Use of Kriging Technique to Study Roundabout Performance

Road intersections are dangerous places because of the many conflicting points between motorized and nonmotorized vehicles. In the case of defined traffic volume, several research groups have proved that roundabouts reduced the number of injuries and fatal accident cases. In recent years, many countries have adopted roundabouts as a standard design solution for both urban and rural roads. Several recent studies have investigated the performance of roundabouts, including some with models that calculated the entering flow (Q sub e) as a function of the circulating flow (Q sub c). Most existing models have been constructed with the use of linear or exponential statistical regression. The interpolative techniques in classical statistics are based on the use of canonical forms (linear or polynomial) that completely ignore the correlation law between collected data. As such, the determined interpolation stems from the assumption that the data represent a random sample. In the research reported in this paper, a geostatistical approach was considered: the relationship Q sub e versus Q sub c is supposed to be a regionalized phenomenon. According to that supposition, collected data do not represent a random sample of values but are supposed to be related to each other with a defined law. This recognition allows the realization of interpolation on the basis of the real law of the phenomenon. This paper discusses the fundamental theories, the applied operating procedures, and the first results obtained in modeling the Q sub e versus Q sub c relationship with the application of geostatistics.