期刊名称:International Journal of Software Engineering and Its Applications
印刷版ISSN:1738-9984
出版年度:2015
卷号:9
期号:1
页码:27-34
DOI:10.14257/ijseia.2015.9.1.03
出版社:SERSC
摘要:A method of distinguishing several events included in sensor data was proposed, when various events were sensed simultaneously to the data several sensors sensed and reported and individual event data were mixed in an environment in which a data stream was sequentially obtained. This study had each sensor distinguish each event through a fast analysis of the data sensed by each sensor in this condition. For this purpose, clustering was made with sensor data, and first, internal variance of event cluster classified by each sensor was calculated, and how this changed with the passage of time was checked. Next, the cluster of each sensor sensing the same event was compared. Through this process, the sensed event could be more clearly distinguished. This study suggested the measure to divide a small quantity of sensors when it senses the several events. This measure can be used as application for the small mobile vehicle or robot to sense the peripheral situation.
关键词:Weighting; Clustering; Time slot; Context inference