Enhancement of Atomic Spin Gyroscopes Under Thermo-Mechanical Coupling: A Physics-Informed Measurement and Estimation Approach

In high-precision inertial measurement, the structural bias drift of atomic spin gyroscopes (ASGs) induced by residual thermo-mechanical coupling remains a critical limiting factor for long-term stability. This article presents a physics-informed adaptive measurement and compensation framework designed to suppress these deterministic and stochastic errors. Unlike pure data-driven models, the proposed method characterizes the physical coupling mechanisms to enhance the traceability of the estimation process. Two key metrological innovations are introduced: a statistical-physical dual-threshold gating mechanism that ensures causality in error detection to shield against nonphysical sensor noise, and an excitation-guided anisotropic injection strategy for decoupling channel-wise measurement uncertainties. This approach enables stable locking of quiescent parameters while maintaining rapid response to transient thermal disturbances. Experimental evaluation indicates that the proposed method reduces the bias instability (BI) by 8.66 dB and suppresses the rate random walk (RRW) coefficient by 32.41 dB. By providing a rigorous uncertainty analysis and decoupling recessive signal dynamics from environmental interferences, this work establishes a robust instrumentation paradigm for deploying quantum sensors in complex environments.