RFID data cleaning method based on lightweight event detection

Abstract

The invention discloses an RFID data cleaning method based on lightweight event detection. According to the method, the advancing path of an RFID tag in an application scene is defined as some lightweight events in a formalized mode, a rule base containing a normal read-write path, a repeated read-write path and a missed read tag path is constructed, and real-time data cleaning is carried out on amiddleware layer of an RFID application system. According to the method, the deployment implementation problem of the RFID application scene reader is comprehensively considered, excessive reader position setting does not need to be carried out from the hardware level, the data cleaning rule can be added, deleted and modified according to the application scene, and any event description languagecan be used for defining the rule. The data cleaning process is located in the middleware layer of the application, belongs to lightweight operation, does not bring too much influence to system performance, processing of original data streams better conforms to the context environment, and the data cleaning result is more accurate and reliable.

Description

technical field [0001] The invention belongs to the field of RFID data cleaning, in particular to a method for cleaning RFID data based on lightweight event detection. Background technique [0002] Object identification and tracking using radio frequency identification technology (RFID) is the most popular technical solution in IoT applications. Different application scenarios use different types of RFID tags and middleware for data transmission and data management. Due to the limitation of application scenarios, uncertain information such as repeated reading and writing, misreading and missing reading of RFID tags will be generated in the process of RFID deployment and implementation. The lack of timely and effective front-end RFID data cleaning leads to low efficiency of the background business system. One type of algorithm for data cleaning at the front end of RFID applications is to use a probability model to fill in missing RFID data in a certain time interval, but thi...

AI Technical Summary

Problems solved by technology

Due to the limitation of application scenarios, uncertain information such as repeated reading and writing, misreading and missing reading of RFID tags will be generated in the process of RFID deployment and implementation. The lack of timely and effective front-end RFID data cleaning leads to low efficiency of background business systems

One type of algorithm for data cleaning at the front end of RFID applications is to use a probability model to fill in missing RFID data in a certain time interval, but this method requires a large amount of calculation; the other type is to use the cross-coverage method of RFID readers for verification. The cost of application deployment is high, and repeated reading and cross reading also need to rely on background middleware, and the scalability is not good

The existing RFID data cleaning algorithm lacks effective context information, does not use tag movement path information for cross-validation, and is not easy to implement and deploy. It is necessary to develop an easy-to-deploy and low-cost RFID data cleaning algorithm in the front-end of RFID applications

Images