| Summary: | This paper studies a novel anti-collision algorithm that is proposed in view of the problem, i.e., label reading collision in radio-frequency identification for Internet of Things. Based on the theoretical foundation of label grouping, the algorithm introduces the Mahalanobis distance and density function to traditional fuzzy C-means clustering grouping algorithm by using EPC code and effectively solves the problem of isolated points of clustering and the optimization problem of initial clustering center. Then, the algorithm realizes effective grouping of labels and distributing identification serial numbers to labels upon the distance from interior labels to the center of clustering. Meanwhile, the efficiency of algorithm can be improved through dynamically setting the frame slot time of readers upon the grouping condition to prevent collision. This paper analyzes the throughput rate theoretically in detail. The simulation results of throughput capacity, throughout rate, and slot efficiency of the algorithm manifest that the algorithm is superior to the most commonly used dynamic binary-tree algorithm and current dynamic ALOHA algorithm in performance.
|
|---|
