Derin Öğrenme Yöntemi Kullanılarak Meteorolojik Parametrelere Dayalı Karayolu Kaplama Yüzey Durumunun Tahmini

Abstract

Karayollarında yol kaplama yüzeyinin durumu trafik güvenliğinin sağlanmasında önemli bir faktördür. Yolun bulunduğu iklim koşullarına göre yol kaplamalarının durumu değişkenlik göstermektedir. Yol kaplamalarının meteorolojik faktörlere göre değişkenlik durumunu kayıt altına almak için hem yol kaplama içerisine yerleştirilen sensorler hem de yol kenarlarına yol meteoroloji bilgi istasyonları kurulmaktadır. Akıllı ulaşım sistemleri kapsamında yol yönetim bilgi sistemlerinin kurulması ve sensörlerden alınan veriler ile gerçek zamanlı yol kaplamasının durumu gözlenebilmektedir. Bu sensor verileri ile yol yüzey durumu farklı yapay zeka yöntemleri ile tahmin edilebilmektedir. Böylece yol yüzey durumunun kuru, ıslak ve buzlu olmasına göre önlemlerin alınmasında karar vericiler için önemli bilgiler sunulmaktadır. Bu çalışmada, meterolojik parametelere bağlı yol yüzey durumu tahmin edilmesi amaçlanmıştır. Bu amaçla, derin öğrenme modelleri geliştirilmiştir. 65966 adet veriseti içerisinde hava sıcaklığı, çiğ noktası sıcaklığı, rüzgar hızı, rüzgar yönü, esinti hızı , kaplama sensör sıcaklığı ve kaplama sensör durumu parametreleri kullanılmıştır. Derin öğrenme modellerinin değerlendirilmesinde accuracy kullanılmıştır. Yapılan değerlendirme sonucunda en iyi modelin accuracy değeri 0.88 olarak belirlenmiştir. Ayrıca en iyi modelin test seti için her bir sınıfa ait accuracy, recall, presicion ve F1-score değerleri hesaplanmıştır.

Keywords

• Akıllı ulaşım sistemleri
• yapay zeka
• derin öğrenme
• karayolu üstyapı
• meteorolojik parametreler

Prediction of Highway Pavement Surface Condition Based on Meteorological Parameters Using Deep Learning Method

Abstract

The condition of the pavement surface on highways is an important factor in ensuring traffic safety. The condition of the road pavements varies according to the climatic conditions of the road. To record the variability of road pavements according to meteorological factors, both sensors placed in the pavement and road meteorology information stations are installed on the roadsides. Within the scope of intelligent transportation systems, the establishment of road management information systems and the status of the road pavement in real-time can be observed with the data obtained from the sensors. With these sensor data, the road surface condition can be estimated with different artificial intelligence methods. Thus, important information is provided for decision-makers in taking precautions according to the dry, wet, and icy road surface condition. In this study, it is purposed to estimate the road surface condition based on meteorological parameters. For this purpose, deep learning models have been developed. Air temperature (tmp), dew point temperature (dwp), wind speed (sknt), wind direction (drct), wind gust (gust), pavement sensor temperature (tfs), and pavement sensor condition (cond) parameters were used in 65966 datasets. Accuracy was used in the evaluation of deep learning models. Consequently, the evaluation, the accuracy value of the best model was determined as 0.88. In addition, accuracy, recall, precision, and F1-score values of each class were calculated for the test set of the best model.

Keywords

• Intelligent transportation systems
• artificial intelligence
• deep learning
• Highway pavement
• meteorological parameters

References

1. Alqudah, Y. A. & Sababha, B. H. (2017). On the Analysis of Road Surface Conditions Using Embedded Smartphone Sensors. In 2017 8th International Conference on Information and Communication Systems (ICICS), pp. 177-181.
2. Bouilloud, L., Martin, E., Habets, F., Boone, A., Le Moigne, P., Livet, J., Marchetti, M., Foidart, A., Franchistéguy, L., Morel, S., J. Noilhan & Pettré, P. (2009). Road Surface Condition Forecasting in France. Journal of Applied Meteorology and Climatology, 48(12), 2513-2527.
3. Chen, Y., Lin, Z., Zhao, X., Wang, G. & Gu, Y. (2014). Deep Learning-Based Classification of Hyperspectral Data. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 7(6), 2094-2107.
4. Chicco, D. & Jurman, G. (2020). The Advantages of The Matthews Correlation Coefficient (Mcc) Over F1 Score and Accuracy in Binary Classification Evaluation. BMC Genomics, 21(1), 1-13.
5. Dogan, F. & Turkoglu, İ. (2018). The Comparison of Leaf Classification Performance of Deep Learning Algorithms. Sakarya University Journal of Computer and Information Sciences, 1(1), 10-21 (in Turkish).
6. Dudak, J., Gaspar, G., Sedivy, S., Pepucha, L., & Florkova, Z. (2017). Road Structural Elements Temperature Trends Diagnostics Using Sensory System of Own Design. In IOP Conference Series: Materials Science and Engineering, pp. 012036.
7. Kocianova, A. (2015). The Intelligent Winter Road Maintenance Management in Slovak Conditions. Procedia Engineering, 111, 410-419.
8. Krsmanc, R., Slak, A. S., & Demsar, J. (2013). Statistical Approach for Forecasting Road Surface Temperature. Meteorological Applications, 20(4), 439-446.

9. Li, Y., Chen, J., Dan, H. & Wang, H. (2022). Probability Prediction of Pavement Surface Low Temperature in Winter Based on Bayesian Structural Time Series and Neural Network. Cold Regions Science and Technology, 194, 103434.
10. Liu, B., Yan, S., You, H., Dong, Y., Li, Y., Lang, J. & Gu, R. (2018). Road Surface Temperature Prediction Based on Gradient Extreme Learning Machine Boosting. Computers in Industry, 99, 294-302.
11. Milad, A., Adwan, I., Majeed, S. A., Yusoff, N. I. M., Al-Ansari, N. & Yaseen, Z. M. (2021). Emerging Technologies of Deep Learning Models Development for Pavement Temperature Prediction. IEEE Access, 9, 23840-23849.
12. Molavi Nojumi, M., Huang, Y., Hashemian, L. & Bayat, A. (2022). Application of Machine Learning for Temperature Prediction in a Test Road in Alberta. International Journal of Pavement Research and Technology, 15(2), 303-319.
13. Rahim, M. A. & Hassan, H. M. (2021). A Deep Learning-Based Traffic Crash Severity Prediction Framework. Accident Analysis & Prevention, 154, 106090.
14. Sofaer, H. R., Hoeting, J. A. & Jarnevich, C. S. (2019). The Area Under the Precision‐Recall Curve as A Performance Metric for Rare Binary Events. Methods in Ecology and Evolution, 10(4), 565-577.
15. URL-1. [Online], Available: (2022, Feb 22).
16. Xu, B., Dan, H. C. & Li, L. (2017). Temperature Prediction Model of Asphalt Pavement in Cold Regions Based on an Improved BP Neural Network. Applied Thermal Engineering, 120, 568-580.
17. Yang, C. H., Yun, D. G., Kim, J. G., Lee, G. & Kim, S. B. (2020). Machine Learning Approaches to Estimate Road Surface Temperature Variation along Road Section in Real-Time for Winter Operation. International Journal of Intelligent Transportation Systems Research, 18(2), 343-355.