Ruimin Ke

Emerging transportation big data are expanding the evidence available for proactive road safety assessment. However, extracting visible or measurable road environment features does not by itself produce the coded safety variables required by structured infrastructure assessment protocols. This review examines how far literature studying emerging data sources has progressed from feature extraction toward protocol-defined, reviewable safety variable rating. Using the International Road Assessment Programme (iRAP) Star Rating protocol as an operational reference, we synthesize the literature around static road environment variables rather than data sources or model families. Studies are interpreted through a four-level readiness framework: enabling evidence, proxy evidence, direct variable extraction, and direct protocol coding. The synthesis covers four variable groups: roadside severity, mid-block attributes, intersection and access variables, and vulnerable road user facilities and land use. The review shows that readiness is uneven across variables and depends on the relationship between available evidence and protocol requirements. Variables tied to visible assets or calibrated geometric measurement, such as barriers, lane width, divided carriageway, pedestrian crossings, and sidewalk presence, have the strongest evidence base. Variables requiring functional interpretation, side- or approach-specific assignment, temporal currency, physical measurement, or judgment under protocol rules remain less ready. Major gaps include validation along the evidence-to-code path, benchmarks organized around safety variables, multimodal workflows that preserve reviewable intermediate evidence, and governance practices for combining sources with different coverage, currency, and authority. Addressing these gaps is necessary to move from fragmented feature inventories toward scalable, reviewable, and protocol-ready assessment.

Improving accessibility in public transportation for older adults requires balancing equity-driven service objectives with operational efficiency, a trade-off that bilevel optimization provides a rigorous framework to address, in which strategic stop selection and operational route design are hierarchically coupled. Metaheuristic approaches such as genetic algorithms are well suited to these problems, but their performance is highly sensitive to the quality and feasibility of the initial population, particularly under stringent routing constraints. This study introduces a hybrid quantum-classical framework for accessibility-oriented bus route design in which the Quantum Approximate Optimization Algorithm is used as a structured sampling operator to initialize a genetic algorithm. The upper-level accessibility model is reformulated as a Quadratic Unconstrained Binary Optimization (QUBO) problem, encoding objective and feasibility information into the Quantum Approximate Optimization Algorithm (QAOA) cost Hamiltonian and biasing sampling toward constraint-aware regions of the decision space. The lower-level routing problem is solved classically through a nested best-response procedure, preserving the hierarchical structure of the bilevel formulation. Initial populations are characterized using Hamming distance distributions and quality-diversity metrics, and convergence behavior is assessed through pairwise Mann-Whitney tests across multiple spatial configurations and route length tolerances. All initialization strategies converge to statistically indistinguishable final solutions, confirming that the bilevel formulation reliably identifies high-quality routes regardless of initialization. The primary distinction lies in search efficiency: QAOA consistently achieves full feasibility at generation zero, produces populations with the best objective quality and closest proximity to the final optimal solution, and maintains greater structural diversity than greedy or simulated annealing alternatives. Simulated annealing achieves competitive GA-phase convergence but incurs preprocessing costs two orders of magnitude larger than other strategies, negating its quality advantage in wall-clock time. These findings demonstrate that the contribution of QAOA lies not in quantum computational advantage but in reshaping the initial search distribution to eliminate infeasible exploration and accelerate convergence, offering a scalable approach for constrained transportation network design.

Improving accessibility in public transportation for older adults requires balancing equity-driven service objectives with operational efficiency, a trade-off that bilevel optimization provides a rigorous framework to address, in which strategic stop selection and operational route design are hierarchically coupled. Metaheuristic approaches such as genetic algorithms are well suited to these problems, but their performance is highly sensitive to the quality and feasibility of the initial population, particularly under stringent routing constraints. This study introduces a hybrid quantum-classical framework for accessibility-oriented bus route design in which the Quantum Approximate Optimization Algorithm is used as a structured sampling operator to initialize a genetic algorithm. The upper-level accessibility model is reformulated as a Quadratic Unconstrained Binary Optimization (QUBO) problem, encoding objective and feasibility information into the Quantum Approximate Optimization Algorithm (QAOA) cost Hamiltonian and biasing sampling toward constraint-aware regions of the decision space. The lower-level routing problem is solved classically through a nested best-response procedure, preserving the hierarchical structure of the bilevel formulation. Initial populations are characterized using Hamming distance distributions and quality-diversity metrics, and convergence behavior is assessed through pairwise Mann-Whitney tests across multiple spatial configurations and route length tolerances. All initialization strategies converge to statistically indistinguishable final solutions, confirming that the bilevel formulation reliably identifies high-quality routes regardless of initialization. The primary distinction lies in search efficiency: QAOA consistently achieves full feasibility atgeneration zero, produces populations with the best objective quality and closest proximity to the final optimal solution, and maintains greater structural diversity than greedy or simulated annealing alternatives. Simulated annealing achieves competitive GA-phase convergence but incurs preprocessing costs two orders of magnitude larger than other strategies, negating its quality advantage in wall-clock time. These findings demonstrate that the contribution of QAOA lies not in quantum computational advantage but in reshaping the initial search distribution to eliminate infeasible exploration and accelerate convergence, offering a scalable approach for constrained transportation network design.

Public transit systems involve a wide range of combinatorial optimization problems arising from planning and operational decisions such as network design, timetabling, vehicle scheduling, and demand-responsive routing. Many of these problems contain complex interactions among routes, trips, vehicles, and passenger flows, which naturally lead to higher-order polynomial optimization structures when represented in binary form. These structures present both opportunities and challenges for emerging quantum optimization methods. This study investigates how representative public transit optimization problems can be translated into quantum optimization models and executed on real quantum computing hardware. We first examine several representative transit optimization problems and analyze the higher-order interaction structures that arise in their binary polynomial formulations. A unified framework is then developed to estimate quantum resource requirements and to analyze how higher-order formulations (HUBO) are transformed into hardware-compatible quadratic formulations (QUBO). Using this framework, we conduct a set of hardware-aware benchmark experiments across four representative transit optimization problems. The benchmark instances are implemented with Qiskit 2.1 and executed on IBM Quantum System One. The experiments analyze how polynomial interaction structures, auxiliary-variable growth, and logical variable counts evolve with increasing problem size. The results provide a hardware-grounded perspective on how higher-order transportation optimization problems translate into executable quantum models and how formulation choices influence quantum resource requirements. These findings offer early insights into the potential role of quantum optimization in future intelligent transportation systems.

Transit Origin-Destination (OD) data are fundamental for optimizing public transit services, yet current collection methods, such as manual surveys, Bluetooth/WiFi tracking, or Automated Passenger Counters, are either costly, device-dependent, or incapable of individual-level matching. Meanwhile, onboard surveillance cameras already deployed on most transit vehicles provide an underutilized opportunity for automated OD data collection. Leveraging this, we present TransitReID, a novel framework for individual-level and occlusion-resistant passenger re-identification (ReID) tailored to transit environments. Our approach introduces three key innovations: (1) an occlusion-robust ReID algorithm that integrates a variational autoencoder-guided region-attention mechanism and selective quality feature averaging to dynamically emphasize visible and discriminative body regions under severe occlusions and viewpoint variations; (2) a Hierarchical Storage and Dynamic Matching (HSDM) mechanism that transforms static gallery matching into a dynamic process, balancing accuracy, memory, and speed in real-world bus operations; and (3) a multi-threaded edge implementation that enables near real-time OD estimation while ensuring privacy by processing all data locally. To support research in this domain, we also construct a new Transit ReID dataset with over 17,000 images captured from bus front/rear cameras under diverse occlusion and viewpoint conditions. Experimental results demonstrate that TransitReID achieves state-of-the-art performance, with R-1 accuracy of 88.3% and mAP of 92.5%, and further sustains 90% OD estimation accuracy in bus route simulations on NVIDIA Jetson edge devices. This work advances both the algorithmic and system-level foundations of automated transit OD collection, paving the way for scalable, privacy-preserving deployment in intelligent transportation systems.

Edge sensing and computing are rapidly becoming part of intelligent infrastructure architecture, leading to operational reliance on such systems in disaster or emergency situations. In such scenarios, there is a high chance of power supply failure due to power grid issues, and communication system issues due to base stations losing power or being damaged by the elements, e.g., flooding, wildfires, and so on. Mobile edge computing in the form of unmanned aerial vehicles (UAVs) has been proposed to provide computation offloading from these devices to conserve their battery, while the use of UAVs as relay network nodes has also been investigated previously. This article considers the use of UAVs with further constraints on power and connectivity to prolong the life of the network while also ensuring that the data is received from the edge devices in a timely manner. Reinforcement learning is used to investigate numerous scenarios of various levels of power and communication failure. This approach is able to identify the device most likely to fail in a given scenario, thus providing priority guidance for maintenance personnel. The evacuations of a rural town and urban downtown area are also simulated to demonstrate the effectiveness of the approach at extending the life of the most critical edge devices.

Large Language Models (LLMs), capable of handling multi-modal input and outputs such as text, voice, images, and video, are transforming the way we process information. Beyond just generating textual responses to prompts, they can integrate with different software platforms to offer comprehensive solutions across diverse applications. In this paper, we present ChatSUMO, an LLM-based agent that integrates language processing skills to generate abstract and real-world simulation scenarios in the widely-used traffic simulator - Simulation of Urban MObility (SUMO). Our methodology begins by leveraging the LLM for user input, which adapts it to relevant keywords needed to run python scripts. These scripts are designed to convert specified regions into coordinates, fetch data from OpenStreetMap, transform it into a road network, and subsequently run SUMO simulations with the designated traffic conditions. The outputs of the simulations are then interpreted by the LLM resulting in informative comparisons and summaries. Users can continue the interaction and generate a variety of customized scenarios without prior traffic simulation expertise. Any city available from OpenStreetMap can be imported, and for demonstration, we created a real-world simulation for the city of Albany. ChatSUMO also allows simulation customization capabilities of edge edit, traffic light optimization, and vehicle edit by users through the web interface.