Mobile Intelligence

27.6 CONCLUSION AND FUTURE WORK

In this chapter, an energy-efficient and distributed template-matching approach to detect large-scale events using sensor nodes has been proposed. Instead of performing template-matching at the base station, the VGN approach matches patterns within the network. The proposed approach uses the information processing-based and sleep mode-based techniques to exploit in-network processing, reduce the amount of communication, and prolong the durability of the network. The event-detection problem is converted into template-matching. Patterns depicting interesting phenomena of the environment are used to detect and classify events of interest. The patterns are stored in a distributed manner within the network by partitioning these patterns into their fundamental pattern elements (pairs). The approach dynamically manages the sensor network's collaboration by concentrating the pattern-matching process within sensor nodes of higher significance. The node collaboration results in converting interesting local sensory information into meaningful global event-recognition consensus. Incorporating a sleep mode strategy facilitates on-demand processing and thus leads to better throughput utilization and prolongs the lifetime of the sensor network.

The simulation results have shown the capability of the VGN approach to match patterns within the network. We have proposed novel metrics to evaluate the performance of pattern matching for sensor networks. The metrics evaluate ...