Research on RFID Localization Algorithm of Refrigerator Foods Based on Improved DenseNet Model

Long-term storage of food items often leads to expiration, as the user gradually forgets about their existence. Under this background, efficiently locating and managing food items in a complex environment is a bottleneck that must be solved for future refrigerators to substantially improve life quality for their users. This letter proposes an RFID fingerprint localization algorithm based on an improved DenseNet model combined with the Attention Mechanism to solve this problem. Compared with traditional CNNs, the improved model is 48.19% smaller in terms of the number of parameters and has better accuracy.