YZYanlin Zhang

Yanlin Zhang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Yanlin Zhang has a total of 21 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Quantifying Distance-Dependent Measurement Errors in Vehicular Trajectory Data Collection

    by Nachuan Li, Yanlin Zhang, Pengyuan Liu, Dana Monzer, Sungyong Chung, Chun-Chien Hsiao, Alireza Talebpour, Marco Nie

    Objectives: Trajectory datasets collected by the onboard sensors of automated vehicles (AVs) are an increasingly common resource for traffic research. Unlike infrastructure-based sensors, an AV observes surrounding vehicles from a continuously changing vantage point, so the accuracy of a measurement depends on how far away its target is. This study asks whether such distance-dependent errors are consequential for downstream traffic analyses. Methods: We develop a general, sensor-agnostic error model that separates two sources of uncertainty conflated in AV-based sensing---the vehicle's error in locating itself and its error in ranging to a target---and represents distance-dependent bias, distance-dependent noise, and temporal correlation in each. We calibrate the model through a controlled two-vehicle field experiment with GPS and radar, then inject the resulting error structure, together with a broader family of hypothetical structures, into the Third Generation Simulation (TGSIM) I-90/94 dataset treated as a proxy for ground truth. Comparing analyses of the error-injected trajectories against the proxy isolates the effect of measurement error on static properties (gap sizes), macroscopic relationships (speed-density-flow), and trajectory-level metrics (speed variance and headway). Findings: Distance-dependent bias displaces gap measurements with or without AV involvement by different amounts, because the two are reconstructed from different combinations of range observations. Under the empirically calibrated error structure, this displacement is sufficient to reverse a statistical conclusion: measured data may indicate that human drivers follow AVs more closely than they follow other human-driven vehicles (HVs), which the ground-truth trajectories do not support. At the macroscopic level, bias produces regime-dependent distortions in flow and density that propagate into capacity assessment. At the trajectory level, it introduces location-dependent biases in speed variance and average headway, so that a vehicle's observed driving behavior depends in part on where it sat relative to the sensor. Unbiased distance-dependent uncertainty is largely attenuated by trajectory smoothing, though it still biases speed variance, since smoothing effectiveness itself varies with the local noise magnitude. Novelty: This study is among the first to systematically quantify how distance-dependent sensor error structures in AV-collected trajectory data bias traffic analysis at both microscopic and macroscopic levels, and to separate the consequences of systematic bias from those of random uncertainty---a distinction that matters because aggregation removes the latter but not the former. Practical Applications: The findings highlight the importance of reporting distance-specific sensor error characteristics, rather than just aggregate accuracy measures, when collecting, processing, and analyzing AV-collected trajectory datasets. Our framework also provides practical guidance for trajectory data quality assessment.

  • Investigation of Discretionary Lane-Changing Decisions: Insights From the Third Generation Simulation (TGSIM) Dataset

    by Yanlin Zhang, Alireza Talebpour, Hani S. Mahmassani, Samer H. Hamdar

    The data-driven characterization of discretionary lane-changing behaviors has traditionally been hindered by the scarcity of high-resolution data that can precisely record lateral movements. In this study, we conducted an exploratory investigation leveraging the Third Generation Simulation (TGSIM) dataset to advance our understanding of discretionary lane-changing behaviors. In this paper, we developed a discretionary lane-changing extraction pipeline and scrutinized crucial factors such as gaps and relative speeds in leading and following directions. A dynamic time warping (DTW) analysis was performed to quantify the difference between any pair of lane-changing behaviors, and an affinity propagation (AP) clustering, evaluated on normalized DTW distance, was conducted. Our results yielded five clusters based on lead and lag gaps, enabling us to categorize lane-changing behaviors into aggressive, neutral, and cautious for both leading and following directions. Clustering based on relative speeds revealed two distinct groups of lane-changing behaviors, one representing overtaking and the other indicative of transitioning into a lane with stable and homogenous speed. The proposed DTW analysis, in conjunction with AP clustering, demonstrated promising potential in categorizing and characterizing lane-changing behaviors. Additionally, this approach can be readily adapted to analyze any driving behavior.


  • Advancing Self-Enforcing Streets Phase 1: The Relationship between Roadway Environment and Crash Severity

    by University of Illinois Urbana-Champaign, Alireza Talebpour, Yanlin Zhang, Tina Radvand, Mahdi Yousefi

    The concept of self-enforcing roadways (SER) has been proposed as a speed management strategy to achieve harmony in drivers’ speed selection and to address safety issues associated with excessive speeding. While numerous examples of SER implementation exist outside of the United States (particularly in European countries), this concept is fairly new in the United States. Only a few states in the U.S. have explored the possibility of implementing SER and evaluated its potential impacts on roadway safety. Accordingly, this study aims at (1) providing a comprehensive review of the literature on SER implementation and impact assessment across the world and (2) examining how roadway elements influence operating speeds and, consequently, crash outcomes, in Illinois, shedding light on the potential impact of roadway design and features on crash severity. A specific before-and-after study is required to analyze carefully the impacts of various SER strategies on safety in the state of Illinois.


  • Close Look into the Spatial-Temporal Distribution of Speed, Lane Changes, and Heavy Vehicles in a Congested Freeway Weaving Section

    by Nachuan Li, Hani S. Mahmassani, Yanlin Zhang, Alireza Talebpour, Samer Hamdar

    Traffic behavior around major freeway weaving sections exhibits complex dynamics associated with multi-directional maneuvers such as lane changes (LCs) accompanied by shockwave-generating braking. Data to study both microscopic and macroscopic properties of congested weaving sections have been generally lacking, leaving an important lacuna in the underlying traffic science. The Third Generation Simulation (TGSIM) trajectory data set collected on multiple freeway locations in the USA provides a rich opportunity to examine the phenomena associated with high-density weaving operations on freeways. The focus of this paper is to examine the spatial-temporal distribution of average speed, LCs, and heavy vehicles (HVs). In addition, we examine the time-shifted association of speed with LCs and HVs. Our analysis reveals considerable variation of speed across lanes and longitudinal locations. LCs are generally associated with higher speeds of the surrounding traffic and correlate with the speed changes on the original and target lanes differently. In addition, differences of speed change have been found for vehicles that execute mandatory LCs (MLCs) and discretionary LCs (DLCs). Finally, while a lower average speed is associated with the existence of HVs, it tends to recover gradually when the HVs move downstream.


  • Introduction to the Third Generation Simulation Dataset: Data Collection and Trajectory Extraction

    by Rami Ammourah, Pedram Beigi, Bingyi Fan, Samer H. Hamdar, John Hourdos, Chun-Chien Hsiao, Rachel James, Mohammdreza Khajeh-Hosseini, Hani S. Mahmassani, Dana Monzer, et al.

    This study aims to provide accurate trajectory datasets capable of characterizing human–automated vehicle interactions under a diverse set of scenarios in diverse highway environments. Distinct methods were utilized to collect data from Level 1, Level 2, and Level 3 automated vehicles: (1) fixed location aerial videography (a helicopter hovers over a segment of interest); (2) moving aerial videography (a helicopter follows the automated vehicles as they move in a much longer segment than in the first method); and (3) infrastructure-based videography (multiple overlapping cameras located on overpasses creating a comprehensive image of the study area). Utilizing the fixed location aerial videography approach, trajectories were extracted on I-90/I-94 in Chicago, IL. The moving aerial videography approach was adopted to extract four datasets on I-90/I-94 and I-294 in Chicago, IL. Finally, two datasets were collected on I-395 and George Washington University Campus in Washington, D.C., using the infrastructure-based videography approach. Extracting multiple complete and accurate vehicle trajectories raises a set of methodological and practical challenges that vary across the three data measurement approaches. The methodological details to extract these trajectories are presented in this paper along with the lessons learned with respect to data collection setup, instrumentation, and experimental design efforts.


  • Characterizing Human–Automated Vehicle Interactions: An Investigation into Car-Following Behavior

    by Yanlin Zhang, Alireza Talebpour

    Automated vehicles are expected to influence human drivers’ behavior. Accordingly, capturing such changes is critical for planning and operation purposes. With regard to car-following behavior, a key question is whether existing car-following models can replicate these changes in human behavior. Using a data set that was collected from the car-following behavior of human drivers when following automated vehicles, this paper offers a robust methodology based on the concept of dynamic time warping to investigate the critical parameters that can be used to capture changes in human behavior. The results indicate that spacing can best substantiate such changes. Moreover, calibration and validation of the intelligent driver model (IDM) suggest its inability to capture changes in human behavior in response to automated vehicles. Thus, an extension of the IDM that explicitly models stochasticity in the behavior of individual drivers is applied, and the results show such a model can identify a reduction in uncertainty when following an automated vehicle. This finding also has implications for a stochastic extension to other models when analyzing and simulating a mixed-autonomy traffic flow environment.