Real-time tracking of moving targets using wireless sensor networks has been a challenging problem because of the high velocity of the targets and the limited resources of the sensors. CPA (closest point of approach) algorithms are appropriate for tracking fast-moving targets since the tracking error is roughly inversely proportional to the square root of the target velocity. However, this approach requires a specific node configuration with reference to the target trajectory which may not always be possible in randomly deployed sensor networks. Moreover, our mathematical analysis of the original CPA algorithm shows that it suffers from huge localization errors due to inaccuracies in sensor location and measured CPA times. To address these issues, we propose an enhanced CPA (ECPA) algorithm which requires only five sensors around the target to achieve the reliability and efficiency we want for computing the bearing of the target trajectory, the relative position between the sensors and the trajectory, and the velocity of the target. To validate the ECPA algorithm, we designed and implemented this algorithm over an actual data-centric acoustic sensor network as well as simulating it in an NS-2 simulator. The results of our field experiments and simulations show that we can achieve our goals of detecting the target and predicting its location, velocity and direction of travel with reasonable accuracy.
In this paper we propose a system that uses a sensor network to detect and respond to tsunamis. Sensor nodes sense underwater pressure data and send it to commander nodes where it is analyzed. Commander nodes use a general regression neural network (GRNN) to predict which barriers need to be fired in order to lessen the impact of the tsunami. We have implemented two versions of a GRNN to perform prediction and a genetic algorithm to optimize the parameters of the neural network. Finally, we analyze the performance differences for each version with respect to both accuracy and earliness of predictions.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.