Mobile Intelligence

27.4 SIMULATIONS

In this section, the VGN approach is analyzed using simulations. The VGN approach provides an energy-efficient strategy to match patterns. Similarity among patterns produces shared areas that are utilized to conserve energy. The VGN approach considers these areas of redundance important to pattern matching, as they produce identical vote sets. Instead of using all the sensor nodes to match patterns, the VGN approach is capable of dynamically selecting a subset of sensor nodes based on their significance to create a team of sensor nodes capable of collaborating to matching event patterns.

In Section 27.3.4, the VGN approach performance under the worst case scenario has been studied. The communication requirement in the worst case scenario is approximate to scenarios where all sensory data are communicated and processed at the base station. The worst case scenario is limited to patterns that are of d(x, y) = 1. These patterns represent a very small portion of the total possible number of patterns, thus their effect is limited. Since the occurrence of the worst case scenario is rare, the VGN approach is always capable of eliminating redundant communication and conserving energy and attaining a performance better than the worst case or base station scenarios.

Nonetheless, the VGN performance is affected by the similarity among patterns. Reference patterns of various similarity result in various VGN performances. It is, therefore, necessary to simulate the VGN approach ...