Abstract:The performance of navigation algorithms significantly determines the trajectory tracking accuracy of the guidance, navigation, and control (GNC) system of an autonomous underwater vehicle (AUV). In closed-loop operation, the interaction among path planning, control, and navigation plays a crucial role in the tracking accuracy of the overall GNC system. A Doppler velocity log (DVL) is often used for AUVs to measure velocity over the ground, positively affecting the closed-loop tracking error. However, a DVL may not be installed in miniaturized AUVs due to limited space and energy. In this paper, a navigation filter for an underactuated miniature AUV (nanoAUV) is considered that is mainly based on acoustic localization using a novel highly-miniaturized ultra-short baseline (USBL) system and a depth pressure sensor. The nanoAUV is being developed for subglacial lake exploration. We compare two unscented Kalman filters (UKF) with different prediction models - the classical strapdown inertial navigation systems (SINS) model and a hydrodynamic motion model (HMM). To enable a fair comparison, filter parameters are auto-tuned with Bayesian optimization (BO) for open and closed-loop performance, which is novel in AUV navigation. The results indicate that BO performs similarly to particle swarm optimization (PSO) regarding sample efficiency for the proposed problem. To quantify the GNC tracking performance, we use extensive Monte Carlo simulations. Results suggest that with BO-tuned navigation filter parameters, the median tracking error is reduced by up to 50% compared to default parametrization.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to optimize the nonlinear Kalman filter (such as the Unscented Kalman Filter, UKF) through automatic parameter - tuning techniques in the nano - Autonomous Underwater Vehicle (nanoAUV) to improve its navigation performance in open - loop and closed - loop operations, especially the trajectory - tracking accuracy when using the Ultra - Short Baseline Positioning System (USBL) and the depth - pressure sensor for navigation without a Doppler Velocity Log (DVL). Specifically, the paper focuses on: 1. **Navigation Performance Optimization**: Adjust the filter parameters through automatic parameter - tuning methods (such as Bayesian Optimization, BO) to improve the performance of the UKF based on the Inertial Navigation System (SINS) prediction model and the Hydrodynamic Motion Model (HMM) prediction model. This includes improving the accuracy of state estimation in the open - loop state and improving the accuracy of trajectory tracking in the closed - loop state. 2. **Filter Structure Comparison**: Compare the differences in navigation performance between the UKF of two different prediction models (SINS and HMM), and verify the performance of these filters under different conditions, such as different sensor noises, water flow speeds and USBL interruptions, through a large number of Monte Carlo simulations. 3. **Effectiveness of Automatic Parameter - Tuning**: Evaluate the performance of Bayesian Optimization (BO) and Particle Swarm Optimization (PSO) in terms of sample efficiency, and verify the feasibility and effectiveness of the automatic parameter - tuning method in practical applications. Through these studies, the paper aims to provide a better solution for the design of the navigation system of small AUVs, especially to improve the robustness and accuracy of the navigation system in sub - glacial lake exploration tasks in extreme environments.

Automated Tuning of Nonlinear Kalman Filters for Optimal Trajectory Tracking Performance of AUVs

Integrated Navigation for an Autonomous Underwater Vehicle Carrying Synthetic Aperture Sonar

AUV Docking Based on USBL Navigation and Vision Guidance

An Underactuated AUV Tracking Algorithm Based on Backstepping Adaptive Sliding Mode Control

An Integrated Auv Navigation Method Using Bayes Filtering

Particle Filter-Based AUV Integrated Navigation Methods

Unscented Kalman Filtering on Manifolds for AUV Navigation -- Experimental Results

Accurate two-step filtering for AUV navigation in large deep-sea environment

Joint Constrained Bayesian Optimization of Planning, Guidance, Control, and State Estimation of an Autonomous Underwater Vehicle

A Hybrid Adaptive Velocity Aided Navigation Filter with Application to INS/DVL Fusion

Seamless Underwater Navigation with Limited Doppler Velocity Log Measurements

System noise variance matrix adaptive Kalman filter method for AUV INS/DVL navigation system

Robust Model-Aided Inertial Localization for Autonomous Underwater Vehicles

An Integrated Navigation Algorithm for Underwater Vehicles Improved by a Variational Bayesian and Minimum Mixed Error Entropy Unscented Kalman Filter

Adaptive Step Size Learning with Applications to Velocity Aided Inertial Navigation System

The INS/GPS integrated navigation based on autonomous underwater vehicle using modified extended Kalman filter

Position USBL/DVL Sensor-based Navigation Filter in the presence of Unknown Ocean Currents

Tracking Control of Autonomous Underwater Vehicles with Acoustic Localization and Extended Kalman Filter

Improving Navigational Accuracy for AUVs Based on an Adaptive Modified Square-Root Cubature Kalman Filter

An Integrated Navigation Method Aided by Position Correction Model and Velocity Model for AUVs