Given the lack of systematic research on bathymetric surveys with multi-beam sonar carried by autonomous underwater vehicles (AUVs) in unfamiliar waters, this paper proposes a method for multi-beam bathymetric surveys based on the constant-depth mode of AUVs, considering equipment safety, operational efficiency, and data quality. Firstly, basic principles for multi-beam bathymetric surveys under the constant-depth mode are proposed based on multi-beam operational standards and AUV constant-depth mode characteristics. Secondly, a vertical effective height model for the vehicle is established, providing vertical constraints and a basis for determining fixed depth in constant-depth missions. Subsequently, according to these basic principles and the vertical effective height model, the operational process for multi-beam bathymetric surveys in unfamiliar waters under the AUV constant-depth mode is outlined. Finally, we validate the proposed method through sea trials in the Xisha Sea of the South China Sea. The test results show that the method proposed in this paper not only ensures the vehicle safety operation and multi-beam data quality, but also improves the operation efficiency by about 68%, demonstrating the reliability of the proposed method and its significant engineering value and guidance implications.

Current methods of fusing side-scan sonar images fail to tackle the issues of shadow removal, preservation of information from adjacent strip images, and maintenance of image clarity and contrast. To address these deficiencies, a novel curvelet-transform-based approach that integrates the complementary attribute of details from side-scan sonar strip images is proposed. By capitalizing on the multiple scales and orientations of the curvelet transform and its intricate hierarchical nature, myriad fusion rules were applied at the corresponding frequency levels, enabling a more-tailored image fusion technique for side-scan sonar imagery. The experimental results validated the effectiveness of this method in preserving valuable information from side-scan sonar images, reducing the presence of shadows and ensuring both clarity and contrast in the fused images. By meeting the aforementioned challenges encountered in existing methodologies, this approach demonstrated great practical significance.

In response to the absence of standardized work practices, work safety measures, efficient work procedures, and suitable line planning methods for exploring seabed topography using autonomous underwater vehicles (AUVs) equipped with side-scan sonar systems, this paper proposes a design framework for side-scan sonar measurement technology based on AUVs. Detailed research has been conducted on line planning methods that differ significantly from those used in shipborne operations. The new line planning method was developed based on the performance characteristics of AUV equipment, the principles of motion modes, and shipborne side-scan sonar operations. The validity of the line planning method was verified through actual sea experiments using different types of AUVs in various terrains. The experimental results demonstrate that the new survey line planning method is distinct from existing AUV path planning methods and is highly applicable for conducting seabed topography exploration operations. Furthermore, the technical design scheme studied in this article is applicable in different sea areas and ensures the safety, operational efficiency, and data quality of AUVs. It holds significant importance for marine science engineering and practical applications. The article also features an analysis of successful data collected from AUVs equipped with side-scan sonar for conducting seabed topography exploration operations.

In light of the prevailing approach in which data from side-scan sonar (SSS) from Autonomous Underwater Vehicles (AUVs) are primarily processed and visualized post mission, failing to meet the requirements in terms of timeliness for on-the-fly image acquisition, this paper introduces a novel method for real-time processing and superior imaging of navigation strip data from SSS aboard AUVs. Initially, a comprehensive description of the real-time processing sequence is provided, encompassing the integration of multi-source navigation data using Kalman filtering, and high-pass filtering of attitude and heading data to exclude anomalies, as well as the use of bidirectional filtering techniques within and between pings, ensuring real-time quality control of raw data. In addition, this study adopts the semantic segmentation Unet network for automatic real-time tracking of seafloor lines, devises a real-time correction strategy for radial distortion based on historical echo data, and utilizes the alternating direction multiplier method for real-time noise reduction in strip images. With the combined application of these four pivotal techniques, we adeptly address the primary challenges in real-time navigation data processing. In conclusion, marine tests conducted in Bohai Bay substantiate the efficacy of the methodologies delineated in this research, offering a fresh paradigm for real-time processing and superior visualization of SSS navigation strip data on AUVs.

Aiming to address the problem that autonomous underwater vehicle (AUV) with multi-beam echo sounder system(MBES) for underwater topographic survey cannot autonomously obtain the sound velocity of the acoustic path to correct bathymetric data, this paper proposes a framework to correct sound velocity error(SVE) for AUV with MBES, which include sound velocity profile(SSP) extension and sound velocity error correction using genetic algorithm (GA) and support vector regression (SVR). The SSP is first calculated by extending the temperature and conductivity profiles to the seafloor based on the trend of temperature and conductivity with depth, which is obtained by using the least squares method to fit the data of conductivity temperature depth (CTD) sensor carried by AUV. Then the SSP is calculated according to the empirical sound speed formula. To eliminate the SVE in the extended SSP, the function of water depth, lateral distance and SVE was first derived. Then, the SVE of the main survey line was corrected, including calculating the discrepancy values between the matching point pairs of the main and auxiliary survey lines, and establishing a regression model about depth error using the SVR algorithm. Finally, the GA algorithm was used to optimize the hyperparameters of the SVR algorithm. The test results demonstrate that the proposed method effectively reduces the impact of SVE on multi-beam bathymetric data in both middle-deep and shallow water areas. The algorithm is not limited by terrain complexity, providing a solution for the lack of SSP in AUV autonomous topographic survey.

To address the challenges of low exploration efficiency, slow convergence speed, and poor decision interpretability in traditional Deep Q-Network (DQN) for maze navigation tasks, this paper proposes a maze navigation decision method based on Large Language Model(LLM) collaborative reinforcement learning. This method leverages LLM to perform dual evaluation of RL agent's single-step decisions(critical step quality) and global trajectories(path quality), and incorporates the evaluation results as priority weights into the experience replay mechanism to achieve priority learning of high-quality experiences. By constructing an LLM-RL collaborative decision framework, utilizing LLM for dual semantic evaluation of RL agent's single-step decisions and global trajectories, integrating evaluation results as priority weights into the experience replay mechanism, and designing a heuristic+LLM verification two-stage analysis mode to reduce computational overhead, priority learning of high-quality experiences is achieved. Through comparative experiments in a 6×6 grid maze environment, results show that compared to standard RL, Prioritized Experience Replay(PER), and Temporal Difference(TD) sampling methods, the proposed method improves navigation success rate by over 25% in the early training stage, reduces the number of training episodes required to reach 80% of final performance by nearly 40%, while maintaining the same optimal asymptotic performance and significantly enhancing decision interpretability. This validates the effectiveness of LLM-enhanced experience replay in improving the sample efficiency of DQN, and provides a practical reference for the fusion of large language models and value-based reinforcement learning.