Tracking and classification with wireless sensor networks and the transferable belief model
Roberts, Matthew Simon
The use of small, cheap, networked devices to collaboratively perform a task presents an attractive opportunity for many scenarios. One such scenario is the tracking and classification of an object moving through a region of interest. A single sensor is capable of very little, but a group of sensors can potentially provide a flexible, self-organising system that can carry out tasks in harsh conditions for long periods of time. This thesis presents a new framework for tracking and classification with a wire less sensor network. Existing algorithms have been integrated and extended within this framework to perform tracking and classification whilst managing energy usage in order to balance the quality of information with the cost of obtaining it. Novel improvements are presented to perform tracking and classification in more realistic scenarios where a target is moving in a non-linear fashion over a varying terrain. The framework presented in this thesis can be used not only in algorithm development, but also as a tool to aid sensor deployment planning. All of the algorithms presented in this thesis have a common basis that results from the integration of a wireless sensor network management algorithm and a tracking and classification algorithm both of which are considered state-of-the-art. Tracking is performed with a particle filter, and classification is performed with the Transferable Belief Model. Simulations are used throughout this thesis in order to compare the performance of different algorithms. A large number of simulations are used in each experiment with various parameter combinations in order to provide a detailed analysis of each algorithm and scenario. The work presented in this thesis could be of use to developers of wireless sensor network algorithms, and also to people who plan the deployment of nodes. This thesis focuses on military scenarios, but the research presented is not limited to this.
75 references, page 1 of 8
views in local repository
downloads in local repository
The information is available from the following content providers: