| In Wireless Sensor Networks, transmission and management of spatio-temporal data has been vitally important because of the limitations of the network. Regardless of the application, energy efficient schemes need to be proposed while adhering to requirements of the application. Toward this, event-based data management both utilizes system resources better and provides easier way to process the information in a meaningful way, since data volume is significantly reduced. Moreover, there is a growing need for many systems to address the efficient management of data streams and distributed systems. Solving the problem of spatio-temporal management of data on high velocity, resource-constraint medium can alleviate potential problems of any growing system.;In this dissertation, we first identify a set of predicates pertaining to meaningful events in the monitored phenomena and expand those predicates to capture co-occurrence of multiple events happening at the same time/place. In addition, we present algorithms and techniques to manage these predicates in the network in a distributed manner, rather than using centralized post-processing methods. Lastly, we propose a middleware system that brings users together with a WSN and an uncertain database. Although, our work spans events in environmental monitoring and object tracking in WSN, the proposed techniques can be applied to different applications in different domains. As proven by our experiments, our implementation schemes for predicate detection are very light-weight and stringent in terms of resource consumption. |
