Fig. 5: Experiment of frequency-varying motion signal recognition in continuous time.

a Schematic diagram of the recognition of continuous-time frequency-varying motion signals. A sequential voltage signal is segmented with a time interval of ∆t = 1, input into the network for prediction, and finally, the corresponding predicted variable curve for this time series is obtained. b Physical connection diagram of the experimental setup, including the data acquisition system (DHDAS), computer, power supply, flapping motor, Arduino microcontroller, and L298N drive module. c Schematic diagram of the experimental setup connection circuit. d Time-history curves of the flapping frequency with a sinusoidal variation in continuous time, showing the test values and predicted values (sliding step = 0.01 s, number of samples n = 20700). Source data are provided as a Source Data file. e Time-history curves of flapping frequency with a stepwise variation in continuous time, showing the test values and predicted values (sliding step = 0.01 s, number of samples n = 5941). Source data are provided as a Source Data file. f Comparison of test values and identified values of flapping frequency under the two variation modes in (d and e). Source data are provided as a Source Data file.