No abstract

Using GIS to evaluate travel behaviour is an important technique to increase our understanding of the relationship between accessibility and transport demand. In this paper, the activity space concept was used to identify the nature of participation in activities (or lack of it) amongst a group of students using a 2 day travel-activity diary. Three different indicators such as the number of unique locations visited, average daily distance travelled, and average daily activity duration were used to measure the size of activity spaces. These indicators reflect levels of accessibility, personal mobility, and the extent of participation respectively. Multiple regression analyses were used to assess the impacts of students socio-economic status and the spatial characteristics of home location. Although no differences were found in the levels of accessibility and the extent of participation measures, home location with respect to a demand responsive transport (DRT) service was found to be the most important determinant of their mobility patterns. Despite being able to travel longer distances, students who live outside of the DRT service area were found to be temporally excluded from some opportunities. Student activity spaces were also visualised within a GIS environment and a spatial analysis was conducted to underpin the evaluation of the performance of the DRT. This approach was also used to identify the activity spaces of individuals that are geographically excluded from the service. Evaluation of these results indicated that although the service currently covers areas of high demand, 90% of the activity spaces remained un-served by the DRT service. Using this data six new routes were designed to meet the coverage goal of public transport based on a measure of network impedance based on inverse activity density. Following assessment of public transport service coverage, the study was extended using a Spatial Multi Criteria Evaluation (SMCE) technique to assess the effect of service provision on patronage.

Earlier work by the author developed a two-dimensional car-following model by taking into account the lateral discomfort experienced between the neighbouring vehicles moving in the same direction. This incorporation of lateral friction in car-following theory may offer a potential breakthrough in the field of microscopic simulation. However, that model was based on a small-scale scrutiny in which the interactions between only two or three vehicles were analysed in isolation. The treatment of the problem with many vehicles in a stream of traffic, say, on a link, was not made. The present paper, therefore, develops a simulation model with various types of vehicular interactions, including lane changing, and performs certain verification checks. The results show that the simulation behaves rationally giving credibility to the approach introduced. Although a number of comparisons with empirical observations are made in the paper, complete validation and experimentation are left for further work. Furthermore, the paper draws research attention to the area so that various existing car-following models can be modified in a similar way to mimic non-lane-based car following in multilane traffic flow. Potential practical implication of the work for traffic engineers is mainly capacity assessments especially in places where the discipline of lane-based-driving is weak.

Vertical traffic-calming applications may possess certain disadvantages. For instance, drivers sometimes apply a harsh brake very late or they may even prefer to accelerate rather than slow down to by-pass the suspension's shock threshold level. This has negative consequences on both vehicular and pedestrian traffic. In this study, the effects of chicanes – one of the most common horizontal speed reduction applications – on driver behaviour were examined and certain design characteristics were evaluated in the traffic laboratory of Akdeniz University in Turkey. To complete the driving tests safely and in an economical fashion, they were performed in a driving simulator. The trajectory data of some 125 drivers were saved by the simulator software. To examine these data in detail, a new method was developed using VisualLisp programming. The trajectory data were then transferred to an ArcGIS analysis program and the driving test data on the virtual chicanes were evaluated according to position and speed data. Finally, the optimum chicane design characteristics and their efficiency were determined.

The lateral position of vehicles at horizontal curves and roundabout circles has been studied, and it has been observed that drivers tend to straighten their travel path as much as possible when negotiating these bends in a corner-cutting fashion. In addition to skid resistance loss, this behaviour results in the early scuffing and removal of road markings. Preliminary data were collected in Northern Ireland at eight different sites over 15 lanes in total. It was found that the expected geometrical location of the wheel path is not the same as the actual observed location at certain sections of highways. On horizontal curves, the most travelled wheel path is shifted towards the convex side of the curve, and this shift increases with decreasing radius, such as at roundabouts, where it reaches a maximum displacement. This shift was about 140 cm in the outer circles of roundabouts and about 60 cm in the inner circles. For horizontal curves the figure was in the region of 30–40 cm.