Research On UHF RFID Positioning Technology Based On Signal Strength

With the development of the concepts of “Industry 4.0” and “Internet +”,the “Internet of Things(IOT)” industry has achieved rapid development and people put forward urgent requirements for indoor positioning technology.With its advantages of cost-effective,high precision,and high reliability,Radio Frequency Identification(RFID)has become one of the fundamental technologies of the IOT,and it is also the preferred indoor positioning technology.Indoor positioning has been a hot topic in RFID research,but most of research focus on calculating the specific position of object.However,in some situations,such as library management,logistic,it is of more useful to get their relative position rather than its specific position.Beside,the positioning accuracy of most positioning scheme can not reach millimeter level,which is not precise enough for relative positioning.A few positioning scheme can achieve high accuracy require dedicated hardware,which make more expensive to deploy in practice.In this thesis,the indoor positioning technology based on UHF RFID is taken as the research background,relative positioning for tagged objects in closed drawers is studied and the main contents are as follow:(1)Determine the relative order of tags inside the closed drawer along the X-axis and Yaxis.To fulfill the localization,an RFID reader is placed at different positions before a drawer and a RSS sequence can be created for each tag.Through analysis,some features of RSS sequence can reflect their relative positions.The X-axis coordinate of each tag is approximated via RSS peak,then use the K-means method to coarsely classify tags into columns and conduct a fine-grained column classification by comparing the similarities of their fluctuation patterns,finally,the order of tags along the X-dimension is determined.Subsequently,the peak values in RSS sequences are used to determine the Y-axis order of tags in the same column.Experiment results demonstrate that,compared with LANDMARC and STPP-RSS methods,the approach of relative positioning can achieve high ordering accuracy.(2)Based on the aforementioned research,a more practical method is proposed.For each tag inside the drawer,the reader obtained the RSS sequence by moving the reader antenna at a constant speed.Based on the matching coefficient,some tags' RSS features are represented by the corresponding quardratic curve,and the remaining tags are still represented by the RSS sequence.According to the peak value in the RSS sequence or quardratic curve,the tag order along the X-axis is obtained.Subsequently,for tags in the non-boundary column,the order along the Y-axis is determined by comparing the peak RSS value of each tag.For tags in the boundary column,the order of each tag along the Y-axis is determined by comparing the RSS values at the X-coordinate of the first tag's peak value.The experimental results demonstrate the approach can achieve high ordering accuracy.