A practical INS/GPS/DVL/PS integrated navigation algorithm and its application on Autonomous Underwater Vehicle. (January 2021)

• Record Type: Journal Article
• Authors: Mu, Xiaokai; He, Bo; Wu, Shuyi; Zhang, Xin; Song, Yan; Yan, Tianhong
• Is Part Of: Applied ocean research. Volume 106(2021)
• Publication Date: 2021-01
• Subjects: Autonomous Underwater Vehicle -- integrated navigation -- adaptive fault-tolerance filter -- Variational Bayesian; Ocean engineering -- Periodicals; 620.416205
• DOI: 10.1016/j.apor.2020.102441
• Languages: English
• ISSNs: 0141-1187

Abstract

Abstract: The integrated navigation system based on multi-sensor data fusion could effectively improve the navigation accuracy for Autonomous Underwater Vehicle (AUV). Various navigation equipment and sensors have different error characteristics. Current research does not take a specific processor based on their error characteristics, causing the optimal estimation is hard to realize in engineering applications. To address this issue, this study presents a robust and practical integrated navigation algorithm. When the AUV works on the surface where the Global Position System (GPS) could obtain the position, the navigation system employs an adaptive fault-tolerance filter to smooth the GPS trajectory, then the processed GPS information would be used to correct the Inertial Navigation System (INS). Otherwise, Variational Bayesian (VB) is adopted to estimate the measurement error covariance of the Doppler Velocity Log (DVL), which would be used for the INS/DVL integration system. Subsequently, the pressure sensor (PS) uses the conventional method to correct the height error of INS. The above information would be fused to obtain the position when the AUV operates underwater. The real experiment data of our independently developed Sailfish AUV is processed to evaluate the algorithm performance. Experimental results illustrate that the proposed algorithm could improve the navigation accuracy and the robustness of resisting unknown measurement uncertainty compared to the conventional method.