Continuous objects detection and tracking in wireless sensor network

A considerable amount of research is done in the area of target detection and tracking in wireless sensor networks. Most of the research concentrates on a single or multiple targets tracking but limited work has been done in tracking and detection of the continuous objects such as forest fires, bio- chemical materials and mudflows etc. These continuous objects pose new challenges because of their specific nature and characteristics such as changing in size and shape, shrinking and expanding, splitting into multiple objects or merging of multiple objects into one another. Continuous objects tracking and detection require extensive communication which consumes most of the energy of the network and the key issue of today's wireless sensor networks is the energy handling, which is the backbone for enhancing the lifetime of the network. In this thesis a new algorithm known as Continuous Object Detection and Tracking (CODAT) is proposed. CODAT's performance is an improvement in accuracy and efficiency while comparing it with the two known algorithms Continuous Boundary Monitoring (COBOM) and Detection and Monitoring for Continuous Objects (DEMOCO). New data structure for reporting data is introduced. Due to this new data structure, the reported data is sent to the sink more efficiently, which eventually reduces the communication cost of the overall algorithm without compromising the accuracy for reconstructing the boundary of a continuous object at the base station. A concept for differentiating between the holes in the phenomenon and overall phenomenon changes at the base station level is also introduced which provides additional information to the user at base station level as an added improvement while maintaining the high accuracy and efficiency. The proposed algorithm efficiency is analyzed by simulations. Simulation results show that CODAT outperforms COBOM in terms of efficiency and outperforms COBOM and DEMOCO in terms of accuracy.