Modeling and Compensation of Dynamic Non-Orthogonality Error for Polarization-Based Integrated Navigation Systems

Polarization-based integrated navigation system (PINS), which integrates polarization sensor (PS) and inertial navigation system (INS), provides an effective solution for attitude and heading information acquisition. However, in practical atmospheric environments, the Rayleigh scattering law can be affected by temporal factors, resulting in a time-varying non-orthogonality error (NOE) between the sun vector and e-vector, thus degrading the performance of PINS. To address the above challenge, a PINS fusion scheme is proposed under the composite disturbance filtering (CDF) framework, which consists of modeling and compensation methods for the NOE. The dynamic characteristics of NOE are analyzed with respect to the temporal variation of the skylight polarization pattern, and a multi-harmonic model of the NOE is thereby established. On this basis, a PS/INS fusion model considering both NOE and inertial measurement unit (IMU) biases is proposed, where a disturbance observer (DO) is designed to compensate for the NOE, and a Kalman filter (KF) is employed to account for the gyroscope biases. The proposed method provides a refined treatment of the multi-source heterogeneous uncertainties in the PINS, improving the system navigation performance, especially under long-term navigation tasks. Finally, the effectiveness of the proposed scheme is validated via both numerical simulations and experimental tests. The experimental results indicate that the PINS with the proposed modeling and fusion scheme is capable of maintaining a heading accuracy of within 0.1° in clear-sky conditions throughout 10 more hours of operation.