Hi, I’m Haolin.

Building a reliable crack record

I proposed a two-level cracking-data quality methodology: compare crack maps against ground references, then separate sensor limitations from detection-algorithm failures. Validation across pavement types, systems, and algorithms in Florida, Mississippi, and Texas provides a known-quality basis for condition histories. The paper is accepted in Transportation Research Record.

I’m tracing how error enters through data collection, segmentation, vectorization, and aggregated metrics—and how much reaches the final condition rating. Repeat-run 3D laser surveys of US-80 provide the empirical benchmark for distinguishing deterioration from measurement noise. The manuscript is in preparation.

I maintain and extend the lab’s Crack Vector Model (CVM), which turns crack masks and range images into node–link graphs with per-link width, length, and orientation. I’m also designing structure-aware thin-crack detection to reduce fragmentation where depth evidence is weak; that study is at an early stage.

Following change across years and sensors

I co-developed feature-based registration for repeated 3D pavement surveys, separating true deterioration from image-offset artifacts. This work was published in Computer-Aided Civil and Infrastructure Engineering in 2025. I’m also co-developing unsupervised deformable crack-map registration for crack-propagation monitoring.

Another study calibrates low-cost RGB imagery against co-registered 3D laser surveys. It tests whether more frequent camera observations preserve the crack measures agencies act on, helping fill gaps between survey cycles and trigger high-resolution updates where change is detected. The deformable-registration and RGB-surrogate manuscripts are in preparation.

Crack condition, severity, and treatment selection

I proposed a link-level crack-density indicator that replaces grid-based density with a buffer around each CVM crack link. It was calibrated and cross-validated on 150 crack maps spanning four crack types and road classes for alligator-crack classification, providing a consistent severity measure for treatment selection.

ASCE Journal of Computing in Civil Engineering, in revision.

Planning robotic crack sealing

I formalized crack-sealing path planning as a sliding-window arc-routing problem, a Rural Postman Problem variant. As a sealing vehicle advances along the lane, a robotic arm must trace every crack within its operating constraints and satisfy a full-coverage requirement.

The work connects treatment prescription with automated preventive maintenance. We are benchmarking heuristic, exact, and reinforcement-learning solvers with a reproducible pipeline on QA-checked road data. The manuscript is in preparation; the benchmark release is planned with it.

OrthoLaneMark: measuring the right pavement region

I built a presence-aware lane-marking benchmark with 1,391 annotated orthographic 3D-laser images, a Gated Region Accuracy metric, and nine re-implemented detectors across four mechanisms. It examines how lane-marking presence reasoning, rather than sub-pixel localization alone, bounds wheelpath-region accuracy for pavement distress measurement.

Journal submission, September 2026; companion presentation accepted for TRB 2027.

Connecting the research to agency practice

As a core member of NCHRP 01-57B, I developed and validated quality-assessment and crack-definition protocols and contributed to the Final Report, Tech Brief, proposed revisions to AASHTO R 85, and crack-data aggregation standard. I co-authored panel-ballot responses and presented to the project panel and at RPUG in 2025 and 2026.

I’m conducting a systematic review of pavement digital twins for life-cycle maintenance and rehabilitation decisions. The framework connects creation, synchronization, and application in a bidirectional loop: observations update the twin, and the twin guides monitoring and treatment. Its initial version was presented at the 2025 ACM SIGSPATIAL Workshop on Spatial Digital Twins.

Peer-reviewed journal articles

1. Wang, H.†, Yang, Z.†, Geary, G. M., & Tsai, Y.-C. (2026). Standardizing cracking data quality assessment and quantification. Transportation Research Record. Accepted (Paper 25-01724); in press.
2. Yang, Z., Mohammadi, M., Wang, H., & Tsai, Y.-C. (2025). A feature-based pavement image registration method for precise pavement deterioration monitoring. Computer-Aided Civil and Infrastructure Engineering, 40(10), 1276–1294. (IF 9.1, Q1)

Peer-reviewed conference and workshop papers

1. Tsai, Y.-C., Wang, H., & Luo, S. (2026). Work in progress: A problem-solving systems approach to AI education: Lifecycle modules, inter-module reasoning, and values in practice. 2026 ASEE Annual Conference & Exposition, Charlotte, NC. doi:10.18260/1-2--59094
2. Wang, H., & Tsai, Y.-C. (2025). Toward reliable pavement digital twins: Multi-scale spatiotemporal monitoring challenges and solutions. Proceedings of the 1st ACM SIGSPATIAL International Workshop on Spatial Digital Twins. doi:10.1145/3764924.3770893

Manuscripts under review or in revision

1. Wang, H.*, Yang, Z.*, Bao, J., Tsai, Y.-C., Geary, G. M., Cary, C., & Frith, D. J. A link-level crack density indicator based on the Crack Vector Model. ASCE Journal of Computing in Civil Engineering. In revision (CPENG-8858).
2. Wang, H.*, Ryu, S.*, Yang, Z., & Tsai, Y.-C. CRAck Structural Difference Index (CRASDI): A crack-topology-aware index for crack map comparison. Computer-Aided Civil and Infrastructure Engineering. Under review.
3. Wang, H., Luo, S., Yang, Z., & Tsai, Y.-C. Presence-aware benchmarking of lane-marking detection methods for pavement distress measurement. Submitted to an ASCE journal, September 2026; companion presentation accepted for the TRB 106th Annual Meeting.

Preprints

1. Zhang, X., Wang, H., Hsieh, Y. A., Yang, Z., Yezzi, A., & Tsai, Y.-C. (2025). Deep learning for crack detection: A review of learning paradigms, generalizability, and datasets. arXiv:2508.10256
2. Zhang, X., Wang, H., & Yang, Z. (2026). Towards successful implementation of automated raveling detection: Effects of training data size, illumination difference, and spatial shift. arXiv:2604.13322

In preparation · First author

1. Wang, H., Yang, Z., & Tsai, Y.-C. Uncertainty quantification for pavement surface crack measurement. Target journal: Reliability Engineering & System Safety (draft complete; submission planned early 2027).
2. Wang, H., et al. Crack-sealing path planning as a sliding-window arc-routing problem: Problem definition and a benchmark of classical, exact, and learned solvers. Target journal: Computer-Aided Civil and Infrastructure Engineering (2027).
3. Wang, H., et al. Feasibility of low-cost RGB imagery as a surrogate for 3D laser pavement surveys: Calibration and cost-effective frequent monitoring. Target journal: Transportation Research Part C: Emerging Technologies (2027).
4. Wang, H., & Tsai, Y.-C. Digital twins for pavement maintenance and rehabilitation: A systematic review of creation, synchronization, and application. Target journal: Automation in Construction (2027).

In preparation · Co-author

1. Fung, J., Wang, H., Lou, Z., Yang, Z., & Tsai, Y.-C. Unsupervised deep learning methodology for spatial-temporal pavement image registration. Target journal: Automation in Construction (2027).

Technical report

1. Tsai, Y.-C., et al. (2026). Validating standard definitions for comparable pavement cracking data. NCHRP Project 01-57B Final Report, with appendices on crack-protocol design, data quality assessment, proposed revisions to AASHTO R 85, and crack-data aggregation. Transportation Research Board. (Contributing author)

CRASDI

Wang, Ryu, Yang, Tsai (2026). Python package for crack-topology-aware crack-map comparison, with reproducibility scripts and per-pair results for nine public datasets.

OrthoLaneMark

Wang, Luo, Yang, Tsai (2026). Presence-aware benchmark for lane-marking detection: 1,391 annotated images, an evaluation framework, nine method implementations, and trained checkpoints.

TRACE

Lou, Yang, Wang, Tsai (2026). Temporal Registered Annual Crack-map Evolution: a 20-sequence preview dataset.

Crack-sealing path-planning benchmark

Benchmarking heuristic, exact, and reinforcement-learning solvers with a reproducible pipeline on QA-checked road data. Release planned with the manuscript.

Ph.D. in Civil Engineering

Construction and Infrastructure Systems Engineering. Advisor: Dr. Yi-chang (James) Tsai. GPA 4.00.

Dissertation: Toward a Digital Twin of Pavement Cracking: Uncertainty-Aware Creation, Structured Diagnosis, and Treatment Prescription (proposal in preparation, Fall 2026).

M.S. in Computer Science & Engineering

B.Eng. in Civil Engineering

GPA 3.89. Recognized as an Excellent Graduate in 2021.

Graduate Research Assistant

Smart City Infrastructure Lab, advised by Dr. James Tsai. Projects funded by NCHRP and GDOT connect survey-data quality, pavement deterioration monitoring, and maintenance decisions.

Undergraduate Researcher

I studied nondestructive acoustic-emission crack monitoring for high-speed-railway ballastless track slabs, linking acoustic signatures to finite-element-simulated crack propagation. This work with Dr. Yu Zhou led to two granted CNIPA patents.

With Dr. Yujie Lu, I investigated a life-cycle maturity model of intelligent construction in urban rail transit through literature review, questionnaires, and hierarchical analysis; the work was recognized as an excellent project.

Mini Traffic: urban-road simulator

Team lead for the Mini Traffic urban-road simulator, awarded third prize in the Shanghai University Student Computer Application Ability Competition in 2019.

Head Teaching Assistant · CEE 3090

Data Analytics in Civil & Environmental Engineering. I deliver lectures on tree-based methods, deep learning, and survival analysis, and lead the Python lab sequence from environment setup through regression, classification, cross-validation, and model selection.

I designed the five-lab Python curriculum—student notebooks, instructor keys, and slide decks—using real civil-engineering datasets: EPA PM2.5 monitors, UCI concrete strength, FHWA LTPP pavement sections, and the full Georgia National Bridge Inventory. An observational-confounding lesson prepares students for the course project.

I co-developed the midterm and term project, coordinate the TA team, assign responsibilities, and review student analyses and code with written feedback.

Previous teaching

CEE 4803 AI for Smart Cities · Spring 2024 and Spring 2025. Lectures, Python labs, assignments, and grading rubrics for undergraduate and graduate students.

CEE 8813 Data Analytics for Transportation Safety · Fall 2024. Python/GIS crash-analytics assignments and term-project mentoring.

CEE 4803 Data Analytics in Civil & Environmental Engineering · Fall 2023 and Fall 2024. Assignments and lecture support in the course that became CEE 3090.

Graduate Mentor · Vertically Integrated Projects

I supervise undergraduate and graduate teams on semester-long research tasks, define scopes, run weekly check-ins, and review deliverables. Projects include crack-definition standardization, spatiotemporal pavement analysis, image registration, lane-marking annotation, and LiDAR-aided sidewalk assessment.

Mentees include Sooha Ryu, Jordan Fung, Jiao Bao, and Trinh Flynt. Mentees are co-authors on the CRASDI, crack-density, and deformable-registration manuscripts.

Teaching interests

Data analytics and machine learning for civil engineering; infrastructure sensing and condition assessment; pavement engineering and asset management; digital twins and spatiotemporal data; optimization and maintenance automation; programming and reproducible research.

Peer review

ASCE Journal of Computing in Civil Engineering · 16 reviews covering 10 manuscripts including revisions, 2024–present.

Transportation Research Record · 2 reviews, 2026.

TRB Annual Meeting · 6 papers for the 2026 and 2027 meetings.

Project and community

Core team member, NCHRP Project 01-57B; presenter to the project panel and at the Road Profile Users’ Group in 2025 and 2026.

Honors & awards

CEE Bill Schutz Graduate Teaching Assistant Award, Georgia Tech (2025); Excellent Graduate, Tongji University (2021); First Prize, Tongji Scholarship for Excellent Students (2020); Bronze Award, Fifth “Excellence Cup” Competition, Tongji University (2020); Third Prize, Shanghai University Student Computer Application Ability Competition (2019); Excellent Student Leader, Tongji University (2018 and 2019).