A knn-based RFID tag classification method and device

Abstract

The invention discloses a kNN-based RFID tag classifying method and device and belongs to the technical field of Internet of things radio frequency identification. The method comprises obtaining the RFID tag data of a known classification result in a readable range from the interface of an RFID reader, and configuring each tag number as a group of training data; calculating a signal strength RSSIand a Doppler shift DOPPLER statistical feature and matching feature for each group of training data to form an input feature vector of kNN learning; and identifying and classifying all readable tagsin a reader range in a warehouse by using a feature vector composed of the statistical feature and the matching feature and in combination with a kNN machine learning method. The method and device effectively distinguish static tags and free tags in an actual application scenario and to-be-sorted tags moving on a conveyor belt at a constant speed.

Description

technical field [0001] The invention relates to a K-Nearest Neighbor (kNN, k-NearestNeighbor)-based RFID tag classification method and device, and belongs to the technical field of radio frequency identification of the Internet of Things. Background technique [0002] The Internet of Things radio frequency identification technology (RFID: Radio Frequency Identification) is widely used in airport luggage sorting, unmanned supermarkets, warehouse inventory, express logistics and other fields. The RFID system pastes the RFID chip containing the item information on the surface of the product, so that each tag within the readable range of the RFID reader can be read by the system, which is used for real-time monitoring of tag status, cargo inventory, and cargo delivery. When the tags are moving on the conveyor belt, it is often necessary to sort the tags to distinguish the matching between the tags read at the same time and the goods. However, the tags above the conveyor belt hav...

AI Technical Summary

Problems solved by technology

When the tags are moving on the conveyor belt, it is often necessary to sort the tags to distinguish the matching between the tags read at the same time and the goods. However, the tags above the conveyor belt have a high probability of reading tags outside the conveyor belt , which leads to wrong sorting of labels, affecting the matching of labels on the conveyor belt and goods, resulting in wrong luggage sorting or wrong delivery addresses of express logistics, etc. In order to avoid these errors, it is hoped that labels that are not on the conveyor belt will be excluded as off-site labels In addition to the sorting algorithm, it is necessary to classify the tags read by the RFID reader to distinguish whether they are on the conveyor belt; in addition, when the item is moved slowly or placed on the shelf, the angle and degree of attention that need to be assigned is Different, therefore, it is also necessary to classify the tags read by the RFID reader dynamically and statically to improve the positioning accuracy of static items, and at the same time increase the attention to slow-moving objects, such as monitoring whether the goods in the unmanned supermarket have been taken Whether it was placed on the wrong shelf, whether it was suspected of being stolen, etc.

Images