Collision-Aware Churn Estimation in Large-Scale Dynamic RFID Systems.
Qingjun Xiao,Bin Xiao,Shigang Chen,Jiming Chen
DOI: https://doi.org/10.1109/TNET.2016.2586308
2017-01-01
IEEE/ACM Transactions on Networking
Abstract:RFID technology has been widely adopted for real-world applications, such as warehouse management, logistic control, and object tracking. This paper focuses on a new angle of applying RFID technology—monitoring the temporal change of a tag set in a certain region, which is called churn estimation. This problem is to provide quick estimations on the number of new tags that have entered a monitored region, and the number of pre-existing tags that have departed from the region, within a predefined time interval. The traditional cardinality estimator for a single tag set cannot be applied here, and the conventional tag identification protocol that collects all tag IDs takes too much time, especially when the churn estimation needs to perform frequently to support real-time monitoring. This paper will take a new solution path, in which a reader periodically scans the tag set in a region to collect their compressed aggregate information in the form of empty/singleton/collision time slots. This protocol can reduce the time cost of attaining pre-set accuracy by at least 35%, when comparing with a previous work that uses only the information of idle/busy slots. Such a dramatic improvement is due to our awareness of collision slot state and the full utilization of slot state changes. Our proposed churn estimator, as shown by the extensive analysis and simulation studies, can be configured to meet any pre-set accuracy requirement with a statistical error bound that can be made arbitrarily small.