Mashrur Chowdhury

Horizontal curves are often associated with roadway crashes because of speed misjudgment and loss of control. With the growing adoption of connected and autonomous vehicles (CAVs), the accurate estimation of safe speed at curves is becoming increasingly important. The widely used AASHTO design method for safe curve speed estimation relies on an analytical equation based on a simplified point mass model, which often uses conservative parameters to account for vehicular and environmental variations. This paper presents a digital twin–based framework for estimating safe speed at curves using a physics-driven virtual environment developed in the Unity engine. In this framework, a real-world horizontal road curve is selected, and vehicle speed data are collected using a radar gun under various weather conditions. A 3-D model of the road curve is constructed in a Unity environment using roadway geometric and elevation data. A parameterized vehicle model is integrated, allowing for variations in mass, acceleration, and center of gravity to reflect different vehicle types and loading scenarios. The simulation identifies the maximum safe speed at which a vehicle can traverse the given curve, providing a more vehicle- and environment-specific estimate of the safe operating speed. The study validated that the safe curve speed estimates generated by the simulation were consistent with the real-world speed values observed at the curve. This study demonstrates how a physics-based digital twin can estimate a safer and more adaptive operating speed for vehicles traversing horizontal curves.

Computer vision plays a critical role in ensuring the safe navigation of autonomous vehicles (AVs). An AV perception module facilitates safe navigation. This module enables AVs to recognize traffic signs, traffic lights, and various road users. However, the perception module is vulnerable to adversarial attacks, which can compromise its accuracy and reliability. One such attack is the adversarial patch attack (APA), an attack in which an adversary strategically places a specially crafted sticker on an object to deceive object classifiers. Such an APA can cause AVs to misclassify traffic signs, leading to catastrophic incidents. To enhance the security of an AV perception system against APAs, this study develops a Generative Adversarial Network (GAN)-based single-stage defense strategy for traffic sign classification. This approach is tailored to defend against APAs across different classes of traffic signs, without prior knowledge of a patch's design, and is effective against patches of varying sizes. In addition, our single-stage defense is computationally efficient, requiring significantly lower computation time than existing multi-stage defenses, making it suitable for real-time deployment in autonomous driving systems. Compared to a classifier without any defense mechanism, our experimental analysis demonstrates that the defense strategy presented in this paper improves our classifier’s accuracy under APA conditions by up to 90% considering traffic sign classes considered in this study. and overall classification accuracy enhances by 55% for all traffic signs considered in this study. Our defense strategy is model agnostic, making it applicable to any traffic sign classifier, regardless of the underlying classification model.

Many Intelligent Transportation Systems (ITS) applications require strong privacy guarantees for both users and their data. Homomorphic encryption (HE) enables computation directly on encrypted messages and thus offers a compelling approach to privacy-preserving data processing in ITS. However, practical HE schemes incur substantial ciphertext expansion and communication overhead, which limits their suitability for time-critical transportation systems. Hybrid homomorphic encryption (HHE) addresses this challenge by combining a homomorphic encryption scheme with a symmetric cipher, enabling efficient encrypted computation while dramatically reducing communication cost. In this paper, we develop theoretical models of representative ITS applications that integrate HHE to protect sensitive vehicular data. We then perform a parameter-based evaluation of the HHE scheme Rubato to estimate ciphertext sizes and communication overhead under realistic ITS workloads. Our results show that HHE achieves orders-of-magnitude reductions in ciphertext size compared with conventional HE while maintaining cryptographic security, making it significantly more practical for latency-constrained ITS communication.

Cyber-physical systems rely on sensors, communication, and computing, all powered by integrated circuits (ICs). These ICs are vulnerable to malicious hardware attacks, with hardware Trojans being one of the stealthiest threats. Trojans are malicious implants in the circuitry, which are often inserted during design or fabrication stages. This stealthy addition remains dormant until triggered and might cause functional disruptions or sensitive information leakage once triggered. Traditional IC validation methods, such as functional testing and logic analysis, usually fail to capture these subtle anomalies because hardware Trojans are intentionally designed to mimic normal circuit behavior. They often remain dormant under standard test vectors, and the changes they bring into power consumption, timing, and area are minimal. We present a dual-domain feature extraction strategy that combines time-domain features with frequency-domain characteristics of power traces. For the detection, we created an artificial intelligence (AI) based robust Trojan detection framework that integrates traditional machine learning models, such as random forest (RF), gradient boosting (GB), naïve bayes (NB), and deep learning models, such as deep neural network (DNN), long short-term memory (LSTM), and graph neural network (GNN). In this study, we consider these models as baseline AI models to detect Trojan-infected circuits via side-channel power analysis. We employed a stacked ensemble classifier that integrates the distinct strengths of the six baseline models used in this study. After evaluating our stacking ensemble-based detection method on the advanced encryption standard (AES)-Trojan benchmark, which covers diverse Trojan types, the results demonstrate that the ensemble method consistently outperformed all six baseline models. The ensemble-based detection method achieved a macro-averaged area under the receiver operating characteristic (ROC) curve (AUC) of 0.987, while remaining golden-chip-free, meaning it does not rely on a trusted reference IC for baseline comparison. Instead it detects anomalies directly from observable characteristics of untrusted chips, such as side-channel emissions.

With the advancement of connected and automated transportation systems, a growing need has emerged for regulatory authorities to update existing laws and statutes, as well as create new ones, to address future implications of connectivity and automation. Specifically, ensuring proper engagement with cybersecurity and data privacy challenges for connected and automated transportation systems will require a comprehensive legal framework at both the federal and state levels. To help policymakers achieve this, a retrieval augmented generation (RAG)-based large language model (LLM) framework, Transportation Cybersecurity and Resiliency (TraCR) AI, has been developed in this study, focusing on extracting pertinent information from existing legislation based on inquiries and crafting LLM-generated responses to highlight potential loopholes for further scrutiny. This study primarily aims to mitigate the hallucinations caused in the domain of LLMs by developing a curated knowledge base of legislative documents and an associated question–answer dataset to improve the effectiveness of query results. This RAG-based framework extracts relevant information to improve the specificity of answers and aids the LLM in providing factually accurate responses with increased reliability. Our analyses reveal that the presented RAG-based framework can aid legislative analysis by generating queries for particular questions and responses. We also compare our RAG framework-generated responses with commercially available LLMs to demonstrate the effectiveness of our approach. TraCR AI outperforms leading commercial LLMs across four distinct metrics, that is, AlignScore, ParaScore, BERTScore, and ROUGE score. This highlights that integrating RAG allows LLMs to produce more factually accurate and up-to-date responses than standalone LLMs. This approach to domain-specific LLM development will improve the quality of legislative analysis that can be used to aid policymakers in meeting the challenge of updating legal codes in accordance with emerging technologies.

This study investigates crash severity risk modeling strategies for work zones involving large vehicles (i.e., trucks, buses, and vans) under crash data imbalance between low-severity (LS) and high-severity (HS) crashes. We utilized crash data involving large vehicles in South Carolina work zones from 2014 to 2018, which included four times more LS crashes than HS crashes. The objective of this study is to evaluate the crash severity prediction performance of various statistical, machine learning, and deep learning models under different feature selection and data balancing techniques. Findings highlight a disparity in LS and HS predictions, with lower accuracy for HS crashes owing to class imbalance and feature overlap. Discriminative mutual information (DMI) yields the most effective feature set for predicting HS crashes without requiring data balancing, particularly when paired with gradient boosting models and deep neural networks, such as CatBoost, NeuralNetTorch, XGBoost, and LightGBM. Data balancing techniques, such as NearMiss-1, maximize HS recall when combined with DMI-selected features and certain models, such as LightGBM, making them well suited for HS crash prediction. Conversely, RandomUnderSampler, HS Class Weighting, and RandomOverSampler achieve more balanced performance, which is defined as an equitable trade-off between LS and HS metrics, especially when applied to NeuralNetTorch, NeuralNetFastAI, CatBoost, LightGBM, and Bayesian mixed logit (BML) using merged feature sets or models without feature selection. The insights from this study offer safety analysts guidance on selecting models, feature selection, and data balancing techniques aligned with specific safety goals, providing a robust foundation for enhancing work-zone crash severity prediction.

Emerging transportation technologies, including electric vehicles (EVs), autonomous vehicles (AVs), and connected vehicles (CVs), are poised to revolutionize mobility and logistics through advancements in connectivity, automation, and electrification. However, the convergence of these systems introduces substantial cybersecurity challenges, including data breaches, spoofing, and infrastructure attacks, due to increasingly complex and interconnected attack surfaces. This article presents a comprehensive survey of cybersecurity threats, vulnerabilities, and evaluation practices within the AV, CV, and EV landscape. We analyze 227 peer-reviewed studies published between 2021 and 2025 and introduce two novel taxonomies: a three-stage cyber-attack lifecycle framework—Stream to Information (S2I), Information to Decision (I2D), and Decision to Actuation (D2A)—and 11 representative attack paths. Our findings reveal that the stream and data processing stages are most frequently exploited, with limited standardization in evaluation metrics and inadequate emphasis on real-world operational consequences. We highlight the growing importance of integrating quantum-safe cryptography and AI-driven anomaly detection as proactive defense mechanisms. Finally, we offer stage-wise design recommendations and identify future research directions, including the need for cross-domain evaluation frameworks, cyber-social risk assessments, and secure integration of emerging multi-modal systems. This survey aims to support policymakers, researchers, and industry stakeholders in developing resilient, secure, and trustworthy next-generation road transportation ecosystems.