RFID Data Denoising by Convolution and Frequency Domain Analysis
With the fast development of the RFID (Radio Frequency Identification) technology, more large scale systems are applied in different fields. However, the reliability and the large size of the RFID data are the major obstacles in the real applications. To guarantee the quality of the obtained data, we proposed a two-step algorithm which contains smoothing and cleansing. The smoothing step is used to increase the reliability of the data and the cleansing step is used to reduce the size of the data. In the smoothing step, Gaussian mask is used to smooth the raw data. In the cleansing step, the frequency domain analysis is introduced to cleanse the data. Furthermore, to make the algorithm suitable for online processing, the complexity of the algorithm is reduced by merging the two steps into one. The algorithm is applied on the simulated noisy RFID data which shows the accuracy of cleansing is high.
RFID noise convolution Fourier Transform
Xiang Gao
Department Computer Science and Engineering,Hong Kong University of Science and Technology,Hong Kong
国际会议
太原
英文
249-253
2011-02-26（万方平台首次上网日期，不代表论文的发表时间）