| Summary: | 碩士 === 東海大學 === 工業工程與經營資訊學系 === 99 === In this research, we proposed a two-step localization method that contains the advantages of the virtual tags and two-step cluster method. The virtual tag could reduce a lot of decoration cost, it is flexible for dynamic environments. But there has a problem, if it couldn’t eliminate the unused tags in an effective way, then the accuracy of localization would be worse. Therefore, here combine the two-step cluster method replace the proximity map concept. Because there are highly associated with data and nearly object in the space environments. The attribute of signal pattern and position information using to replace the difference of signal in traditional. According to the high similarity in same group, and the high variability in different group. The first step could filter the highly similarity tags with the tracking tags using the signal pattern, then second step added the attribute of position, it eliminate the far area. Finally, this algorithm compared with other algorithm based on difference of signal .
|
|---|
