A Novel Information Fusion Method of RFID Strain Sensor Based on Microstrip Notch Circuit

According to the current development situation, the application of chipless radio frequency identification (RFID) sensors in structural health monitoring has the disadvantages of difficult detection, poor effect, and single design function, limiting the further development of this technology. Therefore, a new type of RFID strain sensor that realizes the compact structure and separated function of miniaturized chipless RFID coded tags is proposed. The integrated disk monopole antennas make wireless measurement possible. Through the single-parameter strain simulation analysis, six characteristic parameters with good linearity are determined. The time sequences data enhancement algorithm and the background noise data enhancement algorithm are used to expand the training dataset. Then, the BP neural network is utilized for data fusion, and the training error finally converges to 0.0005. A comparison experiment between wired and wireless is designed, and the wireless experiment is optimized through wired experimentation. The results of wireless measurement experiments show that, combined with multiparameter information fusion technology, the average error between the proposed sensor and the actual strain is 6.04%, and the minimum error is 0.25%. The multiparameter fusion monitoring method of the strain sensor corrects the error of single-parameter measurement, and its accuracy and robustness are improved.