Sensor networks have become attractive in research and applications. Sensor networks provide unprecedented opportunities for monitoring and controlling homes, cities and the environment. In addition, these networks have potential applications in military, medical science, environment, etc. One of the most important applications of these networks is in target tracking. A target moves in an arbitrary but continuous path. Target's location is sampled at specific times. Target's path is determined based on these samples. Many algorithms are suggested for target tracking in sensor networks. Since, limited energy resources are the most important problems in wireless sensor networks, the goal of different algorithms is increasing accuracy and decreasing energy consumption. In many of these algorithms, data transmission is performed through a leader. This leader is a node which is chosen based on a certain policy. At a given time only one leader is present in the network, hence, the transmission of same data for one target is avoided. Therefore, energy consumption is reduced. In the existing algorithms, in some cases, the nearest sensor to target's path is chosen as the leader. Nonetheless, it is possible that the leader doesn't have enough energy and handoff should be done. Increasing the number of communications leads to increasing the energy consumption in the network. On the other hand, it is possible that no appropriate node is found. Therefore, the network misses the target's path and tracking accuracy is reduced. In this thesis, we have tried to provide accurate target tracking and prevent waste of network resources. For reducing energy consumption, minimum numbers of nodes are activated and communicate with each other. At a given time, only one sensor is active as the leader. A two phase technique is used to determine the leader. The initial phase establishes the leader when a target first arrives in the network. The second phase, is employed when one leader is present in the network. The leader estimates target's path and predicts it in the next time before the target is going out of the leader's sensing range. Then, a sensor which can sense the target in the next period time and have the most energy level is chosen as the next leader. We call this procedure, forced handoff. Forced handoff is continued until the target leaves the ... Key Words Wireless sensor networks, Target tracking, Forced handoff, Location prediction, Energy
