Research On Optimization Methods Of Parameter Design For PCB RFID Tag Antenna

With the development of RFID technology research and application,the demand for various types of RFID tag antennas for tracking and identifying targets has increased sharply,and the design tasks of tag antennas are arduous.Aiming at the problems of long design time and unclear parameter optimization direction caused by relying on full-wave electromagnetic simulation to calculate antenna parameters and manual experience to reset antenna structure length in the parameter design of traditional PCB RFID tag antenna,the design process of low return loss,wideband and miniaturization of PCB RFID tag antenna with folded dipole structure is taken as the research object,an optimization methods of parameter design for PCB RFID tag antenna is presented.The parameter design method of traditional PCB RFID tag antenna is improved from the two aspects of parameter prediction and parameter optimization,mainly including:1.A parameter prediction method for a simple folded dipole antenna is proposed.Firstly,the impedance transformation hypothesis based on antenna size and the impedance prediction model hypothesis based on linear relationship were proposed respectively,then the hypothesis was improved and verified by the simulation data,finally,a complete method for parameter prediction of folded dipole antenna is developed.The experimental results show that Compared with the traditional full-wave electromagnetic simulation,the method dramatically shortens the design time of the antenna without the loss of accuracy on various fold and frequency.2.A parameter prediction method based on Bi LSTM neural network is proposed for more complex antenna structures.Firstly,according to the current design trend the structure of the bent dipole antenna was optimized and a new form of PCB RFID tag antenna was formed.Then,the sequence of antenna structure was analyzed,and a surrogate model based on Bi LSTM neural network model structure was established for antenna parameter prediction.Finally,the simulation data was collected to train the surrogate model and a comparative experiment was set to verify its performance.The experimental results show that the Bi LSTM surrogate model is obviously better than the BP surrogate model and RBF surrogate model in the accuracy of parameter prediction of PCB RFID tag antenna,and the time consumption is significantly lower than the HFSS simulation.3.An improved NSGA-II algorithm for parameter optimization of PCB RFID tag antenna is proposed.Normal distribution crossover operator was introduced into the traditional NSGA-II algorithm,and constraints were added to meet the size requirements of tag antenna parameter design.Meanwhile,the fitness function was adjusted to reduce the influence of Bi LSTM surrogate model error on the optimization process.Finally,the update ability of the crossover operator and the optimization ability of the improved NSGA-II algorithm in the parameter design of RFID tag antenna parameters were verified respectively.The results show that the improved NSGA-II algorithm has a larger target space and Pareto front is more uniform and the optimization effect is better.Based on the above improvement,for the purpose of optimizing the parameter design process of traditional PCB RFID tag antenna,the antenna parameters are given by parameter prediction instead of full-wave electromagnetic simulation,and Instead of manually resetting the antenna structure length,the improved NSGA-II algorithm is used to iterate the Pareto optimal set.Finally,a design example was given.The optimized parameter design method of PCB RFID tag antenna is used to complete the design of PCB RFID tag antenna with low return loss,wideband and miniaturization.