基于FSM的物联网大数据清洗算法

在定长滑动窗口算法的基础上,应用有限状态机,按照标签的状态,对RFID标签数据进行清洗。算法解决了在清洗算法中重复窗口的问题,通过创建多阶哈希表存储标签读写记录,针对在海量标签数据中存在的多读、冗余读的现象,通过改进滑动窗口的数据清洗方法,解决其在处理效率方面的不足。对阈值等影响清洗算法准确率的相关因素进行仿真,结果表明基于FSM的海量RFID标签数据清洗算法可以有效清洗标签数据,且性能优于SMURF等算法。

Big Data Cleaning Algorithm for IoT based on FSM

Based on the fixed-length sliding window algorithm,a finite state machine is applied to cleaning the RFID tag data according to the state of the tag.This method solves the problem of duplicated windows in the cleaning algorithm,creates a multi-level hash table to store the read-and-write records of the tags.And aiming at the phenomenon of multiple reads and redundant reads in the massive tag data,and by improving the data cleaning method of the sliding window,its deficiency in processing efficiency is solved.The relevant factors that affect the accuracy of the cleaning algorithm,such as thresholds,are simulated.The simulation results indicate that the cleaning algorithm based on FSM for massive RFID tag data can effectively clean the tag data,and its performance is better than that of SMURF and other algorithms.