A Framework for Level-of-Detail Tracking and Surveillance in Wireless Sensor Networks

Abstract

In the recent years, wireless sensor networks have become a very active research area. Technological advances paved the way for wireless networks being used in many military and civilian applications, including object tracking, security surveillance, habitat monitoring, and traffic control. However, there are intrinsic constraints on design and optimization of sensor networks. Current research mainly focuses on minimizing energy consumption when collecting and transferring data, but improvement is limited by the current sensing paradigm in which users are passive receivers and interpreters of sensory data. A shift is needed from a supplier-initiated, data-centric approach to a consumer-initiated, demand-driven paradigm. This dissertation proposes a Level-Of-Detail (LOD) sensing framework, in which sensing activities are fully determined and adjusted by the user's need for sensory data in real-time. This information-on-demand approach eliminates unnecessary data collection and transfer, and provides an ideal balance between system lifetime, response time, and attention economics. A software test-bed is created for concept proof purpose. The test-bed employs a client-server architecture and allows users to carry out various experiments in an interactive, three-dimensional virtual environment. Experiments results demonstrate substantial reduction in power consumption using the proposed LOD technology.