Research On The Grouping Identifying Repeatedly RFID Anti-Collision Algorithm

There are two issues when Framed Slotted ALOHA algorithm is used to readlarge-scale tags. First, the error caused by the tag estimation algorithm becomes bigger asthe number of tags increases. Secondly, for the frame size is limited, when the tag is denseand each tag randomly selects one slot, it will lead to “Tag Starvation” issue, then theduration of the identifying will last long. Many scholars proposed grouping-basedalgorithms. The algorithms divide all tags into many small tag sets and then identify eachset one by one until all tags are identified. However, these algorithms still use thetraditional tag estimation strategy which may lead to estimation error. More over, when thenumber of tags changes dynamically, some new or old tags may not be read efficiently.In order to solve the issues of identifying of large-scale and dynamically-changeRFID tags, we propose a grouping-identifying-repeatedly RFID anti-collision algorithm.In this algorithm, we use a divide-and-conquer based tag estimation strategy and agrouping-repeatedly strategy. The former strategy estimates the number of tags of eachsubset instead of the whole tag group, which improves the accuracy of tag estimationlargely. The latter strategy divides all unidentified tags again and again during theidentifying process until all tags are identified, which is able to detect the dynamicchanges of unidentified tags efficiently, avoiding the issues of low efficiency andleakage.Five RFID anti-collision algorithms, including Dynamic Framed Slotted ALOHAalgorithm (DFSA), Adaptive Framed Slotted ALOHA algorithm (AFSA), EnhancedDynamic Framed Slotted ALOHA algorithm (EDFSA), Grouping-based Dynamic FramedSlotted ALOHA algorithm (GB-DFSA) and our proposed Grouping-Identifying-Repeatedly algorithm (GIR), are simulated in this thesis. The simulation traces theidentifying process dynamically and provides visualization for the identifying efficiencyanalyses.For the purpose of finding the efficiency differences between AFSA algorithm,EDFSA algorithm, GB-DFSA algorithm and our proposed GIR algorithm, we compare the identifying efficiency of our proposed algorithm when the number of tags and frame sizechange respectively with the other three algorithms. All results show that in large-scaleand dynamically-change RFID tag identifying environment, our proposed algorithmoutperforms the other three algorithms in system efficiency and stability largely.