System Reliability Evaluation of Expressway Horizontal Alignment Design Considering Trucks and Passenger Cars

• Published in: Transportation research record Vol. 2679; no. 5; pp. 836 - 848
• Main Authors: Ramesh, Anjana G., Goyani, Jaydip, Arkatkar, Shriniwas, Easa, Said M.
• Format: Journal Article
• Language: English
• Published: Los Angeles, CA SAGE Publications 01.05.2025
• Subjects: horizontal alignment; roadway design; geometric design; safety; speed
• ISSN: 0361-1981; 2169-4052
• DOI: 10.1177/03611981241310128

The design of safe highway alignments is a complex task, with horizontal curves being the most critical components from both safety and operational perspectives. This is because, from the tangent to curve transitions, maintaining a constant operating speed in line with driver expectations rather than the designer’s judgment is essential from a safety standpoint. The relationship between the highway elements and static/dynamic characteristics of the various vehicles on the highway geometry motivates the development of reliable safety strategies for the curve design and safety evaluation. In this study, we selected passenger cars (PCs) and trucks as non-compliant drivers and passenger taxis (PTs) as compliant drivers to evaluate highway safety. We chose 22 horizontal curves from the Ghat region of India’s Mumbai to Pune Expressway. On selected curves, 5940 samples of vehicle spot speeds using a radar gun for PCs/trucks and data from 16 PTs using an e-tracker were collected. We then investigated a stopping sight distance and speed-based reliability framework for assessing horizontal curve design at the curve and alignment system levels. The results show that PCs in non-compliant conditions operate faster than PTs in compliant and trucks in non-compliant conditions. Therefore, the curve has a higher probability of non-compliance, Pnc, (lower reliability) for PCs in non-compliant conditions, followed by PTs in compliant and trucks in non-compliant conditions. The higher Pnc indicates more chances of crashes on the subject curves. Subsequently, we conducted a sensitivity analysis for the curve radius ranging from 400 to 900 m. The results show that the mid-ordinates must be changed to obtain a lower Pnc by offering a longer sight distance.