Article ID: | iaor20133143 |
Volume: | 55 |
Issue: | 1 |
Start Page Number: | 256 |
End Page Number: | 269 |
Publication Date: | Apr 2013 |
Journal: | Decision Support Systems |
Authors: | Jin Beihong, Zhuo Wei, Hu Jiafeng, Chen Haibiao, Yang Yuwei |
Keywords: | artificial intelligence: decision support, optimization |
Many applications in the Internet of Things (IoT) depend on the occurrences of events with temporal and spatial constraints to determine the further actions. A major challenge encountered is how to specify and detect the spatio‐temporal events. The paper adopts Pub/Sub middleware to help IoT applications to capture spatio‐temporal events. Specifically, the paper presents a composite subscription language CPSL and builds the corresponding Pub/Sub middleware Grus. The subscriptions in CPSL can specify diverse temporal, spatial and logical relationships of events, in particular, can describe the moving events related to mobile objects, and Grus is responsible for detecting whether events are matched with subscriptions in a distributed way. In addition, Grus provides the optimization strategies for subscriptions involving unary spatial operators. The paper also evaluates Grus's matching performance and costs through simulation experiments. The experimental results show that Grus can achieve satisfying performance and acceptable overheads, and the optimization strategies can efficiently speed up the detection of spatial events.