水下无人系统学报

An Engineering Approach for Obstacle Avoidance Path Planning of Unmanned Vessel Based on Optimized Artificial Potential Field

ZHANG Yabo, WANG Hongrui, ZHANG Haiyan, XIAO Qiang, LIU Yuxin

, Available online , doi: 10.11993/j.issn.2096-3920.2025-0030

[Abstract](8) [FullText HTML] (5) [PDF 1259KB](2)

Abstract:
Aiming at the local path planning problem of obstacle avoidance for unmanned surface vessel, relying on the artificial potential field framework, a local path planning method for obstacle avoidance based on the dynamic construction of the water surface situation in longitude and latitude coordinates is proposed. Initially, the basic operations in the longitude and latitude coordinate system are sorted out and organized, and then the forms of the gravitational and repulsive force functions of the traditional potential function method are derived. The problems existing in the traditional potential function method and its improved methods, such as the difficulty in determining the virtual target point in the project and the inability to accurately predict the trajectory of the controlled object, are expounded. An improved potential function local path planning algorithm relying on the dynamic construction of the water surface situation is designed. Finally, the designed method is verified by simulation and sea trials. The simulation and test results show that the proposed engineering method of obstacle avoidance path planning can guide the unmanned surface vessel to complete the obstacle avoidance task, and has strong reliability and robustness.

Ship Radiated Noise Line Spectrum Enhancement Based on Adaptive Filtering-Deep Learning Fusion

CAI Tingting, YU Sunze, ZHAO Mei

, Available online , doi: 10.11993/j.issn.2096-3920.2025-0040

[Abstract](9) [FullText HTML] (5) [PDF 2026KB](2)

Abstract:
To address the challenge of identifying line spectrum features of ship radiated noise in complex underwater acoustic environments, a line spectrum enhancement model combining adaptive filtering and deep learning (DL) is proposed. This model integrates a double-layer collaborative adaptive filtering module with a deep learning module to form a multi-level feature enhancement framework. The adaptive filtering module combines frequency-domain adaptive filtering(FDAF) with the maximum correntropy criterion (MCC) to enhance the suppression of non-stationary noise while effectively reducing broadband background noise. The deep learning module employs a bidirectional long short-term memory (BiLSTM) to extract the local temporal dependencies of line spectra. It also incorporates an attention residual mechanism to focus on weak line spectra and utilizes a Transformer encoder to capture long-range correlations in the time-frequency domain. The model effectively combines the advantages of filtering and deep learning, both suppressing noise and enhancing the detection of weak spectral lines. Simulation results demonstrate that the proposed method outperforms single adaptive filtering or deep learning models in terms of overall line spectrum enhancement and weak line spectrum enhancement. The effectiveness of this method is further validated through lake test data.

Underwater High-speed Target short-time Target Motion Analysis Based on Pseudolinear Kalman filter

ZHAO Junhao, MA Hui, WANG Mingzhou, ZHANG Jun, CAO Hao

, Available online , doi: 10.11993/j.issn.2096-3920.2024-0171

[Abstract](14) [FullText HTML] (11) [PDF 688KB](1)

Abstract:
This paper addresses the problems encountered in detecting underwater high-speed moving targets, namely the long active detection cycle leading to discontinuous data, and the inability of passive detection to obtain target range, thus making it difficult to quickly and effectively calculate target motion parameters. To overcome these issues, a short-term motion analysis method for underwater high-speed moving targets based on pseudo-linear Kalman filtering is proposed. This method employs only one initial active detection at the beginning and is primarily a passive-acoustic based target motion analysis(TMA) method within the horizontal plane, supplemented minimally by active detection. Mathematical models are established, and iterative equations are provided. Based on the prior assumption that the target maintains constant-velocity motion for the majority of the observation time, simulation analysis is conducted under constant-velocity motion conditions. The root mean square error (RMSE) of the predicted trajectory is calculated using the Monte Carlo method. Simulation results demonstrate that, compared to bearing-only TMA using purely passive detection, this method achieves significantly smaller errors and enables rapid convergence of the target motion parameter solution within a short timeframe, with convergence achieved within no more than 10 seconds.

Style Transfer-Based Data Augmentation for Side-Scan Sonar Image Classification

BAI Zhongyu, XU Hongli, RU Jingyu, QIU Shaoxiong

, Available online , doi: 10.11993/j.issn.2096-3920.2025-0045

[Abstract](12) [FullText HTML] (3) [PDF 2511KB](1)

Abstract:
Side-scan sonar (SSS) has been widely employed in ocean exploration due to its stability and efficiency on autonomous underwater vehicles (AUVs). However, the difficulty in acquiring SSS images and the limited availability of training samples significantly limit the performance of the deep neural network (DNN)-based SSS image classification. To address this challenge, this paper proposes a multi-scale attention network (MSANet) that utilizes optical-acoustic image pairs for data augmentation to enhance the generalization ability of SSS image classification models. Specifically, shallow and deep features are extracted from different layers of the encoder to comprehensively capture content and style information. Subsequently, a multi-scale attention module (MSAM) is introduced to extract both local and global contextual information of style features along the channel dimension. These style features are then effectively fused with optical features to achieve precise spatial alignment of optical and acoustic features. Finally, the fused features from different layers are rescaled and fed into a decoder to generate high-fidelity synthetic SSS image samples, which are used to train the SSS image classification network. Experimental results on real-world SSS datasets demonstrate that the proposed style transfer approach effectively generates high-quality synthetic SSS image samples, thereby improving the DNN-based SSS image classification performance.

Adaptive Observation Methods for Temperature and Salinity Data Based on the Dolphin Deep-sea Profiling Buoy

LIU Chu, HOU Fei, YANG Zerong, ZHOU Lingxiao, ZHU Xinke

, Available online , doi: 10.11993/j.issn.2096-3920.2025-0025

[Abstract](11) [FullText HTML] (7) [PDF 1936KB](1)

Abstract:
During the profiling observation process of deep-sea profiling buoys, appropriate sensor sampling frequency must be implemented to reduce power consumption in data acquisition and transmission due to low-power design requirements. Based on the independently developed "Dolphin" deep-sea profiling buoy, this study establishes an optimized model for temperature-salinity data collection strategy using adaptive genetic algorithm through simulation analysis combining peripheral device power consumption specifications and conductivity, temperature, depth(CTD) profile data from Argo program. Leveraging the strong correlation characteristics between adjacent profile data, the proposed method plans new-round sampling schemes by analyzing historical profile data features, while adaptively adjusting sampling frequency according to data variation trends during acquisition. The simulation analysis based on real-world data demonstrates that the profile data acquired through this observation method exhibits minimal deviation from actual measurements, while significantly reducing the overall power consumption of the profiling buoy compared to the pre-programmed deep-layer observation scheme.

Research on drag reduction optimization of foldable solar fins for UUV

WANG Chenyu, PENG Likun, CHEN Jiabao, CHEN Jia, WANG Huarui, PAN Wei

, Available online , doi: 10.11993/j.issn.2096-3920.2024-0168

[Abstract](10) [FullText HTML] (5) [PDF 2228KB](0)

Abstract:
Focusing on the endurance bottleneck faced by unmanned underwater vehicles in missions such as ocean observation and resource exploration, this paper concentrates on the hydrodynamic performance optimization of a novel foldable solar wing. To balance computational efficiency and optimization accuracy, a parametric model of the wing is established in CAESES software with variables including wing point coordinates, rounding factors of wing edges, wing gaps, and gaps between the wing and the hull. Innovatively, a hybrid optimization framework combining Sobol global sampling and the NSGA-II optimization algorithm is constructed: Firstly, the Sobol algorithm is used to generate 80 sample points within the threshold space of each variable to fully explore the design space, followed by multi-generation optimization through NSGA-II. To avoid the accuracy degradation of traditional surrogate models, a coupled computational process integrating high-precision hydrodynamic solutions and optimization algorithms is established, enabling automatic co-simulation between CAESES and STAR-CCM+ software. Hydrodynamic analyses are conducted on UUVs equipped with wings of different shapes to explore the impact of different parameter combinations on total drag. The optimization results indicate that a certain height difference between the two wing sections protruding from the hull is beneficial for reducing total drag. Flow field analysis shows that the optimized shape effectively suppresses energy dissipation caused by turbulence. The proposed technical route of parametric modeling, intelligent optimization, and high-precision verification not only reduces the straight-line drag of the new configuration UUV but also provides a methodological reference for the optimization of complex appendages, possessing significant engineering value for improving the energy utilization efficiency of underwater equipment.

Evaluation of Information Usefulness in Human-Machine Interaction Based on Information Template Libraries

CHEN Weichang, XIAO Gang, DU Linlin, MA Jingjing, WU Tao

, Available online , doi: 10.11993/j.issn.2096-3920.2025-0022

[Abstract](11) [FullText HTML] (4) [PDF 557KB](0)

Abstract:
In response to the problem of incomplete coverage of interactive content in human-computer interaction evaluation, based on the analysis of the characteristics of information transmission in different sentence structures, the concept of information usefulness is proposed and included in the indicator system of human-computer interaction evaluation. Based on this, the core elements of interactive content are used as information templates to construct a hierarchically classified library of information templates and to design a method for calculating information utility by comparing with the library. Case studies are conducted to verify the rationality of the construction of the information template library and the effectiveness of the information utility calculation method, which can provide references and insights for the evaluation of human-computer interaction.

Application of state estimation algorithm for UUV docking

CHEN Weixin, LIU Tao, ZHANG Tao, LIU Feng

, Available online , doi: 10.11993/j.issn.2096-3920.2024-0161

[Abstract](22) [FullText HTML] (11) [PDF 1317KB](2)

Abstract:
Underwater autonomous dynamic docking of unmanned undersea vehicle (UUV) is one of the key technologies for achieving long-range cooperative of UUV. Aiming at the problem of insufficient estimation of the motion state of the docking device in docking, an interacting multi-model adaptive unscented Kalman filter method is proposed to estimate the docking device motion state. Considering that the measurement error of the motion state of the docking device obtained by the UUV sensor is large, the UUV nonlinear observation model is established, and the adaptive unscented Kalman filter (AUKF) algorithm is used to update the observation noise model in real time to reduce the observation error. Considering that the difficulty of the describing the relative motion of UUV and docking device with a single model, the motion state model set of the docking device is established, and the interactive multi-model algorithm is used to describe the motion state of the docking device to improve the filtering accuracy and realize the accurate estimation of the motion state of the docking device by UUV in underwater docking. Based on UUV docking test data, the estimation results of unscented Kalman filter, adaptive unscented Kalman filter and interacting multi-model adaptive unscented Kalman filter are compared. The results show that the accuracy and stability of interacting multi-model adaptive unscented Kalman filter are better than the other two algorithms. It can be applied to underwater autonomous docking scenarios with UUV to improve the success rate.

System effectiveness evaluation of acoustic glider based on ADC model optimization

TANG Shuai, FAN Peiqin, ZHANG Chi, ZOU Jiayun

, Available online , doi: 10.11993/j.issn.2096-3920.2024-0175

[Abstract](8) [FullText HTML] (7) [PDF 711KB](0)

Abstract:
With the development of unmanned technology, acoustic glider is increasingly becoming an advantageous platform for the marine environment observation and underwater target detection. Evaluating the effectiveness has become one of the important topics. Based on the characters and key factors of mission and process, an acoustic glider system effectiveness assessment index system is constructed. This paper improves and optimizes ADC model by considering the effects of marine environment and comprehensive support on acoustic glider. The analytic hierarchy process(AHP)is used to determine the weights of three-level index. Finally, the feasibility of the model are verified by arithmetic examples, which show that the optimized ADC method makes the evaluation results more realistic and provide methodological reference for subsequent of the effectiveness of unmanned equipment.

Prediction method for buckling of deepwater explosion cylindrical shell based on random forest

FU Gaojun, MA Feng, ZHU Wei, JIA Xiyu, WANG Shuang

, Available online , doi: 10.11993/j.issn.2096-3920.2024-0162

[Abstract](13) [FullText HTML] (8)

Abstract:
Under deep-water explosion conditions, pressure-resistant structures such as cylindrical shells will have a different failure mode from that in shallow water - instability buckling. In order to study the conditions for the occurrence of instability buckling of cylindrical shells under deep water explosion conditions and realize the prediction of the buckling state, a numerical simulation model was firstly established to simulate and analyze the results of the buckling of cylindrical shells under the conditions of different charge amount, blast distance and water depth. Based on the simulation results, a random forest model was designed to predict the buckling state. The results show that under the loading conditions of axial explosion in deep water environment, the cylindrical shell may present two macroscopic responses of unbuckling and buckling, and the prediction model constructed by the random forest algorithm can better realize the prediction of the instability state of the cylindrical shell under the specific structural parameters, and the prediction accuracies under the two structures reach 0.9375 and 0.875 respectively, which can provide references for the study of the buckling conditions of the cylindrical shell.

Analysis of the Impact of Shock Waves on the Safe Exit of the Rocket-assisted Vehicle Nose Cap during the Thermal Emission Process of a Concentric Canister Launcher

LIU Gangqi, YUAN Xin, GAO Shan, CUI Canli, HUANG Yuxuan

, Available online , doi: 10.11993/j.issn.2096-3920.2024-0156

[Abstract](13) [FullText HTML] (7) [PDF 1755KB](2)

Abstract:
In response to the impact of shock waves on the safety of the rocket-assisted vehicle nose cap during the concentric tube thermal launch process, computational fluid dynamics (CFD) software was used to numerically simulate the ignition and launch process. The propagation process of shock waves and gas generated by solid rocket motors in the concentric tube was analyzed in detail, and the force variation curve of the nose cap under the action of shock waves was obtained, revealing the force mechanism of the nose cap inside the tube under the action of shock waves. The test data of the shock wave opening process during the field test of a certain product further illustrates the force variation process of the nose cap in the shock wave environment. The research results contribute to a clear understanding of the mechanism of force changes on the nose cap under the shock wave during the thermal emission process of concentric cylinders, and can be used to guide the safety design of the nose cap exiting the cylinder.

The 3D Reconstruction Method of Submarine Cables Based on High-Speed ROV Cruising with Multibeam Sonar

XU Haining, WANG Yong, JING Qiang, DING Tongzhen, YU Fei, SHEN Qingye, CAO Shengzhe

, Available online , doi: 10.11993/j.issn.2096-3920.2025-0036

[Abstract](31) [FullText HTML] (5)

Abstract:
Submarine cables, serving as the critical conduits for power transmission in offshore wind farms, are pivotal to the system's stability. However, due to their complex environments, three-dimensional (3D) reconstruction technology for these cables has become a key method for their inspection and maintenance. Currently, conventional 3D reconstruction methods for submarine cables are costly and less effective in deep-sea environments. Therefore, this paper proposes a 3D reconstruction method for submarine cables based on high-speed Remotely Operated Vehicle (ROV) cruising with sonar, drawing on the concept of synthetic aperture and simplifying calculations through spatial carving. This method comprehensively processes the multiple sonar observation information obtained during the ROV cruising to collectively reflect the spatial occupancy. In the simulation experiments of submarine cable 3D reconstruction, a comparison with mainstream methods was conducted. It is evident that the proposed method not only reduces the cost of submarine cable reconstruction by using conventional multibeam sonar but also achieves higher reconstruction accuracy, demonstrating significant application value and promotion potential.

Test Method for Surface and Underwater Condition of Deep-sea Special Pressure Structure

CHEN Shagu, GAO Yuan, WU Zhirui, WANG Kun, ZHOU Cheng

, Available online , doi: 10.11993/j.issn.2096-3920.2024-0153

[Abstract](82) [FullText HTML] (37) [PDF 1893KB](8)

Abstract:
The head cover is a special pressure structure for deep-sea unmanned systems, which needs to balance long-term underwater pressure and rapid separation function on the water surface. In order to study the stress characteristics and separation performance of deep-sea special pressure structure under surface and underwater condition, full-scale model of the head cover was developed for hydraulic and separation testing. Firstly, based on the existing deep-sea environment simulation test system, a test method is proposed to simulate the deep seawater pressure environment using a cabin device with skin balloons in response to the long-term seawater pressure environment testing requirements faced by the head cover during underwater. Furthermore, a safe and reliable inclined flange connection structure model rapid separation test system was established to meet the separation test requirements of head cover when in the water surface state. The results of the full-scale model test showed that the special pressure structure surface and underwater condition test method is reasonable and feasible. It can not only be used for pressure test and separation test research of the head cover, but also provide some reference for the design and testing of similar pressure structures in other deep-sea equipment.

Multi-Degree-of-Freedom Equipment Shock Response Model Based on Deep Learning

HUANG Qinyi, ZHU Wei, MA Feng, CHEN Si, WANG Shuang

, Available online , doi: 10.11993/j.issn.2096-3920.2024-0143

[Abstract](84) [FullText HTML] (111) [PDF 1516KB](15)

Abstract:
To address the challenge of analyzing the response of multi-degree-of-freedom naval equipment under explosive shock loads, this study proposes a deep learning-based shock response prediction model. Traditional single-degree-of-freedom models cannot effectively analyze the complex shock responses of multi-degree-of-freedom systems. Leveraging deep learning technology, particularly the data feature extraction and nonlinear modeling capabilities of neural networks, this model learns the relationship between the shock spectrum and input shock loads from numerical simulation data, achieving efficient and accurate calculation of shock response spectra at critical points within naval structures. This approach fills the gaps of existing models in handling multi-degree-of-freedom equipment and meets the demand for rapid, accurate analysis of complex system shock responses. Experimental results demonstrate that the model can accurately predict the shock response spectra of multi-degree-of-freedom equipment, with a relative error of less than 8% compared to simulation data, effectively overcoming the limitations of traditional models in multi-degree-of-freedom system analysis.

Latest Articles

Journal News> >

Download> >