Kaza-olay algılama algoritmalarının mikroskobik benzetim ile modellenmesi ve değerlendirilmesi: D 100 Karayolu örneği, İstanbul

Year 2018, Volume: 1 Issue: 2, 87 - 107, 30.10.2018

İlgin Gökaşar

Abstract

Üç kaza tespit algoritması (Kaliforniya, Minesota
ve SNS algoritmaları) D 100 Karayolunun 2.5km'lik bir kısmında PARAMICS
mikroskobik benzetim yazılımıyla modellenmiştir. PARAMICS modeli ile
değerlendirilen üç algoritmanın da Istanbul trafiği için çok uygun olduğu
gösterilmiştir. Kaliforniya, Minesota ve SNS Algoritmaları orta trafik
talebinde DR değerlerini sırasıyla %25, %31,5 ve %25 yükselmiştir. MTTD
değerleri orta trafik talebinde sırasıyla 48.99%, 59.26% ve 48.80% azalmıştır. Minesota
Algoritması en düşük MTTD (0.88) değerinde çok iyi performans göstermektedir.
Bunun yanında, SNS Algoritması düşük trafik düzeyinde sadece 1 yanlış alarm
verilmiştir. Tüm test edilmiş algoritmalar, dedektörler arası eşit ve 800m´den
az olduğunda Trafik Kontrol Merkezinden alınan gerçek zamanlı trafik verileri
kullanılarak uygulanabilir.

Keywords

Kaza-Olay Algılama Algoritmaları, Mikroskobik benzetim

Modeling and Evaluation Incident Detection Algorithms using Microscopic Simulation: A Case of D 100 Highway, Istanbul

Abstract

Three incident detection algorithms (California, Minnesota and SNS) have been tested on 2.5km-length section of D100 Highway using PARAMICS microscopic simulation software. PARAMICS model evaluation showed that all three algorithms are well suited for the Istanbul traffic. California, Minnesota and SNS Algorithms are better in medium traffic demand since detection rate values are increased by 25%, 31.5% and 25%, respectively. Mean time to detect (MTTD) values are also good in medium traffic demand in that they decrease by 48.99%, 59.26% and 48.80%, respectively.  Actually, Minnesota Algorithm had superior performance with the lowest MTTD value (0.88 min). In addition, there is only one false alarm generated by SNS Algorithm in low traffic demand. All of the tested algorithms when properly calibrated can be implemented using real time traffic data provided that distances between the detectors are equally spaced and smaller than 800 m by the traffic control centers.

Keywords

Incident detection algorithms, Microscopic simulation

References

• [1]. Tina Collier, Ginger Goodin, Managed Lanes: A Cross-Cutting Study, No. FHWA-HOP-05-037, USA, 2004.
• [2]. Kaan Ozbay, Pushkin Kachroo, Incident Management In Intelligent Transportation Systems, Electrıcal and Computer Engineerıng Faculty Publications, USA, 1999.
• [3]. Payne, H. J; Helfenbein, E. D; Knobel, H. C, Development and testing of incident detection algorithms, volume 2: Research methodology and detailed results, Report No. FHWA-RD- 76-20, USA, 1976.
• [4]. Payne, H. J; Tignor, S. C, Freeway Incident-Detection Algorithms Based On Decision Trees With States. TRB, National Research Council, 1978, pp. 30-37.
• [5]. Mak, Chin Long; Henry SL Fan, Transferability of expressway incident detection algorithms to Singapore and Melbourne. Journal of Transportation Engineering 2005, 31, 2, 101-111.
• [6]. Mountain Plains Consortium (MPC), Evaluation of Incident Detection Algorithms Report No: MPC-01-122, USA, 2011.
• [7]. Paramics Traffic Microsimulation Software Manual Version 6.7.2., USA, 2011.
• [8]. Chassiakos, A. P; Stephanedes, Y. J, Smoothing algorithms for incident detection. Transportation Research Record. 1993, 1360, 50-57.