Echo Frequency Estimation Technology for Passive Surface Acoustic Wave Resonant Sensors Based on a Genetic Algorithm

Abstract

Passive wireless surface acoustic wave (SAW) resonant sensors are widely used in measuring pressure, temperature, and torque, typically detecting sensing parameters by measuring the echo signal frequency of SAW resonators. Therefore, the accuracy of echo signal frequency estimation directly affects the performance index of the sensor. Due to the exponential attenuation trend of the echo signal, the duration is generally approximately 10 μs, with conventional frequency domain analysis methods limited by the sampling frequency and data points. Thus, the resolution of frequency estimation is limited. Here, signal time-domain fitting combined with a genetic algorithm is used to estimate SAW echo signal frequency. To address the problem of slow estimation speed and poor timeliness caused by a conventional genetic algorithm, which needs to simultaneously estimate multiple parameters, such as signal amplitude, phase, frequency, and envelope, the Hilbert transform is proposed to remove the signal envelope and estimate its amplitude, and the fast Fourier transform subsection method is used to analyze the initial phase of the signal. The genetic algorithm is thereby optimized to realize the frequency estimation of SAW echo signals under a single parameter. The developed digital signal processing frequency detection system was monitored in real time to estimate the frequency of an SAW echo signal lasting 10 μs and found to have only 100 sampling points. The proposed method has a frequency estimation error within 3 kHz and a frequency estimation time of less than 1 s, which is eight times faster than the conventional genetic algorithm.