LAUSR

The localization of autonomous underwater vehicles (AUVs) in anchor-free environments has always been a difficult problem due to the lack of global positioning systems and absolute references. In general, AUVs localize themselves by dead reckoning (DR), whereas the localization error grows without bound. To alleviate the growth of the localization errors, we propose intermittent belief propagation based dead reckoning (IBPDR) as a cooperative localization (CL) framework. In IBPDR, AUVs use DR to localize themselves and periodically correct DR's deviation with CL methods. The intermittent feature of IBPDR reduces communication costs among AUVs by decreasing the frequency of CL. In the IBPDR framework, we design a particle-based underwater-adaptive belief propagation (UABP) algorithm for CL. The UABP algorithm is naturally distributed and viable in nonlinear and non-Gaussian situations. Thus, it is suitable for CL issues. Furthermore, the UABP algorithm is robust to the accumulated inertial measurement errors and reduces communication costs among AUVs. Moreover, we propose a particle-based current-aided filter to further improve the localization accuracy by comparing AUVs’ ambient current observations with the available current maps. Simulation results validate the proposed algorithms by comparisons with alternative approaches in localization accuracy, communication costs, and robustness to abnormal cases, such as packet loss, ranging bias, and outliers.