Decoupling compensation accuracy from sensor misalignment: A generalized black-box model for gravity gradiometers

Abstract

Moving-base gravity gradiometers are widely used in applications such as mineral exploration and passive navigation; however, their high sensitivity to platform motion presents significant challenges in compensating for motion-induced errors. Existing methods rely on precise alignment between motion sensors and the gradiometer’s coordinate system, making compensation accuracy heavily dependent on sensor installation precision. To address this issue, we propose a novel generalized black-box modeling method for motion error compensation in gravity gradiometers. By incorporating sensor misalignment directly into the model, our approach effectively decouples compensation accuracy from installation precision. Constructed based on the gradiometer’s dynamic characteristics and the excitation and input–output behavior of its sensitive components, the black-box model inherently avoids redundant operational terms, reduces multicollinearity, and lowers computational resource requirements. Validation through numerical simulations in a dynamic 0.1 g environment demonstrates that the output consistency between the black-box and analytical models reaches 10⁻¹⁰, achieving an compensation accuracy of 0.1 E. Furthermore, experimental results using a gravity gradiometer prototype confirm the model’s effectiveness in real-world conditions. The proposed method significantly reduces motion errors induced by 1 mg linear motion and 0.001 rad/s angular motion, lowering the noise level to 4 ng/$\sqrt{Hz}$, approaching the static inherent noise level of 1 ng/$\sqrt{Hz}$. These results verify the black-box model’s effectiveness and robustness in dynamic environments, highlighting its potential to enhance the practical performance of gravity gradiometers by mitigating motion-induced errors without stringent requirements on sensor installation precision.