New smoothing scheme combining RTSS and FBC for PPP/INS integrated system in ocean environments

Abstract For high-precision positioning and attitude determination of unmanned systems in marine environments such as unmanned surface vessels (USV) and unmanned underwater vehicles (UUV), precise point positioning (PPP) and inertial navigation system (INS) integrated systems are usually recommended. The post-mission smoothing algorithms can fully utilize observations to provide optimal estimates and address the limitations inherent in Kalman filtering (KF) in satellite signal degraded environments. We propose a new post-processing smoothing scheme combining Rauch-Tung-Striebel smoother (RTSS) and Forward-Backward combination (FBC). The performance of the traditional RTSS, FBC, and the newly proposed FBC-RTSS smoothing schemes are compared and analyzed through three experiments in different ocean scenarios, namely the USV (optimal satellite visibility condition), UUV Ⅰ (good ocean condition), and UUV Ⅱ (rough ocean condition) tests. The results show that the RTSS, FBC, and FBC-RTSS post-processing smoothing schemes exhibit enhanced positioning and attitude determination accuracy compared to the forward Kalman filtering (FKF) real-time scheme. The newly proposed FBC-RTSS smoothing scheme exhibits superior performance compared to both the RTSS and FBC smoothing schemes. For the USV and UUV I tests under fine satellite observation conditions, the post-processing smoothing schemes have limited improvement in positioning accuracy but have a significant improvement in attitude accuracy. However, for the UUV II test under conditions of frequent satellite signal interruptions, the post-processing smoothing schemes significantly elevate both the positioning and attitude accuracy. The RTSS, FBC, and FBC-RTSS smoothing schemes have improved the 3D positioning accuracy by 46.5%, 31.3%, and 63.2%, and heading accuracy by 52.4%, 51.8%, and 55.5%, respectively, compared to FKF scheme for the UUV II test.