An Accurate Anti-Collision Algorithm of Chipless RFID Tags Based on Deep Learning

Dynamic Frame Slot ALOHA algorithm is a commonly used algorithm to solve the tag collision problem in Radio Frequency Identification (RFID) systems. However, the existing ALOHA algorithm has a long consumption time for adjusting to the optimal frame length and results in too many idle time slots during the re-identification process. In this article, a modified algorithm for dynamic frame slot ALOHA based on deep learning is proposed. The system regards the number of tags in the reader's range at different times as a time series, and utilizes the superiority of the Long Short-Term Memory (LSTM) neural network in time series prediction to adjust the frame length. The simulation results show that the optimized DFSA algorithm based on LSTM can maintain a good throughput rate even when the number of tags is large, with less slot waste than the traditional DFSA algorithm.