RFID Antenna Paramter Optimization Method Based on Improved GA-BP Algorithm

In order to improve the prediction accuracy of the algorithm for antenna parameters, an improved genetic algorithm-back propagation (GA-BP) algorithm based on Adagrad optimizer is proposed. By introducing the Adagrad optimizer and threshold strategy in the iterative process, the position information of the optimal individual in the degenerated population is redirected, solving the problem of insufficient local optimization ability of the GA-BP algorithm, significantly reducing the error loss and accelerating the convergence speed. Based on this method, the printing quality and electromagnetic parameters of the radio frequency identification (RFID) tag antenna are modeled and analyzed. The results show that the improved GA-BP algorithm can search for extremes steadily while avoiding the local extremum trap, and outperforms the traditional BP algorithm and GA-BP algorithm in terms of error and con-vergence efficiency, and can achieve higher prediction accuracy for the optimal control of RFID tag antenna print quality and S11 char-acteristic curve prediction. Compared with the BP and GA-BP algorithms, the improved GA-BP algorithm can reduce the average ab-solute error in the application of printing quality of RFID tag antennas by 91. 92% and 85. 64%, respectively. The reductions are 13. 77% and 13. 19%, respectively, when the improved GA-BP algorithm is applied to electromagnetic parameter prediction.