ALOHA Improvement Algorithm for Dynamic Frame Time Slots with Deep Learning

Radio Frequency Identification (RFID) is a radio-based communication technology used to collect and identify data from tagged objects. In recent years, RFID technology has been widely applied in both industrial and daily life, driving a continuous increase in the demand for tag reading systems. In large-scale tag environments, RFID systems often encounter collisions due to simultaneous tag responses, which results in severe collisions within the same reading frame and decreases reading efficiency. The key to solving this issue lies in improving the speed and accuracy of the tag number estimation algorithm. This paper presents a novel tag number estimation algorithm that combines the Dynamic Framed Slotted ALOHA (DFSA) algorithm with Transformer and LSTM neural networks. Compared to traditional algorithms, this approach ensures more accurate tag number predictions, reduces the time consumption of the reading system, and enhances system throughput. Simulation results show that the proposed algorithm significantly improves the recognition accuracy of RFID systems while minimizing slot wastage.