Little has been known about traffic safety at a disaggregate level such as accidents on a road. Much of the work today has been done at an aggregate level, which compares accident rates for different cities, countries, or regions. This means that the available traffic safety indicators are not appropriate to describe the situations at a disaggregate level. Therefore, there is need to develop new techniques of obtaining a better understanding of the traffic safety situation on a given type of a road. The limited available work on traffic safety models at a disaggregate level has been established for situations of the developed world. Such results cannot be applied to situations in developing countries due to the differences in the behaviour of both traffic systems. Under the current research, an attempt is being made to determine ow some engineering measures affect levels of safety on a given road. Traffic mobility, which the engineer can influence, is used as a measure of exposure to risk of accident casualties on a given road corridor in developing countries. Uganda is used as a case study. Relationships between traffic safety and traffic mobility were developed by use of person-through put concept as a measure of traffic mobility. This parameter is a measure of person-travel in a given space and time. It incorporates as many independent variables affecting traffic safety as possible, thereby reducing the complex problem of multivariate analysis. Traffic mobility on a road stretch of 19Km in Uganda was quantified using UTEM model. This road corridor has two main sections. One section has a total of two lanes whereas the other has three lanes. In a four-year period from 1996 to 1999 there were about 180 fatalities, 530 serious injuries and 230 slight injuries that occurred between 0700 - 2200 hours on this stretch of the road. Regression analysis techniques were used to develop relationships between road accident casualties and traffic mobility based on this time period of four years. The results suggest that, for heterogeneous traffic conditions along study road corridor, total fatalities seem to be significantly related to person throughput. Total serious injuries seem to be significantly related to only the average travel speeds whereas total slight injuries seem to be significantly related to both the average travel speeds and the proportion of passenger service vehicles. The injury models developed predict that slow speeds contribute to accident frequency, average-to-slightly high speeds reduce frequency, and very high speed increases frequency. Both serious and slight injury rates are seemingly lowest at an average travel speed of 47Km/h. Thus an average travel speed of 47Km/h may be the most efficient level of service in terms of traffic safety along the study road corridor. This level of service was simulated using the UTEM model by: (i) Restricting the speed limits of all vehicles in traffic mix to 80Km/h and (ii) Installation of traffic signals with appropriate traffic control schemes. (A)
Samenvatting