Renato Lamberti

Crash Prediction Models for Rural Motorways

In this paper, crash prediction models for estimating the safety of rural motorways are presented. Separate models were developed for total crashes and severe (fatal plus all injury) crashes. Generalized linear modeling techniques were used to fit the models, and a negative binomial distribution error structure was assumed. The study used a sample of 2,245 crashes (728 severe crashes) that occurred from 2001 to 2005 on Motorway A16 between Naples and Canosa in Italy. Many characteristics of the motorway were substandard. The motorway allowed investigation of a wide spectrum of geometric configurations. The models were developed by the stepwise-forward procedure with explanatory variables related to traffic volume and composition, horizontal alignment, vertical alignment, design consistency, sight distance, roadside context, cross section, speed limits, and interchange ramps. The decision to keep a variable in the model was based on two criteria. The first was whether the t-ratio of the variables estimated coefficient was significant at the 5% level. The second criterion was related to the improvement of goodness-of-fit measures of the model that includes that variable. Goodness-of-fit measures were the parameter Rα2 and Akaikes information criterion. All the parameters have a logical and expected sign. The most important result was that design consistency measures significantly affected road safety, not only on two-lane rural highways, but also on motorways.

Indirect Skid Resistance Measurement for Porous Asphalt Pavement Management

This paper focuses on developing a method for road managers to assess the sideways-force coefficient (SFC) in porous asphalt with low-cost standard tests. SFC can be used as a component in road surface condition surveys and asset management decisions. A standard piece of equipment for determining SFC is the sideways-force coefficient routine investigation machine (SCRIM), which can be truck or trailer mounted. This equipment is often beyond the budget of most small, low-volume road agencies. An empirical model for the indirect estimation of SFC for porous road surfaces was developed as a result of this study. With handheld equipment—the portable skid resistance tester [to derive the British pendulum number (BPN)] and the sand patch method [height in sand (HS) test]—a correlation of these test results can be made, and a reasonable approximation to SFC with the use of the SCRIM method can be attained. BPN, the unit of measurement of the skid tester, is a representation of the microroughness of the wearing surface, and the sand patch method (HS test) results yield the macroroughness of the pavement. The study was conducted over 20-km segments of low-volume roadway in southern Italy. The initial results are promising, with a maximum percentage of error of less than 15.2%. Further study is needed to adapt the model to other road surface conditions.

Models of Operating Speeds for Low-Volume Roads

Recent studies have documented a noticeable disparity between the speeds for which roads are designed and the actual operating speeds. An important reason for driving errors on two-lane rural roads is the misjudgment of the real course of the road by road users. Unfavorable visibility conditions are one of the problems of road infrastructure that can be detected on existing low-volume roads. Therefore an operating speed model was developed to calculate the contribution made by sight distance, and to be a useful tool for planners in the study of the safety of existing roads. Seven low-volume roads in the province of Salerno, Italy, were studied. The study identified operating speeds on 84 curves; the relative sight distance, radius, radius of the previous curve, length of the tangent before the curve, curvature change rate, development of the curve, and number of speed changes made during the study were determined for each curve. Because of the different characteristics of the roads analyzed, once the data were collected it was possible to develop two operative speed models. The first was suitable for roads characterized by consecutive curves alternating with small tangents and the second model was mainly intended for roads characterized by a small number of curves that were preceded and followed by long tangents.