data. Determination of the starting and ending points (time window) of lane change operation has a significant impact on the lane change duration and dynamics. In this study, a new approach is proposed to identify the time window of lane change operation using the absolute value of the derivative of the cumulative lateral speed which reduces unintended drifts of vehicles in lateral direction. Then, the lane changes are classified into free, forced and cooperative using Hidas definition (Hidas, 2005) which does not closely replicate the field conditions. This study proposes three different methods for the classification of lane changes based on microscopic traffic variables and vehicle kinematics in the vicinity of lane changing process and these methods are verified quantitatively. Statistical tests are performed to check if the types of lane changes are statistically different. Finally, two types of classification are considered such as free lane change and constraint lane change, and a log-normal distribution is fitted for the lane change duration of these two types.

Transfer learning for spatio-temporal transferability of real-time crash prediction models

Real-time crash prediction is a heavily studied area given their potential applications in proactive traffic safety management in which a plethora of statistical and machine learning (ML) models have been developed to predict traffic crashes in real-time. However, one of the fundamental issues relating to the application of these models is spatio-temporal transferability. The present paper attempts to address this gap of knowledge by combining Generative Adversarial Network (GAN) and transfer learning to examine the transferability of real-time crash prediction models under an extremely imbalanced data setting. Initially, a baseline model was developed using Deep Neural Network (DNN) with crash and microscopic traffic data collected from M1 Motorway in the UK in 2017. The dataset utilised in the baseline model is naturally imbalanced with 257 crash cases and 16,359,163 non-crash cases. To overcome data imbalance issue, Wasserstein GAN (WGAN) was utilised to generate synthetic crash data. Non-crash data were randomly undersampled due to computational limitations. The calibrated model was then applied to predict traffic crashes for five other datasets obtained from M1 (2018), M4 (2017 & 2018 separately) and M6 Motorway (2017 & 2018 separately) by using transfer learning. Model transferability was compared with standalone models and direct transfer from the baseline model. The study revealed that direct transfer is not feasible. However, models become transferable temporally, spatially, and spatio-temporally if transfer learning is applied. The predictability of the transferred models outperformed existing studies by achieving high Area Under Curve (AUC) values ranging between 0.69 and 0.95. The best transferred model can predict nearly 95% crashes with only a 5% false alarm rate by tuning thresholds. Furthermore, the performances of transferred models are on par with or better than the standalone model. The findings of this study proves that transfer learning can improve model transferability under extremely imbalanced settings which helps traffic engineers in developing highly transferable models in future.

Utilizing UAVs Technology on Microscopic Traffic Naturalistic Data Acquirement

Research into collecting and measuring reliable, accurate, and naturalistic microscopic traffic data is a fundamental aspect in road network planning scientific literature. The vehicle trajectory is one of the main variables in traffic flow theory that allows to extract information regarding microscopic traffic flow characteristics. Several methods and techniques have been applied regarding the acquisition of vehicle trajectory. The forthcoming applications of intelligent transport systems on vehicles and infrastructure require sufficient and innovative tools to calibrate existing models on more complex situations. Unmanned aerial vehicles (UAVs) are one of the most emerging technologies being used recently in the transportation field to monitor and analyze the traffic flow. The aim of this paper is to examine the use of UAVs as a tool for microscopic traffic data collection and analysis. A comprehensive guiding framework for accurate and cost-effective naturalistic traffic surveys and analysis using UAVs is proposed and presented in detail. Field experiments of acquiring vehicle trajectories on two multilane roundabouts were carried out following the proposed framework. Results of the experiment indicate the usefulness of the UAVs technology on various traffic analysis studies. The results of this study provide a practical guide regarding vehicle trajectory acquirement using UAVs.

Developing safety performance functions for freeways at different aggregation levels using multi-state microscopic traffic detector data

Safety Performance Functions (SPFs) have been widely used by researchers and practitioners to conduct roadway safety evaluation. Traditional SPFs are usually developed by using annual average daily traffic (AADT) along with geometric characteristics. However, the high level of aggregation may lead to a failure to capture the temporal variation in traffic characteristics (e.g., traffic volume and speed) and crash frequencies. In this study, SPFs at different aggregation levels were developed based on microscopic traffic detector data from California, Florida, and Virginia. More specifically, five aggregation levels were considered: (1) annual average weekday hourly traffic (AAWDHT), (2) annual average weekend hourly traffic (AAWEHT), (3) annual average weekday peak/off-peak traffic (AAWDPT), (4) annual average day of the week traffic (AADOWT), and (5) annual average daily traffic (AADT). Model estimation results showed that the segment length and volume, as exposure variables, are significant across all the aggregation levels. Average speed is significant with a negative coefficient, and the standard deviation of speed was found to be positively associated with the crash frequency. It is noteworthy that the operation of the high occupancy vehicle (HOV) lanes was found to have a positive effect on crash frequency across all the aggregation levels. The model results also showed that the AAWDPT and AADOWT models consistently performed better (the improvements range from 3.14%–16.20%) than the AADT-based SPF, which implies that the differences between the day of the week and peak/off-peak periods should be considered in the development of crash prediction models. The model transferability results indicated that the SPFs between Florida and Virginia are transferrable, while the models between California and the other two states are not transferrable.

Introducing novel statistical-based method of screening and combining currently well-known surrogate safety measures

ABSTRACT Until now various surrogate safety measures (SSM) have been developed and research in this area is ongoing. However, it is unclear which indicators are able to reflect danger more accurately. This paper proposes a novel method of performance comparison by taking the advantage of statistical analyses of SSMs data. This analytical approach is inspired by the dominant behavior of drivers in car-following scenarios. Afterward, those indicators that previously identified as eligible are merged using a recently developed safety indicator, named collision probability (CP). The proposed method is applied to microscopic traffic data collected from the I-80 freeway in Next Generation Simulation (NGSIM) project. The results indicated that the proposed framework helps screen SSMs and calculate the probability of rear-end collisions using a combination of the selected SSMs. The results might be helpful for improving the precision and accuracy of in-vehicle collision avoidance warning system and safety evaluations without crash data.

Crash Risk Assessment for Heterogeneity Traffic and Different Vehicle-Following Patterns Using Microscopic Traffic Flow Data

This paper investigates the impacts of heavy vehicles (HV) on speed variation and assesses the rear-end crash risk for four vehicle-following patterns in a heterogeneous traffic flow condition using three surrogate safety measures: speed variation, time-to-collision (TTC), and deceleration rate to avoid a crash (DRAC). A video-based data collection approach was employed to collect the speed of each individual vehicle and vehicle-following headway; a total of 3859 vehicle-following pairs were identified. Binary logistic regression modeling was employed to assess the impacts of HV percentage on crash risk. TTCs and DRACs were calculated based on the collected traffic flow data. Analytical models were developed to estimate the minimum safe vehicle-following headways for the four vehicle-following patterns. Field data revealed that the variation of speed first increased with HV percentage and reached the maximum when HV percentage was at around 0.35; then, it displayed a decreasing trend with HV percentage. Binary logistic regression modeling results suggest that a high risk of rear-end collision is expected when HV percentage is between 0.19 and 0.5; while, when HV percentage is either below 0.19 or exceed 0.5, a low risk of rear-end collision is anticipated. Analytical modeling results show that the passenger car (PC)-HV vehicle-following pattern requires the largest minimum safe space headway, followed by HV-HV, PC-PC, and HV-PC vehicle-following patterns. Findings from this research present insights to transportation engineers regarding the development of crash mitigation strategies and have the potential to advance the design of real-time in-vehicle forward collision warnings to minimize the risk of rear-end crash.

Understanding the effects of vehicle platoons on crash type and severity

Crash type is an informative indicator to infer driving behaviors and conditions that cause a crash. For example, rear-end and side-swipe crashes are typically caused by improper vehicle interaction such as sudden lane-changing or speed control while hit-object crashes are likely the result of a single driver’s mistake. This study investigated the impact of vehicles travelling as a group (platoon) and its configuration (i.e., types of vehicles consisting of the platoon) on crash type and severity since the vehicles could affect each other when travelling in close proximity. This study applied Generalized Structure Equation Modeling (GSEM) to capture the complex relationships among the various crash factors such as traffic condition, driver characteristics, environmental conditions, and vehicle interaction to the crash attributes including type and severity.This study collected over 3 million individual vehicle data from 39 traffic count sites in California to estimate the vehicle interactions and driving behaviors. The microscopic traffic data are matched to 1417 crash reports. Results showed that vehicles traveling in platoons are associated with more rear-end and side-swipe crashes. Speed difference in the platoon had a positive effect on hit-object crashes if the platoon comprises vehicles of homogeneous type – either trucks or non-trucks. In addition, human factors such as age and gender were identified as significant influential factors in all type of crashes, however truck involvement particularly played an important role amongst side-swipe crashes. Crash severity was negatively affected by total flow, and rear-end crashes were more likely to be severe compared with hit-object crashes. Based on findings, this study suggests practical operational strategies to reduce traffic instability associated with platooned vehicle patterns. Understanding the high-risk factors for different crash types and severities would provide valuable insights for decision-makers and transportation engineers to develop targeted intervention strategies in consideration of road users and traffic conditions such as fleet mix and speed.

Challenges faced in heterogeneous traffic data collection: a comparison of traffic data collection technologies

Traffic data are the fundamental inputs to traffic flow analysis and simulation studies, which facilitate decision making in the fi eld of traffic engineering. Hence, the accuracy of traffic data is of paramount importance. This study compares new technologies available for traffic data collection considering their accuracy and applicability in the Sri Lankan context. Traffic in Sri Lanka is of heterogeneous nature, as opposed to the homogeneous nature observed in most developed countries. Hence, collection of traffic data poses several challenges that affects its accuracy. Three techniques, the infrared driven TIRTL instrument, the video image processing-based TRAZER application and the traffic data collection method using the Google distance matrix application programming interface (API), with respect to their data collection accuracy are reviewed in this study. The fundamental macroscopic traffic data variables (fl ow and speed) were evaluated against control surveys. It was found that each technology has its strengths and weaknesses and needs to be used appropriately. The TIRTL instrument fared better on road sections on level terrain when the crossfall did not obstruct the infrared beams. Such occasions provided a rich set of microscopic traffic data. The TRAZER software delivered data up to a 100 % accuracy. However, this required the user to go through a lengthy postprocessing routine to extract the final set of traffic data. Google traffic data collection provides highly accurate results when estimating link speeds. This method is ideal for collection of bulk data with spatio-temporal variations and the process can be fully automated to reduce the human resource requirement.

Combining shockwave analysis and Bayesian Network for traffic parameter estimation at signalized intersections considering queue spillback

This paper focuses on traffic parameters estimation at signalized intersections based on a framework combining shockwave analysis (SA) and Bayesian Network (BN) using vehicle trajectory data. Detailed queuing evolution and spillback across adjacent intersections are considered. According to shockwave analysis, the analytical probability distribution of individual vehicle’s travel time is derived based on different initial conditions. This probability distribution is parameterized by the fundamental diagram (FD) parameters, traffic volume, and cycle state (queue length). A three-layer recursive BN model is then proposed to construct the state evolution process as well as the relationships between traffic volume, cycle state, FD parameters, sampled vehicles’ arrival times and intersection travel times. As traffic volume and initial queue cannot be measured directly from sampled trajectory data, the expectation maximization (EM) algorithm and particle filtering (PF) are introduced to solve this recursive BN model. By shockwave analysis, such estimated traffic parameters are then used to estimate the maximum queue length and traffic volume of each cycle. The proposed method is evaluated using microscopic traffic simulation data as well as empirical data. Numerical results show that the proposed method achieves promising accuracy even under low penetration rates, with the mean absolute percentage error (MAPE) of the estimation bounded by 15% and generally around 10%.

Calculating dynamic thresholds for critical time to collision as a safety measure

Surrogate safety measures have been recognised as suitable tools for a warning strategy. As one such measure, the time to collision (TTC) is the time remaining to a collision if the collision course and speed remain unaltered. To apply the TTC in discriminating dangerous situations, a critical threshold (TTC*) must be determined. However, a method for calculating this threshold is yet to be presented. Previously, the critical threshold has been considered as a constant value, but the value of TTC* changes the calculated probability for a crash at each time instant. In this work, a method was developed for calculating TTC* based on driver characteristics, environmental conditions, the type of preceding object and the microscopic traffic parameters of the subject vehicle based on an adaptive neuro fuzzy inference system and motion mechanics. For this purpose, data were first collected from a driving simulator. Then, to compare the collision probability calculated based on a dynamic TTC* (DTTC*) and a static TTC* (STTC*), microscopic traffic data were obtained from Modares highway in Tehran, Iran. Assessments indicated that, statistically, there is a significant difference between DTTC* and STTC* results. This finding might help to enhance the capability of in-vehicle collision-avoidance systems to prevent rear-end collisions.

Extraction and analysis of microscopic traffic data in disordered heterogeneous traffic conditions

ABSTRACT This paper discusses the development of a novel off-line vision-based system to obtain the naturalistic trajectory database of traffic streams from recorded video footages. The developed system uses a semi-automatic mechanism that provides manual interference for vehicle identification and classification and executes automated tracking of the identified vehicles. A trajectory database of typical disordered heterogeneous traffic stream was collected to evaluate the performance of the developed system. Results show that the developed system significantly enhances the process of trajectory data collection in such traffic conditions. The collected trajectory database is then used to investigate two crucial aspects that characterize the disordered heterogeneous traffic (i) interaction of different types of vehicles in the longitudinal and staggered following scenario, and (ii) lateral shift propensity of different types of vehicles. The analysis emphasizes the behavioral difference between different types of vehicles, and utility of the developed system to address the prevailing research gaps.

An evaluation of time-to-collision as a surrogate safety measure and a proposal of a new method for its application in safety analysis

ABSTRACT Various surrogate safety measures have been introduced, of which Time-to-Collision (TTC) is the most well-known and commonly applied indicator. TTC first was defined as the time remaining prior to a collision if the path and speed of vehicles are maintained as a constant. Various modifications have been applied to this indicator to increase the efficiency of TTC. However, there is still no specific platform that shows how all of these modifications can be applied simultaneously. Thus, this paper first reviews the modifications that have been made to TTC; subsequently, a comprehensive framework is proposed for the application of this indicator for different types of conflicts, while considering all previous modifications. The framework is then applied for microscopic traffic data collected in Tehran. Results indicate that there is a great difference between the results of different versions of TTC for collision probability prediction. However, the proposed procedure might be beneficial in increasing the precision and accuracy of detecting dangerous encounters.

Spatio-Temporal Synchronization of Cross Section Based Sensors for High Precision Microscopic Traffic Data Reconstruction

The next generation of Intelligent Transportation Systems (ITS) will strongly rely on a high level of detail and coverage in traffic data acquisition. Beyond aggregated traffic parameters like the flux, mean speed, and density used in macroscopic traffic analysis, a continuous location estimation of individual vehicles on a microscopic scale will be required. On the infrastructure side, several sensor techniques exist today that are able to record the data of individual vehicles at a cross-section, such as static radar detectors, laser scanners, or computer vision systems. In order to record the position data of individual vehicles over longer sections, the use of multiple sensors along the road with suitable synchronization and data fusion methods could be adopted. This paper presents appropriate methods considering realistic scale and accuracy conditions of the original data acquisition. Datasets consisting of a timestamp and a speed for each individual vehicle are used as input data. As a first step, a closed formulation for a sensor offset estimation algorithm with simultaneous vehicle registration is presented. Based on this initial step, the datasets are fused to reconstruct microscopic traffic data using quintic Beziér curves. With the derived trajectories, the dependency of the results on the accuracy of the individual sensors is thoroughly investigated. This method enhances the usability of common cross-section-based sensors by enabling the deriving of non-linear vehicle trajectories without the necessity of precise prior synchronization.

Modified Traffic Flow Model with Connected Vehicle Microscopic Data for Proactive Variable Speed Limit Control

Most previous prediction based Variable Speed Limit (VSL) control strategies focused on improving traffic mobility based on the macroscopic traffic data. Nowadays, the emerging technologies provide access to the microscopic traffic flow data, which better captures the details of traffic flow dynamics in the VSL controlled environment. Thus, in this paper, the microscopic traffic flow data were utilized as a supplement to predict the evolutions of traffic flow parameters. The proposed VSL control algorithm adopts the Model Predictive Control (MPC) framework, which employs a modified version of the classic traffic flow model METANET to take advantage of the microscopic data in traffic flow predictions. The microscopic traffic simulation software VISSIM was used to establish an experimental simulation platform and perform real time traffic responsive control based on field data. The proposed control strategy was evaluated against the no-VSL control and macroscopic-based VSL controlled scenario. The results show that utilizing the proposed modified METANET model reduced the error in speed prediction accuracy and improved system mobility performance.

Traffic parameters estimation for signalized intersections based on combined shockwave analysis and Bayesian Network

This paper proposes a framework to combine shockwave analysis (SA) with Bayesian Network (BN) for traffic flow parameters estimation at signalized intersections using vehicle trajectory data. Traffic SA is applied as the basis to derive the analytical expression of individual vehicle trajectory given the fundamental diagram (FD) parameters (e.g., capacity, jam density, free flow speed) as well as the traffic volume. According to SA, the analytical probability distribution of each vehicle’s travel time (defined as the time of a vehicle travelling between an upstream and a downstream location at a signalized intersection) is derived. This probability distribution is parameterized by traffic volume and FD parameters. We then apply a three-layer BN model to construct the relationship between traffic volume, FD parameters, sampled vehicles’ arrival times and intersection travel times. Since traffic volume cannot be measured directly from sampled vehicle trajectories, an expectation maximization (EM) algorithm is introduced to perform learning for estimating the cycle-by-cycle volume (or arrivals) and FD parameters. Such estimated parameters are then used, via SA, to estimate vehicles’ intersection travel times. The proposed combined SA-BN model is evaluated using microscopic traffic simulation data and the NGSIM dataset. Sensitivity analysis of the estimation performance with respect to penetration rates is conducted. In general, the proposed method yields promising estimation accuracy, with the mean absolute percentage error (MAPE) of the estimation bounded by 15%, generally around 5%, as compared with the true value, even under low penetration rates.

Increasing the efficiency of vehicle ad-hoc network to enhance the safety status of highways by artificial neural network and fuzzy inference system

The future of traffic control and management may manifest itself in the form of digital and automatic methods, which could help to decrease driver errors. One of the most important technologies in this area is the Vehicle Ad-Hoc Network (VANET), in which each vehicle is a node that can receive and transmit information from and to other vehicles. Such systems can be used for safety or nonsafety purposes. This study investigates the safety applications of this system and aims to develop a method to detect and record dangerous situations for each vehicle based on microscopic traffic data. To detect danger, VANET must transmit data between vehicles and traffic safety indicators must be applied. To increase the efficiency of VANET, artificial neural network (ANN) and fuzzy inference system (FIS) were used. This study focuses on rear-end collisions in car-following situations. For this purpose, microscopic traffic data were collected along a 400-meter long segment of the Modares highway in Tehran, Iran. Based on the analysis, applying FIS to develop a new safety index was found to help detect rear-end collisions. Results also indicate that the average mean of errors based on the motion predictions generated by ANN is negligible and that an adequate history of motion for a vehicle diminishes such errors. Therefore, applying artificial intelligence can improve the workability of VANET in detecting a dangerous car-following situation that might lead to a rear-end collision.

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.

Theory of safety surrogates using vehicle trajectories in macroscopic and microscopic settings: Application to dynamic message signs controlled traffic at work zones

This paper presents a new mathematical framework for obtaining quantitative safety measure using macroscopic as well as microscopic traffic data. The safety surrogate obtained from the macroscopic data is in terms of analysis performed on vehicle trajectories obtained from the macroscopic data. This method of obtaining safety measure can be used for many different types of applications. The safety surrogate for the traffic dynamics are developed in terms of a new concept of Negative Speed Differentials (NSD) that involve a convolution of vehicle speed function obtained from vehicle trajectories and then performing the integration of the square of the output for its negative values. The framework is applicable to microscopic traffic dynamics as well where we can use car following models for microscopic dynamics or the LWR model for macroscopic dynamics. This paper then presents the use of this new safety surrogate on the development of a feedback control law for controlling traffic in work zones using Dynamic Message Signs. A hybrid dynamics model is used to represent the switching dynamics due to changing DMS messages. A feedback control design for choosing those messages is presented as well as a simple simulation example to show its application.