一种有效的多数据源RFID冗余数据清洗技术

摘要

Aim. The introduction of the full paper does two things: (1) it reviews Refs. 1 through 6 and then points out that popularizing RFID ( Radio Frequency Identification) technology to a large scale is difficult, owing to low quality of raw RFID readings, especially redundant readings; (2) it proposes what we believe to be a relatively more effective redundancy data cleaning technology. Sections 1 and 2 explain what we propose. Section 1 is entitled "abstraction of scenarios producing RFID redundant readings"; its core consists of: (1) Fig. 1 shows the abstraction of the scenarios producing RFID redundant readings; (2) it gives six definitions. Section 2 is entitled "algorithms for detecting and eliminating redundant data"; its core consists of: (1) Fig. 2 shows the schematic of merge -redundant cleaning technology; (2) Fig. 3 shows the schematic of cross -redundant cleaning technology, whose algorithm is based on reference tag strategy and RSSI (Received Signal Strength Indicator). The experimental results, presented in Figs. 4a) through 4f), and their analysis show preliminarily that our proposed technology can clean redundant RFID data effectively and, more importantly, will not destroy the integrity of RFID data.%相关文献指出低质量的RFID原始数据,特别是大量的冗余数据制约了RFID技术的大规模应用,因此需要一种更有效的冗余数据清洗技术.文章根据RFID数据以及典型应用场景的特点,提出了一种多数据源冗余数据清洗技术,其中包括融合冗余数据清洗方法和交叉冗余数据清洗方法.融合冗余数据清洗方法基于滑动窗口与集合论,交叉冗余数据清洗方法借鉴了参考标签思想并且结合信号强度特征.实验证明,提出的数据清洗技术不但可以有效地完成RFID冗余数据清洗,同时能够为清洗后的数据保留准确的数据来源信息.