基于动态Q学习的防碰撞算法的研究  

Research on Anti-collision Algorithm Based on Dynamic Q Learning

作　　者：陈浩[1] 张海峰[1] 应屹航 

机构地区：[1]杭州电子科技大学电子信息学院,浙江杭州310018 [2]杭州国芯科技股份有限公司,浙江杭州310012

出　　处：《杭州电子科技大学学报（自然科学版）》2016年第1期27-31,共5页Journal of Hangzhou Dianzi University：Natural Sciences

摘　　要：在射频识别系统中防碰撞是很关键的技术之一,针对EPC-C1G2协议的防碰撞算法中Q值调节不够灵敏的不足,提出了一种双参数且参数动态变化调节Q值的改进算法(EDQ算法).在EDQ算法中,将单一参数C分为双参数Cs和Cc,分别用来调节空闲和碰撞时隙的Q值.通过大量的实验分析,提出了在不同Q值下双参数Cs和Cc的最优值.仿真实验表明,算法能够减少时隙的浪费,从而增大系统吞吐率,减少标签的识别时间.Anti-collision is the key technology in the radio frequency identification system, as the insufficient of Q adjustment is not sensitive enough in anti-collision algorithm for EPC-C1G2 protocol, an improved algorithm（EDQ algorithm） is proposed for adjusting the Q value of the dual-parameter and parameter can be dynamic changing. In the EDQ algorithm, the single parameter C is divided into two parameters C, and Co, which are used to adjust the Q value of the idle and collision slots. Through the analysis of a lot of experiments, the optimal values of the dual-parameter Cc and C, are presented under different Q value. The simulation experiments show that the algorithm can reduce the waste of time slots, thus increasing the system throughput rate and reducing the time of label identification.

关 键 词：射频识别 防碰撞算法 Q值 系统吞吐率 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]