Wireless sensor networks comprise an emerging technology that has found many applications in different areas. Area surveillance is one of the most important applications of wireless sensor networks. Area surveillance consists of detection, classification, and tracking of intruding objects in an area. Some algorithms are proposed in the literature to use wireless sensor networks in this application. Since the main constraint on sensor nodes is the energy supply, existing methods attempt to achieve an acceptable level of accuracy along with minimum energy consumption to maximize the lifetime of the sensor network. In this thesis, new algorithms for tracking and classifying a target using a wireless sensor network are proposed. The new tracking algorithm ensures good accuracy, low computational overhead on the sensor nodes, and efficient energy dissipation compared to the previous algorithms. Using the estimated target’s trajectory, the new classification algorithm determines the associated class of the detected target. In this thesis, targets are classified as animals, human beings, and automobiles. Simulation results show that although there is no need for any excessive processing on sensor nodes, the algorithm has an acceptable accuracy.
