Abstract: The efficiency of stochastic ALOHA to solve tags collision is very low, and deterministic tree searching algorithm has the limitation that the number of the tags in the area doesn’t change. This algorithm gets over this disadvantage, According to the reader’s identifying result each time, the tags can modify their responsive probability in the incremental or degressive mode. Finally, its identifying efficiency can steadily attain 0.322 in the dynamic condition and a lot of tags simultaneously appearing. This article describes it by Markov chain. This article also analyses what value should be in the case that the identifying efficiency of this algorithm initially reaches the best and how much the changing range of the increase or decrease should be in the dynamic situation to make this algorithm more appropriate and efficient.

Key words: Exponential index, RFID, Anti-collision algorithm, Markov chain

摘要： 解决RFID多标签冲突的随机ALOHA方法效率较低，确定性树型方法要求区域内标签数量不变。本算法克服了这些局限，根据阅读器每次识别的结果，标签以递增或递减方式修改其应答概率。最终，该算法识别效率在动态以及标签数量庞大的情况下也可以稳定地达到0.322。本文用马尔可夫链理论对该算法模型进行了刻划。重点针对标签以线性方式进入时，在识别效率能初步达到最优的情况下，标签可以取得的极小状态级别数k以及标签应答概率动态变化时，变化的幅度如何才能更加合理进行了分析。

关键词: 指数索引, RFID, 反碰撞算法, 马尔可夫链