Robust Precise Time Difference Estimation Based on Digital Zero-Crossing Detection Algorithm

Precise time difference (time interval) measurement has been widely applied in communication, detection, and metrology. Among the various methods of time difference measurement, the analog zero-crossing detection method and the full digital processing methods are commonly used in ultrasonic flow measurement application. Conventional full digital processing methods, which take advantage of total signals' waveforms, can have high accuracy in high signal to noise ratio (SNR) when employing a high enough sampling frequency and extremely high performance digital signal processors; but their accuracies are severely limited by the hardware costs and can be easily affected by the noise and the distortion of signals. As a viable method, the analog zero-crossing detection method can achieve high resolution, whose accuracy, however, is easily influenced by the detection threshold and baseline shifts of the received signals, as well as the noises. In this paper, the digital zero-crossing detection algorithm is proposed for ultrasonic flow measurement to achieve high-precision and high-robustness time difference estimation at relatively low hardware costs, i.e., a relatively low sampling rate (e.g., 10 Msamples/s) and a relatively low computational load. The digital zero-crossing detection algorithm is performed with the signal reconstruction and the least-square fitting technologies. The simulation and experiment results show that compared with conventional methods, the proposed method presents higher accuracy and higher robustness, especially in the case of low SNR.