Abstract
Wireless sensor network (WSN) is a popular research topic in manifold applications including medical care, industrial control, entertainment, security, military defense, and other possible fields. Developers therefore ceaselessly brings forth different sensor platforms to deal with these applications. However, little eyesight is focused on the platform’s actual performance. Since few researches and methods nowadays are able to evaluate WSN platforms, users tend to roughly guess the overall performance through looking up the data sheets or speculations of the composed devices. Oversights within these processes might consume redundant cost or decrease the work efficiency. Besides methodologies for evaluation, lacking of WSN emulation environment and benchmarking suites are also critical issues for platform selection. In this thesis, we propose practical methodologies to quantitate the metrics that affect WSN system’s data quality: conversion accuracy and timing precision. The measured results show that these deviations are larger than the desk study value. Besides, we address that for a certain application, the geometric mean of corresponding metrics can be utilized as its overall fitness. Regarding the WSN system emulation, environmental signal reconstructions are fulfilled according to sampling theory and spline interpolation. These signals are digitized and stored afterward to support devices for data playback operation. Also, data processing, instrument control, and emulation graphical user interface are handled by National Instruments Lab- VIEW software.By these emulation and evaluation methods, we hope users can make efficient platform selections and testings. Moreover, pressing WSN developers to make more efforts in systematic considerations and design platforms more authentically. Consequently, making contributions to the progress and improvement of the overall wireless sensor network technology field.