The Fault Detection and Virtual Reconstruction Method for Underwater Navigation System

The underwater navigation system is prone to failure under the interference of complex environment. In this article, a fault detection and virtual reconstruction method is proposed for strap-down inertial navigation system (SINS)/ultrashort baseline (USBL) navigation system, which is based on principal component analysis (PCA) and extreme learning machine (ELM). First, according to the problem of USBL information failure, the PCA model is established. Meanwhile, the online fault identification method is proposed based on the PCA model. Second, in order to maintain navigation performance during the fault, the ELM method is introduced to construct virtual USBL information. Meanwhile, an improved ELM algorithm is proposed. Finally, the simulation and river test are designed. Compared with PCA-SINS and PCA-ELM, the proposed PCA-IELM method improves the horizontal position error (HPE) by 99.77% and 84.95%, respectively. Therefore, the PCA-IELM method proposed in this article has a good performance in the SINS/USBL navigation system.