A non-rigid terrain transformation particle filter suitable for terrain-aided navigation without sound velocity profile

The deepening development of marine resources has led to an increasing demand for Autonomous Underwater Vehicles (AUV), with higher performance requirements. Among these requirements, high -precision navigation serves as a crucial foundation. Therefore, terrain -aided navigation (TAN) systems are commonly used due to their low measurement difficulty, high precision, and strong autonomy. Particle filtering -based TAN methods have gained widespread attention because they can adapt to the nonlinear and non -Gaussian characteristics of terrain information. However, existing methods rely on external environmental information support, particularly sound velocity profile (SVP), to provide accurate measurement for navigation. During filtering, only random errors in measured terrain are considered, which limits the autonomy of TAN. To address this issue, we proposes a particle filter -based TAN method that incorporates a non -rigid terrain transformation model (called NR-PF); this model utilizes distance attenuation coefficient and incident angle transformation coefficient that have nonlinear relationships with beam incidence angles to approximate complex sound ray along curves into straight lines. This method can correct non -rigid terrain transformation errors caused by the absence of SVP, and eliminate its impact on TAN. Finally, we conducted simulation experiments to quantitatively and qualitatively analyze the effectiveness of the algorithm.