This thesis studies the transportation network, and is divided into three sections. Initially an algorithm is described which finds the user-equilibrium assignment for networks with linear congestion functions where the cost of travel on a link is dependent on the flow in the whole network. Secondly it investigates the sensitivity of the cost of travel and of the flow distribution in the network, to changes in the link congestion function. Combinatorial methods are used for evaluating the results of the sensitivity analysis. This is done with the aim of obtaining fast and efficient algorithms for the evaluation of cost sensitive and paradoxical links. Finally, for networks where the demand is elastic, it describes the catastrophic behaviour of the point representing the user-equilibrium flow distribution under certain cost conditions.

It is often acknowledged that movement exists in multiple, interdependent forms and that we live in an Information Age. However, mobilities perspectives on contemporary cycling tend to neglect the a) interconnections between transport (physical mobility of people and objects) and communication (mobility of symbolic information) b) paradigmatic shifts in modernity that affect how and why we communicate about transport. This thesis responds to such neglect. Firstly, it places urban cycling in an internet context by examining practices and perceptions of policy blogging, asking why do individuals blog about cycling-related transport policy and to what effect? Secondly, it analyses the answers to these questions through the theoretical lens of the risk society and reflexive modernisation theses. Empirical data is the result of 46 semi-structured interviews with bloggers and expert system representatives, mostly in London, New York and Paris. Blogging about cycling-related transport policy is shown to be an individualised response to the perceived failings of expert systems, as well as in Giddens’ words, a ‘reflexive project of the self’. Citizens who may otherwise only be policy subjects or passive consumers of transport, emerge as policy, media and civil society actors by virtue of their ability to publish information, which forms the basis of social relations. Through blogging, they produce and mobilise knowledge. Knowledge claims mediated by blogging interact with expert systems responsible for transport, which in turn adapt; routine institutional practices evolve; a new order emerges; blogging makes a difference. That difference is however limited, not least because the public remains reliant on expert systems. Ultimately, despite the obvious importance of physical mobility to cycling, this thesis seeks to move beyond it. Information and communication technologies have radically altered how we - researchers, the public, expert system representatives - communicate about and understand cycling, and as such, this project argues for a renewed emphasis on mobilities in a genuinely plural sense of the word as being about more than physically moving from A to B.

The work presented in this dissertation provides a framework for object detection,tracking and vehicle classification in urban environment. The final aim is to produce a system for traffic flow statistics analysis. Based on level set methods and a multi-phase colour model, a general variational formulation which combines Minkowski-form distance L2 and L3 of each channel and their homogenous regions in the index is defined. The active segmentation method successfully finds whole object boundaries which include different known colours, even in very complex background situations, rather than splitting an object into several regions with different colours. For video data supplied by a nominally stationary camera, an adaptive Gaussian mixture model (GMM), with a multi-dimensional Gaussian kernel spatio-temporal smoothing transform, has been used for modeling the distribution of colour image data. The algorithm improves the segmentation performance in adverse imaging conditions. A self-adaptive Gaussian mixture model, with an online dynamical learning rate and global illumination changing factor, is proposed to address the problem of sudden change in illumination. The effectiveness of a state-of-the-art classification algorithm to categorise road vehicles for an urban traffic monitoring system using a set of measurement-based feature (BMF) and a multi-shape descriptor is investigated. Manual vehicle segmentation was used to acquire a large database of labeled vehicles form a set of MBF in combination with pyramid histogram of orientation gradient (PHOG) and edge-based PHOG features. These are used to classify the objects into four main vehicle categories: car, van (van, minivan, minibus and limousine), bus (single and double decked) and motorcycle (motorcycle and bicycle). Then, an automatic system for vehicle detection, tracking and classification from roadside CCTV is presented. The system counts vehicles and separates them into the four categories mentioned above. The GMM and shadow removal method have been used to deal with sudden illumination changes and camera vibration. A Kalman filter tracks a vehicle to enable classification by majority voting over several consecutive frames, and a level set method has been used to refine the foreground blob. Finally, a framework for confidence based active learning for vehicle classification in an urban traffic environment is presented. Only a small number of low confidence samples need to be identified and annotated according to their confidence. Compared to passive learning, the number of annotated samples needed for the training dataset can be reduced significantly, yielding a high accuracy classifier with low computational complexity and high efficiency.

Traffic congestion today is a major problem in almost all of the metropolitan areas of the world. An increasing level of congestion results in negative impacts on the urban environment. These include environmental pollution, energy problems and traffic accidents. The analysis of these problems and the predictions of the impacts of any transport policies that could be devised to deal with them are very critical to their success. Traffic problems are almost the same in most modern cities either in developed countries or less economically developed countries. The driving cycle for a vehicle is the representation of a speed–time sequenced profile, which is developed for a specific area or city. It is an important requirement in the evaluation of the driver's behaviour and the performance of vehicles for a number of applications, mainly in the area of environmental studies. For example, fuel consumption and emissions' predictions need information input on the characteristics of driving patterns of traffic. The applications of driving cycle analysis can be extended however, to many more other areas. The motivation for this research is to investigate the detailed impacts of travel demand management (TDM) measures, that are already in application. This is to improve the network performance, using driving cycle analysis. It is important to explicitly assess these measures using a micro-level detailed approach in order to comprehend overall results in terms of emissions and network performance. These understandings will benefit government agencies and policy makers in their planning and appraisals. It will also benefit public transport providers to improve their service in attracting and retaining their customers. The developments of the real world driving cycles in Edinburgh and Abu Dhabi have been presented in this research. The analysis of real world data, which has been obtained from monitoring traffic conditions in both cities using the GPS tracking of traffic, is presented. This data was collected from trips which have been carried out on a number of traffic corridors in both cities. The assessment of various parameters of traffic (i.e. speed, time percentage spent on acceleration, deceleration, idling, cruising and cycle duration) and their statistical validity, produced a real world driving cycle for the buses as well as the private cars. Two TDM measures have been considered; bus lanes and traffic calming measures. At each corridor, a handheld GPS device was used to record speed, acceleration, deceleration and distances driven. This data enabled the analysis of driving cycles for the buses and for the private cars. The driving cycle analysis and investigations have further been investigated using regression analysis techniques. The results suggest that the approach shows potential but further research is needed with more data available. The results suggest that the driving cycle analysis approach would be very useful to have a better understanding of driving behaviour and also the detailed impacts of the transport policies on traffic. In terms of bus lanes and traffic calming measures, the results show some positive impacts of these policies, while there are evidences of some negative impacts as well. These findings would be very valuable for the policy decision makers. It is recommended from this research that the driving cycle analysis could be utilised effectively in the assessment of TDM measures. Further investigations and analysis of driving cycle is urgently recommended in a number of research directions. Combined GIS and GPS data could also enhance the development in this research.

The thesis seeks to analyse the electric bicycle (e-bike) transition phenomenon in China by applying the Multi-Level Perspective (MLP) Transition Theory and Multi-scalar Perspective MLP. The research is performed with abductive case studies drawn from mixed methods. Firstly, we synthesised secondary data to investigate the e-bike transition at the national level (China) and the city level (“mini” case studies of Beijing and Fuzhou) to explore the research questions of 1) Can socio-technical transition occur without deliberate policy support (RQ1)? 2) How can we explain the rapid emergence and enduring popularity of e-bikes in China (RQ2)? Then we performed exploratory research in Cardiff, UK and Nanjing, China using semi-structured interviews. The interview results are used to help design survey questionnaire in Nanjing case study. It is the key research step and addresses the following research questions: 1) How are e-bikes embedded in the current transport regime (RQ3)? 2) How much longer can e-bikes continue to be embedded in the transport regime (RQ4)? 3) What are the mechanisms underlying the rapid emergence of e-bikes (RQ5)? To analyse the data collected from the survey, Generalised Linear Models and Binomial Generalised Linear Models are adopted to investigate current mode choice behaviour and predict future choice. In the theoretical aspect, the thesis applied the Multi-scalar Perspective MLP, filling the gap that traditional MLP does not take into account geographical, socio-political heterogeneity. In addition, we paid attention to the individual role in travel mode behaviour. From a practical perspective, the thesis uses substantial empirical data to provide a comprehensive understanding of the e-bike transition. We invited various groups in the survey, including e-bike users, bicycle users, car drivers, pedestrians and traffic police. The thesis explores a wide range of influencing factors, such as user anxiety, feelings related to e-bike adoption, e-bike user charging behaviour, other travel mode users and traffic police attitude towards e-bike development which have not been studied in previous e-bike literatures. The main findings are: 1) The fast emergence of e-bikes in China is spontaneous, without direct policy support from governments; 2) E-bike transition in China begins with transformation pathway (P1), followed by de-alignment and re-alignment pathway (P2); 3) E-bike users in China are mainly young career-aged commuters and have a much higher education level than average, which are different from other countries; 4) E-bikes are well embedded in the current transport regime and they are adopted widely in many aspects of people’s daily lives, including commuting, going shopping, and collecting children; 5) The underlying reasons for the selection of e-bikes are they provide affordable personal mobility due to the advantages of effort saving, flexible trip times, time saving in traffic jams, and high accessibility, whereas environmental and health factors are negligible; 6) E-bikes are possibly an intermediate mode on Nanjing’s motorisation pathway.

The aim of this research thesis was to understand how the urban built environment interacts with utility cycling behaviours in space and time. Using mixed methods entrenched in the philosophy of pragmatism, the research contributed to an as yet under-developed research evidence-base within the British context by addressing the gap in knowledge relating to: the usability of spacetime and action space theories within visual analytics context in facilitating the knowledge discovery process from spatio-temporal datasets; empirical evidence on perceived and actual movement behaviour of urban utility cyclists; methodological advancement in collecting, refining, analysing and visualising detailed utility cycling behaviours in a British urban environment. Findings suggest that 57.4% of cyclists’ bike trips were found on the cycle network and with 42.6% of cyclists still cycling outside the designated cycle network; it is therefore imperative that policy initiatives aimed towards strategic investment in cycling behavioural research and infrastructure. The findings also showed a higher concentration of cycling uptake around the south-eastern part of Newcastle upon Tyne suggesting this area may need more investment than other areas in Tyne and Wear. Systematic comparison of GPS data and travel diary data suggest 8.4% under reporting of the former. The null hypothesis that urban transport network restrictions do not have any significant influence on movement of commuter cyclists was rejected upon examination and it was found that observed routes tend to be significantly longer than their shortest path counterparts. Profiling activity spaces of utility cyclists utilising different geographies was found to be useful in the examination of cycling behaviours for the purpose of providing visual aid for planners and policy makers to identify areas for improvement and informed investment in support of sustainable transport. Several efforts were being made to enhance data availability to inform policy strategies, and facilitation of feasible solutions for improving the urban cycling infrastructure and encouraging more people to cycle as part of their daily commute, for which this research aimed to contribute by providing evidence on the use of the area’s cycling infrastructure by utility cyclists and spatial variability of cycling in space and time.

This PhD investigates driver behavioural adaptation to in-vehicle navigation systems (IVNS). Behavioural adaptation is receiving an increasing amount of research attention in traffic psychology, but few studies have directly considered the concept in relation to IVNS. The thesis aims were addressed using a range of quantitative and qualitative methodologies. Using two online surveys, over 1300 drivers (including over 1000 IVNS users) were sampled, to identify a range of positive, neutral and negative aspects of end-user behavioural adaptation to IVNS in terms of both safety and navigational efficiency. The first survey (N=450) aimed at drivers in general, showed that IVNS users believe they commit some common driving errors (e.g. misreading signs when leaving a roundabout) significantly less frequently than ordinary drivers who do not use these systems, but that they also feel they drive without fully attending to the road ahead significantly more frequently. The second survey (N=872) was aimed at IVNS users only, and further explored distracted driving. This survey found that the majority of IVNS users have interacted with their system while driving (e.g. to enter a destination), and that some do so frequently. It also showed that system reliability is a key issue affecting most current IVNS users, revealing that some drivers have followed inaccurate as well as illegal and potentially dangerous, system-generated route guidance information in a range of different contexts. A longitudinal diary study (N=20) then collected rich qualitative data from a sample of worker drivers who regularly used their IVNS in unfamiliar areas. The data collected illustrated the diverse contexts in which drivers experience aspects of behavioural adaptation to IVNS identified in the surveys. Both the IVNS user-survey and diary study also identified key demographic individual difference variables (most notably age and computing skill) that were associated with the extent to which driver’s experienced different manifestations of behavioural adaptation to IVNS. Moreover, other individual difference variables (e.g. complacency potential, system-trust, confidence) were found to be associated with more specific behavioural adaptations. Two simulator studies investigated system interaction while driving. The first (N=24) demonstrated the poor degree of correspondence between drivers’ perceptions of driving performance when entering destinations while driving (relative to normal driving) and objective performance differences between these conditions. The second simulator study (N=24) showed that safety and training based interventions designed to reduce the extent to which drivers use IVNS while driving or to improve their performance if they do had only a modest effect on dependent measures. This thesis represents the first attempt in the literature to bring together research from diverse areas of human factors and traffic psychology to consider behavioural adaptation to in-vehicle navigation systems. By associating a range of these issues with behavioural adaptation to IVNS, it has indirectly increased the scope of several salient, previous research findings. Moreover, by investigating many of these issues in depth, using both quantitative and qualitative methodological approaches, it has set the foundation for future work. Such work should aim to explore many of the issues raised, and develop effective remediating or mitigating intervention strategies for negative behavioural adaptations that could adversely affect driving safety, as well as to encourage and support those which may be considered more positive.

Traditionally, Vehicle Routing Problems (VRPs) are modelled with fixed traversal times. The amount of time it takes to drive from one end of a road to the other is unchanged throughout the day. Nearly always, the reality of the situation that is being modelled is very different, with road speeds varying heavily, especially with “rush hour" traffic. Modelling VRPs with time varying congestion means that even slight changes early in a vehicle tour can have major knock-on effects that are hard to predict. Recalculating the total traversal time of vehicles whenever their tours are changed drastically increases metaheuristic calculation times compared to non-time varying models. In this thesis we use a simple technique of calculating the localised change and inferring the global effects resulting from neighbourhood moves. Only if the localised change suggests that the global result is satisfactory do we then calculate the actual global result. Inevitably using these estimates does not give as accurate results as always calculating the changes, but we aim to show that the loss of solution quality is overshadowed by the significant savings in calculation time. We present a series of experiments comparing simple metaheuristics with and without using estimates and show consistent savings in calculation time whenever estimates are used compared to when they are not. These savings shown to increase as the size of the problem (in terms of the number of customers) increases. In addition to synthetic problems, we also present a problem based on real world vehicle traversal times and show that our estimates prove just as accurate, if not more so, at retaining solution quality as the synthetic methods. Lastly, we briefly discuss further methods of solving VRPs that could also benefit from our work here.

The period between the two World Wars, 1919 to 1939, saw a major expansion in the use of motor vehicles on the roads of Britain and a consequent increase in road accidents and fatalities. Studies of road transport in this period have concentrated largely on the expanding urban areas. Little detailed work has been completed in rural areas. This study goes some way to fill this gap in our knowledge. It will examine the rural county of Herefordshire where the use of motor vehicles did increase greatly, although not to the extent of the country as a whole. This thesis will address the management of the safety of road traffic which was exercised by the County Council in the inter war years. The heart of the study is the evaluation of the size of the problem and the efforts of the County Surveyor to improve the safety of the county road network. Increasing use of the road network was encouraged by the decline in use of the rail network. The A40 trunk road in the south of the county saw traffic increases of approximately one and a half times over the two decades. However the secondary roads such as the B4214 north of Bromyard saw much greater increase. Cars increased by six and three quarter times whilst lorries by the even larger margin of nine times over the two decades. The improvements carried out to the county road network consisted of road widening, rather than major road realignments. The County Council members were strongly resistant to the loss of agricultural land. However significant new improvements to the road surface were effected throughout the county albeit in short lengths. Four different types of road surface were tried. In fact the county was at the forefront of road surface development in this period. The County Surveyor’s preferred choice was a bitumen/stone mix laid cold which became the accepted solution for the country as a whole in the 1930s. The attitude adopted by the police force, essentially reactive, in the management of road safety will be examined. Whilst traffic volumes increased in the inter-war years there was no corresponding increase in the 1930s in the recorded number of fatal road accidents. The thesis will conclude with an assessment of the degree of public satisfaction with the efforts of the County Surveyor and the Chief Constable in maintaining a safe county road network in the inter-war years. The public, through the medium of the press and the County Council minutes, indicated that they considered that the efforts of the Surveyor and Chief Constable were acceptable.

The countersteering effect in motorcycles describes the apparent need to steer in the wrong direction in order to cause the chassis of the vehicle to lean over in the required direction just prior to executing a turn. The inherent danger with this procedure is that in a emergency situation where a motorcyclist must execute a sudden swerve to avoid a collision, the required behaviour is counterintuitive and panic may cause the rider to make the wrong initial movement thereby reducing their chance of avoiding a collision. As the importance of the countersteering effect is not taught in UK motorcycle training courses, the current work has attempted to establish whether doing so could significantly improve the ability of riders in swerve to avoid manoeuvres. An initial survey of motorcycle riders suggested some confusion about the nature of countersteering. To explore this further, four groups of riders with different levels of experience and training: novice, experienced, advanced and expert, were tested over a simple swerve to avoid course that was based on the procedure in the current UK motorcycle test. All the riders used the same motorcycle with on-board instrumentation to record the steering effort and the response of the machine. The tests were also videoed to gain extra information about rider behaviour. The results suggest that those riders that had been trained in exploiting countersteering were better able to avoid the obstacle and significantly better at returning the machine to the desired path thereby avoiding a potential secondary collision. It appeared that those riders who had learned by experience were still not proficient when faced with a sudden swerve to avoid scenario.