Micro-thrust measurement technology plays a pivotal role in the detection of space gravitational waves. However, ultra-low frequency stochastic drift poses a significant challenge to the precision of micro-thrust measurements. To mitigate the low-frequency drift during the process of micro-thrust measurement, this paper introduces a multiple scale drifting compensation method for ultra-low frequency micro-thrust measurement based on autoencoder prediction. A short time micro-scale drift skeleton of the raw signal is extracted using a bidirectional Butterworth low-pass filtering approach. An optimized autoencoder model is employed to predict a long-time sequence of the low-frequency drift skeleton. Then, the predicted skeleton sequence is removed from the raw signal to obtain a micro-thrust measurement signal with drifting compensation. It is validated that the autoencoder prediction-based multiple scale compensation of low-frequency drift in micro-thrust measurement successfully reduced the noise level of the measured micro-thrust data to 0.30 μm/Hz^1/2 at a frequency of 8×10^-5 Hz, confirming its effectiveness in suppressing low-frequency noise.

Micro-thrust Measurement; ultra-low frequency drift compensation; Bidirectional Filtering; Autoencoder prediction