Enhanced Multi-AUV Localization Network: Achieving Alternative Estimation of Unreliable Sensor Inputs With Delay-Biased Observation Assistance

In case that the sensor inputs become unreliable due to compass or Doppler Velocity Log (DVL) failures, the posterior state estimation will rely more heavily on external observation information. However, the complex underwater environment introduces not only the instability of internal navigation sensors but also uncertainties and delays in acquiring external information via acoustic communication. Most existing unknown input estimation methods are derived and designed under the assumption that the observations at the current time step are always reliable. This ideal assumption is often violated in realworld underwater cooperative localization (CL) networks. In this study, we propose a method to address a long-duration sensor fault and enhance the stability and reliability of the systems output for unknown inputs and position estimates, assisted by external observations with delay bias. The performance of the proposed algorithm is tested through lake trials involving various failure modes, including isolated and combined DVL and compass faults, reflecting realistic challenges faced by low-cost autonomous underwater vehicles (AUVs). The results fully validate its advantages, demonstrating improved localization accuracy without a significant increase in computational complexity at the order-of-magnitude level.