Developing Models for Estimating Winter Road Weather and Surface Conditions–An Empirical Investigation
Bibliographic record
Abstract
Inclement weather poses a threat to road safety and mobility for motorists in cold regions during winter. To facilitate more efficient winter maintenance decision support and reduce weather-related collisions, many transportation agencies have adopted one of the most critical highway infrastructures; namely, road weather information systems (RWIS). While RWIS are effective in collecting real-time road surface conditions (RSC) information, they are costly to install and operate. Equally important, RWIS provide point measurements that are often unrepresentative of distant surrounding areas. Acknowledging the limitations in present knowledge and methods pertaining to improving its spatial coverage, this research proposes a new systematic framework that uses one of the most advanced geostatistical interpolation techniques, namely, regression kriging (RK), to estimate continuous RSC between different pairs of existing RWIS stations.This research contains two phases: Phase I first evaluates the feasibility of applying RK to road surface temperature (RST) and road surface index (RSI) estimations. A comparison study using different spatial interpolation methods, including inverse distance weighting, global polynomial interpolation, local polynomial interpolation, and thin plate spline, is conducted to further verify the robustness of the RK method proposed herein. Phase II of the thesis extends the application of the previously developed model in Phase I to estimate RSC using stationary RWIS data only. A sensitivity analysis is also carried out to investigate the influence of RWIS stations density on model performance. Lastly, a recommendation to optimize the RWIS network is introduced by incorporating a renowned combinatorial particle swarm optimization method with the objective of minimizing the total kriging estimation errors.The case study areas are Highways 2 and 16, which are major traffic corridors between Edmonton and Calgary (approximately 300 km) and between Edmonton and Edson (approximately 150 km), respectively. The datasets used in this study are from twelve surveys on four winter nights on Highway 16 and six surveys on two winter nights on Highway 2. Weather events are classified based on the wind speed and snow on ground information to investigate the generalization potential of the models developed herein.The main findings of this thesis are summarized as follows.The findings of Phase I indicate that the kriging models developed in this thesis have a strong predictive ability in estimating road weather and surface conditions, as indicated by low average root mean square errors (RMSE) of 0.254oC and 0.046oC for RST and RSI estimations, respectively. The results also suggest that the RSC estimations can be greatly enhanced with the help of additional covariates included in the models. Furthermore, there exists a strong dependency between the variability in data sets and weather event categories, which can be further used to generalize the findings of this study. The comparison analysis further confirms the robustness of the RK models, whereby improving the accuracy of estimation by up to 50% when compared to other methods. The findings in Phase II of the thesis suggests that the use of stationary RWIS data alone can generate reliable results (i.e., RMSE less than 1oC) when a known semivariogram model is available. The sensitivity analysis also reveals that the increase in the number of RWIS stations will improve the accuracy of estimation until it reaches a certain level, when the magnitude of benefits decreases and stabilizes. Lastly, a proposed RWIS location allocation optimizer is recommended to minimize the total kriging estimation error, for transportation authorities to delineate new site locations for improved monitoring capabilities.The proposed approaches provide a unique opportunity for continuous monitoring and visualization of road weather and surface conditions, to promote more efficient mobilization of winter maintenance resources. It is also anticipated that the findings of this research will, undoubtedly, contribute to improving the overall quality of winter road maintenance services and create a safer and more mobile environment for all travellers.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.013 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".