Lateral Placement Behaviour along Horizontal Curve Sections under Mixed Traffic: Evidence from North Easter Hill range of India

Abstract

Where a vehicle sits across the carriageway on a curve is closely tied to how safely and how efficiently that curve  performs. This paper examines that behaviour on a four-lane median-divided hill highway. Lateral placement was  measured at three gates, namely the entry, centre and exit, on ten horizontal curves of the Itanagar–Naharlagun section  of NH-415, yielding 26,898 vehicle observations. Each gate was analysed on its own, and the spread of placement at  each was fitted to a scaled Beta model built from the measured mean and variance. The section effect was then tested  twice. Averaged over the ten curves, the Friedman test returned nothing for any class; such averaging washes out the  very within-curve movement that the study set out to capture. The individual vehicles told a different story. With the  same vehicle followed through all three gates, a matched Friedman test made the section effect highly significant for  every dominant class, namely car, SUV, motorcycle and scooter, with p below 0.001 throughout. Its shape did not  change from one class to the next. Vehicles ran widest near the centre and drew closest to the median at the exit, a  swing of barely 0.2 to 0.4 m. The section therefore shifts placement in a reliable but small way, and accounts for only  a modest part of the overall spread. Valley-side and hill-side curves also differed, most sharply at the entry, although  the roadside type here cannot be separated from the direction of travel. Read together, the results make the case for  treating the curve section, rather than the whole curve, as the unit of design and maintenance attention on  heterogeneous hill corridors.