arXiv subjects

Shaurya Agarwal

Publications and source records attributed to Shaurya Agarwal.

At least 19 recordsLinked to original sources

arXiv2026

Trust as a Field: A Macroscopic Representation for Vehicular Networks

Trust assessment is a fundamental component of cooperative and connected vehicle systems. However, existing approaches operate primarily at the level of individual vehicles, making it difficult to reason about trust evolution across road segments. In this paper, we propose a spatio-temporal trust-field framework that aggregates microscopic vehicle-level trust into a continuous representation over space and time. The trust field is formally defined on road segments. We conducted simulation-based experiments using synthetic trajectories generated under controlled conditions, enabling analysis of trust-field behavior in simple road scenarios. Beyond theoretical modeling, we study an implication of the trust-field concept: reconstructing the full trust field from sparse roadside-unit (RSU) measurements. We compare (i) a coordinate-based deep learning baseline that learns a generic trust field from sparse samples and (ii) a field-informed deep learning method that treats trust as a latent quantity carried by vehicles and enforces measurement consistency through the aggregation mechanism. The field-informed approach more accurately recovers trajectory-aligned low-trust patterns and yields improved reconstruction error.

cs.RO↗

arXiv2026

Camera and LiDAR BEV Fusion for Cooperative 3D Object Detection on TUMTraf V2X

We describe a Camera and LiDAR fusion detector developed for the TUMTraf V2X cooperative 3D object detection track of the DriveX 2026 challenge. The detector fuses three roadside cameras with a fused infrastructure-plus-vehicle point cloud in a shared bird's-eye-view space and predicts boxes through a CenterPoint-style head with a generalized IoU regression loss and an IoU quality re-ranking head. Trained on the provided train and validation splits, the model reaches a 3D mAP of 0.85 on the public Codabench test split. While iterating on the system, we observed that 44 of the 50 test frames are also present in the released train (40) and validation (4) splits with their labels. We therefore conducted two additional studies to quantify how this overlap affects the final score: (1) a finetuning run that oversamples the 44 overlapping frames, reaching 0.89 mAP, and (2) a post-processing run that replaces predictions on those frames with the released ground truth, reaching 0.99 mAP (uploaded to our Codabench account for testing but not published on the leaderboard). All three configurations and their per-class results are reported.

cs.CV↗

arXiv2026

Physics-Informed Teacher-Student Ensemble Learning for Traffic State Estimation with a Varying Speed Limit Scenario

Physics-informed deep learning (PIDL) neural networks have shown their capability as a useful instrument for transportation practitioners in utilizing the underlying relationship between the state variables for traffic state estimation (TSE). Another efficient traffic management approach is implementing varying speed limits (VSLs) on transportation corridors to control traffic and mitigate congestion. However, the existing training architecture of PIDL in the literature cannot accommodate the changing traffic characteristics on a freeway with VSL. To tackle this challenge, we propose a novel framework integrating teacher-student ensemble training with PIDL neural networks for TSE under VSL scenarios. The physics of flow conservation law is encoded locally in the teacher models by PIDL, and the student model uses a multi-layer perceptron classifier (MLP) to identify traffic characteristics and selects the ensemble member of PIDL neural networks for TSE. This integrated framework provides a natural solution for capturing the heterogeneity of VSL and accurately addressing the TSE problem. The case study results validate the proposed ensemble approach, demonstrating its superior performance in TSE compared to other popular baseline methods, as indicated by relative L2 error.

cs.LG↗

arXiv2026

Time-varying System Identification of Bedform Dynamics Using Modal Decomposition

Measuring sediment transport in riverbeds has long been a challenging research problem in geomorphology and river engineering. Traditional approaches rely on direct measurements using sediment samplers. Although such measurements are often considered ground truth, they are intrusive, labor-intensive, and prone to large variability. As an alternative, sediment flux can be inferred indirectly from the kinematics of migrating bedforms and temporal changes in bathymetry. While such approaches are helpful, bedform dynamics are nonlinear and multiscale, making it difficult to determine the contributions of different scales to the overall sediment flux. Fourier decomposition has been applied to examine bedform scaling, but it treats spatial and temporal variability separately. In this work, we introduce Dynamic Mode Decomposition (DMD) as a data-driven framework for analyzing riverbed evolution. By incorporating this representation into the Exner equation, we establish a link between modal dynamics and net sediment flux. This formulation provides a surrogate measure for scale-dependent sediment transport, enabling new insights into multiscale bedform-driven sediment flux in fluvial channels.

eess.SY↗

arXiv2026

Spatial-Temporal Nonlocal Traffic Dynamics: Analytical Properties, Adaptive Kernel Formulation, and Empirical Validation

This paper presents a new spatial-temporal nonlocal traffic flow model formulated to overcome the boundedness limitations inherent in classical local formulations. The model introduces an adaptive kernel that captures both spatial and temporal nonlocal interactions, allowing the velocity at a given point to depend on aggregated downstream traffic conditions over a finite time horizon. This structure provides a more realistic representation of driver anticipation and reaction behavior. In addition to developing the model, we establish several key analytical properties that clarify the theoretical foundations of the proposed nonlocal framework. To assess its practical relevance, we conduct a detailed empirical validation using high-resolution NGSIM trajectory data. The results demonstrate that the spatial-temporal nonlocal model significantly improves the reconstruction of traffic density fields compared with traditional local macroscopic models, particularly in regimes where anticipation effects dominate. These findings highlight the potential of spatial-temporal nonlocal traffic dynamics as a robust theoretical and data-driven framework for capturing complex traffic behavior.

math.NA↗

arXiv2026

UrbanTwin: Synthetic Roadside LiDAR Datasets

This article presents UrbanTwin datasets, high-fidelity, realistic replicas of three public roadside lidar datasets: LUMPI, V2X-Real-IC, and TUMTraf-I. Each UrbanTwin dataset contains 10K annotated frames corresponding to one of the public datasets. Annotations include 3D bounding boxes, instance segmentation labels, and tracking IDs for six object classes, along with semantic segmentation labels for nine classes. These datasets are synthesized using emulated lidar sensors within realistic digital twins, modeled based on surrounding geometry, road alignment at lane level, and the lane topology and vehicle movement patterns at intersections of the actual locations corresponding to each real dataset. Due to the precise digital twin modeling, the synthetic datasets are well aligned with their real counterparts, offering strong standalone and augmentative value for training deep learning models on tasks such as 3D object detection, tracking, and semantic and instance segmentation. We evaluate the alignment of the synthetic replicas through statistical and structural similarity analysis with real data, and further demonstrate their utility by training 3D object detection models solely on synthetic data and testing them on real, unseen data. The high similarity scores and improved detection performance, compared to models trained on real data, indicate that the UrbanTwin datasets effectively enhance existing benchmark datasets by increasing sample size and scene diversity. In addition, the digital twins can be adapted to test custom scenarios by modifying the design and dynamics of the simulations. To our knowledge, these are the first digitally synthesized datasets that can replace in-domain real-world datasets for lidar perception tasks. UrbanTwin datasets are publicly available at https://dataverse.harvard.edu/dataverse/ucf-ut.

cs.CV↗

arXiv2025

UrbanTwin: Building High-Fidelity Digital Twins for Sim2Real LiDAR Perception and Evaluation

LiDAR-based perception in intelligent transportation systems (ITS) relies on deep neural networks trained with large-scale labeled datasets. However, creating such datasets is expensive, time-consuming, and labor-intensive, limiting the scalability of perception systems. Sim2Real learning offers a scalable alternative, but its success depends on the simulation's fidelity to real-world environments, dynamics, and sensors. This tutorial introduces a reproducible workflow for building high-fidelity digital twins (HiFi DTs) to generate realistic synthetic datasets. We outline practical steps for modeling static geometry, road infrastructure, and dynamic traffic using open-source resources such as satellite imagery, OpenStreetMap, and sensor specifications. The resulting environments support scalable and cost-effective data generation for robust Sim2Real learning. Using this workflow, we have released three synthetic LiDAR datasets, namely UT-LUMPI, UT-V2X-Real, and UT-TUMTraf-I, which closely replicate real locations and outperform real-data-trained baselines in perception tasks. This guide enables broader adoption of HiFi DTs in ITS research and deployment.

cs.CV↗

arXiv2025

LiGuard: A Streamlined Open-Source Framework for Rapid & Interactive Lidar Research

There is a growing interest in the development of lidar-based autonomous mobility and Intelligent Transportation Systems (ITS). To operate and research on lidar data, researchers often develop code specific to application niche. This approach leads to duplication of efforts across studies that, in many cases, share multiple methodological steps such as data input/output (I/O), pre/post processing, and common algorithms in multi-stage solutions. Moreover, slight changes in data, algorithms, and/or research focus may force major revisions in the code. To address these challenges, we present LiGuard, an open-source software framework that allows researchers to: 1) rapidly develop code for their lidar-based projects by providing built-in support for data I/O, pre/post processing, and commonly used algorithms, 2) interactively add/remove/reorder custom algorithms and adjust their parameters, and 3) visualize results for classification, detection, segmentation, and tracking tasks. Moreover, because it creates all the code files in structured directories, it allows easy sharing of entire projects or even the individual components to be reused by other researchers. The effectiveness of LiGuard is demonstrated via case studies.

cs.CV↗

arXiv2025

High-Fidelity Digital Twins for Bridging the Sim2Real Gap in LiDAR-Based ITS Perception

Sim2Real domain transfer offers a cost-effective and scalable approach for developing LiDAR-based perception (e.g., object detection, tracking, segmentation) in Intelligent Transportation Systems (ITS). However, perception models trained in simulation often under perform on real-world data due to distributional shifts. To address this Sim2Real gap, this paper proposes a high-fidelity digital twin (HiFi DT) framework that incorporates real-world background geometry, lane-level road topology, and sensor-specific specifications and placement. We formalize the domain adaptation challenge underlying Sim2Real learning and present a systematic method for constructing simulation environments that yield in-domain synthetic data. An off-the-shelf 3D object detector is trained on HiFi DT-generated synthetic data and evaluated on real data. Our experiments show that the DT-trained model outperforms the equivalent model trained on real data by 4.8%. To understand this gain, we quantify distributional alignment between synthetic and real data using multiple metrics, including Chamfer Distance (CD), Maximum Mean Discrepancy (MMD), Earth Mover's Distance (EMD), and Fr'echet Distance (FD), at both raw-input and latent-feature levels. Results demonstrate that HiFi DTs substantially reduce domain shift and improve generalization across diverse evaluation scenarios. These findings underscore the significant role of digital twins in enabling reliable, simulation-based LiDAR perception for real-world ITS applications.

cs.CV↗

arXiv2025

A Stochastic Differential Equation Framework for Modeling Queue Length Dynamics Inspired by Self-Similarity

This article develops a stochastic differential equation (SDE) for modeling the temporal evolution of queue length dynamics at signalized intersections. Inspired by the observed quasiperiodic and self-similar characteristics of the queue length dynamics, the proposed model incorporates three properties into the SDE: (i) mean reversion with periodic mean, (ii) multiplicative noise, and (iii) fractional Brownian motion. It replicates key statistical features observed in real data, including the probability distribution function (PDF) and PSD of queue lengths. To our knowledge, this is the first equation-based model for queue dynamics. The proposed approach offers a transparent, data-consistent framework that may help inform and enhance the design of black-box learning algorithms with underlying traffic physics.

eess.SY↗

arXiv2025

DINOSTAR: Deep Iterative Neural Object Detector Self-Supervised Training for Roadside LiDAR Applications

Recent advancements in deep-learning methods for object detection in point-cloud data have enabled numerous roadside applications, fostering improvements in transportation safety and management. However, the intricate nature of point-cloud data poses significant challenges for human-supervised labeling, resulting in substantial expenditures of time and capital. This paper addresses the issue by developing an end-to-end, scalable, and self-supervised framework for training deep object detectors tailored for roadside point-cloud data. The proposed framework leverages self-supervised, statistically modeled teachers to train off-the-shelf deep object detectors, thus circumventing the need for human supervision. The teacher models follow fine-tuned set standard practices of background filtering, object clustering, bounding-box fitting, and classification to generate noisy labels. It is presented that by training the student model over the combined noisy annotations from multitude of teachers enhances its capacity to discern background/foreground more effectively and forces it to learn diverse point-cloud-representations for object categories of interest. The evaluations, involving publicly available roadside datasets and state-of-art deep object detectors, demonstrate that the proposed framework achieves comparable performance to deep object detectors trained on human-annotated labels, despite not utilizing such human-annotations in its training process.

cs.CV↗

arXiv2024

An Extreme Value Theory Approach for Understanding Queue Length Dynamics in Adaptive Corridors

This paper introduces a novel approach employing extreme value theory to analyze queue lengths within a corridor controlled by adaptive controllers. We consider the maximum queue lengths of a signalized corridor consisting of nine intersections every two minutes, roughly equivalent to the cycle length. Our research shows that maximum queue lengths at all the intersections follow the extreme value distributions. To the best knowledge of the authors, this is the first attempt to characterize queue length time series using extreme value analysis. These findings are significant as they offer a mechanism to assess the extremity of queue lengths, thereby aiding in evaluating the effectiveness of the adaptive signal controllers and corridor management. Given that extreme queue lengths often precipitate spillover effects, this insight can be instrumental in preempting such scenarios.

eess.SY↗

arXiv2024

Self-Similar Characteristics in Queue Length Dynamics: Insights from Adaptive Signalized Corridor

Self-similarity, a fractal characteristic of traffic flow dynamics, is widely recognized in transportation engineering and physics. However, its practical application in real-world traffic scenarios remains limited. Conversely, the traffic flow dynamics at adaptive signalized intersections still need to be fully understood. This paper addresses this gap by analyzing the queue length time series from an adaptive signalized corridor and characterizing its self-similarity. The findings uncover a $1/f$ structure in the power spectrum of queue lengths, indicative of self-similarity. Furthermore, the paper estimates local scaling exponents $(α)$, a measure of self-similarity computed via detrended fluctuation analysis (DFA), and identifies a positive correlation with congestion patterns. Additionally, the study examines the fractal dynamics of queue length through the evolution of scaling exponent. As a result, the paper offers new insights into the queue length dynamics of signalized intersections, which might help better understand the impact of adaptivity within the system.

eess.SP↗

arXiv2024

Existence of Trust-field in Vehicular Ad Hoc Networks: Empirical Evidence

Vehicular Ad Hoc Networks (VANETs) play a crucial role in enhancing road safety and traffic efficiency by enabling communication between vehicles (V2V) and between vehicles and infrastructure (V2I). Robust trust management is necessary to ensure the reliability of information in decentralized systems. This paper presents the notion of a ``Trust Field" in VANETs, conceptualized as the behavior of the nodes that represents trust levels evolving in both spatial and temporal dimensions. Using the LogitTrust model, we provide empirical evidence of how trust fields in vehicular networks change over time in different scenarios, including when malicious nodes are present. The results of our study demonstrate that the trust domain can adjust to fluctuations in network conditions, thereby offering a comprehensive metric for assessing the reliability of nodes. This innovative method improves the dependability of VANET applications by efficiently detecting and mitigating malicious actions.

math.DS↗

arXiv2024

Data-driven discovery of quasiperiodically driven dynamics

The analysis of a timeseries can provide many new perspectives if it is accompanied by the assumption that the timeseries is generated from an underlying dynamical system. For example, statistical properties of the data can be related to measure theoretic aspects of the dynamics, and one can try to recreate the dynamics itself. The underlying dynamics could represent a natural phenomenon or a physical system, where the timeseries represents a sequence of measurements. In this paper, we present a completely data-driven framework to identify and model quasiperiodically driven dynamical systems (Q.P.D.) from the timeseries it generates. Q.P.D. are a special class of systems that are driven by a periodic source with multiple base frequencies. Such systems abound in nature, e.g., astronomy and traffic flow. Our framework reconstructs the dynamics into two components - the driving quasiperiodic source with generating frequencies; and the driven nonlinear dynamics. We make a combined use of a kernel-based harmonic analysis, kernel-based interpolation technique, and Koopman operator theory. Our framework provides accurate reconstructions and frequency identification for three real-world case studies.

math.DS↗

arXiv2023

Incorporating Nonlocal Traffic Flow Model in Physics-informed Neural Networks

This research contributes to the advancement of traffic state estimation methods by leveraging the benefits of the nonlocal LWR model within a physics-informed deep learning framework. The classical LWR model, while useful, falls short of accurately representing real-world traffic flows. The nonlocal LWR model addresses this limitation by considering the speed as a weighted mean of the downstream traffic density. In this paper, we propose a novel PIDL framework that incorporates the nonlocal LWR model. We introduce both fixed-length and variable-length kernels and develop the required mathematics. The proposed PIDL framework undergoes a comprehensive evaluation, including various convolutional kernels and look-ahead windows, using data from the NGSIM and CitySim datasets. The results demonstrate improvements over the baseline PIDL approach using the local LWR model. The findings highlight the potential of the proposed approach to enhance the accuracy and reliability of traffic state estimation, enabling more effective traffic management strategies.

cs.LG↗

arXiv2023

Physics Informed Deep Learning: Applications in Transportation

A recent development in machine learning - physics-informed deep learning (PIDL) - presents unique advantages in transportation applications such as traffic state estimation. Consolidating the benefits of deep learning (DL) and the governing physical equations, it shows the potential to complement traditional sensing methods in obtaining traffic states. In this paper, we first explain the conservation law from the traffic flow theory as ``physics'', then present the architecture of a PIDL neural network and demonstrate its effectiveness in learning traffic conditions of unobserved areas. In addition, we also exhibit the data collection scenario using fog computing infrastructure. A case study on estimating the vehicle velocity is presented and the result shows that PIDL surpasses the performance of a regular DL neural network with the same learning architecture, in terms of convergence time and reconstruction accuracy. The encouraging results showcase the broad potential of PIDL for real-time applications in transportation with a low amount of training data.

cs.LG↗

arXiv2023

On the Limitations of Physics-informed Deep Learning: Illustrations Using First Order Hyperbolic Conservation Law-based Traffic Flow Models

Since its introduction in 2017, physics-informed deep learning (PIDL) has garnered growing popularity in understanding the evolution of systems governed by physical laws in terms of partial differential equations (PDEs). However, empirical evidence points to the limitations of PIDL for learning certain types of PDEs. In this paper, we (a) present the challenges in training PIDL architecture, (b) contrast the performance of PIDL architecture in learning a first order scalar hyperbolic conservation law and its parabolic counterpart, (c) investigate the effect of training data sampling, which corresponds to various sensing scenarios in traffic networks, (d) comment on the implications of PIDL limitations for traffic flow estimation and prediction in practice. Detailed in the case study, we present the contradistinction in PIDL results between learning the traffic flow model (LWR PDE) and its variation with diffusion. The outcome indicates that PIDL experiences significant challenges in learning the hyperbolic LWR equation due to the non-smoothness of its solution. On the other hand, the architecture with parabolic PDE, augmented with the diffusion term, leads to the successful reassembly of the density data even with the shockwaves present.

cs.LG↗