Bio-inspired Enhanced Underwater Dead Reckoning Solution via IMU/Polarization/DVL Integration

Dead reckoning (DR) is a fundamental navigation technique for unmanned underwater vehicles (UUVs), yet its accuracy is constrained by attitude estimation, particularly heading determination, in complex and unstructured environments. This paper proposes an enhanced underwater DR framework integrating the Inertial Measurement Unit (IMU), polarization, and Doppler Velocity Log (DVL). To mitigate attitude drift, the DVL-derived velocity is employed for attitude updates instead of conventional integration-based methods. In addition, polarization cues are leveraged for heading correction, forming the polarization-based DR (POL-based DR) system. Following that, we develop a multi-sensor fusion strategy incorporating POL-based DR, IMU, and DVL, significantly improving both heading and positioning accuracy. Simulation and ocean experiments are conducted to validate the effectiveness of the proposed approach.