Efficient RFID data cleaning model based on dynamic clusters of monitored objects

Because wireless radio frequency signal is adopted during the RFID (radio frequency identification) communication, many false negative and false positive readings may be produced leading to inaccurate event detection. In RFID-based applications, monitored objects often progress in the form of group. In this paper, by analyzing the group changes based on the defined association degree and dynamic clusters, a series of models and algorithms about association degree maintenance and data cleaning are proposed. Especially, by compressing the graph model, an association degree maintenance strategy based on splitting and recombining list model is discussed to improve the time and space performance. Simulated experiments have demonstrated the efficiency and accuracy of the proposed data cleaning model.