Some factors affecting journey-times on urban roads
This thesis discusses the results of two types of study, (a) of the relationships between journey-time and traffic flow during peak periods on two sections of urban road, each about 0.4 miles in length, and (b) of relationships between journey-times and a number of land-use variables over about 50 miles of suburban main roads in off-peak periods. The results of study (a) show that the most significant relationships occurred when the data was analysed using non-linear or two-regime models. In the latter case the two regimes were separated by a critical flow. Above the critical flow changes in flow were correlated with changes in journey-time and below the critical flow the changes appeared to be random. Further evidence is provided to show that the use of different sampling intervals can give rise to different empirical relationships with the same data. A queuing model postulated by Davidson (1968) has been shown to give a satisfactory agreement with the results of study (a). A model is developed in this thesis which accounts for consistent changes in saturation flow arising from the use of different sampling intervals in Davidson's model. Additionally, an empirical index has been derived which reflects the changes in activity during the peak hour, and journey-time is highly correlated with this index. Study (b) shows that after the effects of major intersections have been accounted for, journey-time was significantly correlated with a number of factors of land-use along the suburban main roads which were studied. Furthermore, a number of mathematical techniques have been used which permit the calculation of indices of the variation of the surrounding conditions and these techniques were shown to produce robust and repeatable results. Journey-time was significantly correlated with an Activity Index which was associated with a number of land-use parameters.