No description available.

No description available.

No description available.

No description available.

No description available.

Traditional network design problems only consider the long-term stationary travel patterns (e.g., fixed OD demand) and short-term variations of human mobility are ignored. This study aims to integrate human mobility characteristics and travel patterns into network design problems using a space-time network structure. Emerging technologies such as location-based social network platforms provide a unique opportunity for understanding human mobility patterns that can lead to advanced modeling techniques. To reach our goal, at first multimodal network design problems are investigated by considering safety and flow interactions between different modes of transport. We develop a network reconstruction method to expand a single-modal transportation network to a multi-modal network where flow interactions between different modes can be quantified. Then, in our second task, we investigate the trajectory of moving objects to see how they can reveal detailed information about human travel characteristics and presence probability with high-resolution detail. A time geography-based methodology is proposed to not only estimate an individual’s space-time trajectory based on his/her limited space-time sample points but also to quantify the accuracy of this estimation in a robust manner. A series of measures including activity bandwidth and normalized activity bandwidth are proposed to quantify the accuracy of trajectory estimation, and cutoff points are suggested for screening data records for mobility analysis. Finally, a space-time network-based modeling framework is proposed to integrate human mobility into network design problems. We construct a probabilistic network structure to quantify human’s presence probability at different locations and time. Then, a Mixed Integer Nonlinear Programming (MINLP) model is proposed to maximize the spatial and temporal coverage of individual targets. To achieve near optimal solutions for large-scale problems, greedy heuristic, Lagrangian relaxation and simulated annealing algorithms are implemented to solve the problem. The proposed algorithms are implemented on hypothetical and real world numerical examples to demonstrate the performance and effectiveness of the methodology on different network sizes and promising results have been obtained.

The WITAS project aims to develop technologies to enable an Unmanned Airial Vehicle (UAV) to operate autonomously and intelligently, in applications such as traffic surveillance and remote photogrammetry. Many of the necessary control and reasoning tasks, e.g. state estimation, reidentification, planning and diagnosis, involve prediction as an important component. Prediction relies on models, and such models can take a variety of forms. Model design involves many choices with many alternatives for each choice, and each alternative carries advantages and disadvantages that may be far from obvious. In spite of this, and of the important role of prediction in so many areas, the problem of predictive model design is rarely studied on its own. In this thesis, we examine a range of applications involving prediction and try to extract a set of choices and alternatives for model design. As a case study, we then develop, evaluate and compare two different model designs for a specific prediction problem encountered in the WITAS UAV project. The problem is to predict the movements of a vehicle travelling in a traffic network. The main difficulty is that uncertainty in predictions is very high, du to two factors: predictions have to be made on a relatively large time scale, and we have very little information about the specific vehicle in question. To counter uncertainty, as much use as possible must be made of knowledge about traffic in general, which puts emphasis on the knowledge representation aspect of the predictive model design. The two mode design we develop differ mainly in how they represent uncertainty: the first uses coarse, schema-based representation of likelihood, while the second, a Markov model, uses probability. Preliminary experiments indicate that the second design has better computational properties, but also some drawbacks: model construction is data intensive and the resulting models are somewhat opaque. / <p>Report code: LiU-Tek-Lic-2002:15.</p>

No description available.

This thesis presents a framework for motor vehicle detection-tracking surveillance systems. Given an optimized object detection template, the feasibility and effectiveness of the methodology is considered for vehicle counting applications, implementing both a filtering operation of false detection, based on the speed variability in each segment of traffic state, and an occlusion handling technique which considers the unusual affine transformation of tracking subspace, as well as its highly fluctuating averaged acceleration data. The result presents the overall performance considering the trade-off relationship between true detection rate and false detection rate. The filtering operation achieved significant success in removing the majority of non-vehicle elements that do not move like a vehicle. The occlusion handling technique employed also improved the systems performance, contributing counts that would otherwise be lost. For all video samples tested, the proposed framework obtained high correct count (>93% correct counting rate) while simultaneously minimizing the false count rate. For future research, the author recommends the use of more sophisticated filters for specific sets of conditions as well as the implementation of discriminative classifier for detecting different occlusion cases.

This master's thesis focuses on 3D reconstruction of vehicles passing in front of a traffic surveillance camera. Calibration process of surveillance camera is first introduced and the relation of automatic calibration with 3D information about observed traffic is described. Furthermore, Structure from Motion, SLAM, and optical flow algorithms are presented. A set of experiments with feature matching and the Structure from Motion algorithm is carried out to examine results on images of passing vehicles. Afterwards, the Structure from Motion pipeline is modified. Instead of using SIFT features, DeepMatching algorithm is utilized to obtain quasi-dense point correspondences for the subsequent reconstruction phase. Afterwards, reconstructed models are refined by applying additional constraints specific to the vehicle reconstruction task. The resultant models are then evaluated. Lastly, observations and acquired information about the process of vehicle reconstruction are utilized to form proposals for prospective design of an entirely custom pipeline that would be specialized for 3D reconstruction of passing vehicles.