A Roving-Object modelling framework for location-tracking applications

This thesis introduces the concept of Roving Object (RO) modelling as a means of managing the uncertainty in the location tracking of human moving objects travelling on a network. The term RO is adopted to refer to human or human-controlled moving point objects. For previous movements of the ROs, the uncertainty stems from the discrete nature of location tracking systems, where gaps are created among the location reports. Future locations of ROs are, by definition, uncertain. The objective in this thesis is to maximize the estimation accuracy while minimizing the operating costs. Humans are creatures of habit. Knowing the routes taken by ROs in and their speeds in different contexts facilitates building RO models for estimating future routes and speeds in similar contexts. The most probable speed and route, indicated by such models is then applied to estimate the RO locations in the future and between location reports in the past with a high degree of certainty. The newly proposed RO model consists of two components; the speed model and the route model. The speed model captures typical RO speeds under different driving conditions on different road types during different times of the day and days of the week, and in different areas of the network. In this thesis Bayesian Networks (BNs) are adopted as the modelling tool. The route RO model captures the typical route taken between each source and destination pair. The model relies on a simple probabilistic approach to represent the most probable route, based on the previous trips. The routes are represented as a series of road segments. A spatiotemporal access method is developed to utilize an R-tree that is constructed by using only static objects such as road segments. The temporal dimension is divided into time slots where each has its own array of hash tables. Each RO entry in the hash table points to the next and previous time slot array elements where the object resides. To test the novel approach, the ROving Objects Trip Simulator (ROOTS) is introduced to create ROs with distinct characteristics in terms of driving style and route preference.