RFID系统欠定盲分离的汉明重分组防碰撞算法

Under-determined Blind Source Separation Anti-collision Algorithm for RFID Based on Hamming Weight Grouping

针对欠定盲源分离RFID(Radio Frequency IDentification)系统在标签数大量增加时产生的系统性能急剧下降等问题,提出了一种基于约束性非负矩阵分解(NMF,Non-negative Matrix Factorization)与汉明重分组技术相结合的多标签防碰撞算法。通过计算阅读器识别范围内标签的ID前M位的汉明重量的不同对标签进行分组,RFID系统的阅读器并按照组号的顺序依次识别。实验仿真表明,在天线数相同的情况下,改进算法与跳跃式动态树算法相比,当天线数为8~18时,标签识别率提高了42.8%~169.5%,标签识别速度提高了40.3%~62.9%,在多标签识别领域具有很好的实用性。

The results of the algorithm ofunder-determined blind source separation RFID（Radio Frequency IDentification） system become worse or even lead to the degradation of overall system performancewith the large increase in the number of tags. Anovel parallelizable identification anti-collision algorithm based on constrained non negative matrix factorization（NMF） and Hamming weight grouping technology was proposed. The tag groupings were made by calculating the Hamming weight of the tag＇s front M bits, which were within the reader identification range. Each group was identified by the RFID reader according to the grouping sequence. Simulation results show that when the antenna number is 8-18, compared with the Jumping and Dynamic Searching algorithmunder the same antenna number, the improved algorithm tag identification rate increased by 42.8%-169.5%, the tag identification speed increased 40.3%-62.9%. It has very good practicability in the field of multi tag identification.