水下无人系统学报

Adaptive Underwater Image Enhancement Guided by Degradation-Perception Priors

SHAN Siyao, WANG Wei

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0048

[Abstract](40) [FullText HTML] (13)

Abstract:
To address the limitations of traditional underwater image enhancement methods, where non-physical-model-based approaches often lack degradation-mechanism constraints and may cause under-enhancement or over-enhancement, while physical-model-based approaches rely on stable estimation of key parameters, this paper proposes an adaptive enhancement method guided by degradation-aware priors. The method constructs physically interpretable priors from local statistical image features, thereby preserving the flexibility of non-physical-model-based enhancement while introducing degradation constraints, without requiring target-dataset network training or explicit inversion of a complete underwater imaging model. Specifically, color deviation is first corrected using a minimum attenuation channel strategy, and adaptive luminance equalization is achieved by estimating a non-uniform ambient light field via guided filtering. Local contrast variance, ambient light field deviation, local energy statistics, and gradient magnitude are then mapped to haze distribution, exposure risk, noise distribution, and structural saliency, respectively, to guide local dehazing and multi-scale texture enhancement. Experiments show that the proposed method performs favorably in color deviation, structural preservation, texture detail enhancement, and computational efficiency.

Adaptive LPV-MPC and NFTSM-ESO Cooperative Control Method for Underactuated AUV

ZHANG Yuxuan, YAN Tianhong, CHEN Shuwen

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0067

[Abstract](39) [FullText HTML] (16) [PDF 1913KB](2)

Abstract:
This paper addresses depth-heading tracking and roll stabilization of underactuated autonomous undersea vehicles (AUVs) subject to missing sway/roll actuation, rudder-induced channel coupling, actuator constraints, and time-varying marine disturbances. A cooperative control framework integrating adaptive linear parameter-varying model predictive control (LPV-MPC), a nonsingular fast terminal sliding mode extended state observer (NFTSM-ESO), and a roll proportional-derivative(PD) inner loop is proposed. A roll-angle-scheduled depth-roll-heading LPV error model is first established by describing the variation of elevator and rudder effectiveness under roll attitude, so that the influence of roll on the depth and heading channels can be incorporated into the prediction model. An error-disturbance-driven adaptive LPV-MPC is then developed to solve a constrained control-increment sequence under rudder magnitude, rudder-rate, and roll soft constraints, thereby producing feasible rudder-angle commands. The NFTSM-ESO rapidly and smoothly estimates lumped disturbances caused by currents, wave-induced moments, hydrodynamic uncertainty, and unmodeled dynamics; the estimates are used for feedforward compensation and weight scheduling. A roll PD inner loop is superimposed on the rudder channel to provide high-bandwidth roll damping and reduce roll-heading coupling in the outer prediction loop. Simulation and ablation results demonstrate that the proposed method outperforms standard MPC+LESO, fixed-weight LPV-MPC, and the scheme without roll inner-loop compensation in depth error, heading error, roll suppression, and constraint feasibility.

Research on Sensorless Heavy-Load Startup Method for Subsea Long-Cable-Fed PMSM

ZHAO Xinyu, DONG Hang, YANG Rui, LI Ming

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0046

[Abstract](43) [FullText HTML] (14) [PDF 1947KB](2)

Abstract:
To address the challenges of heavy-load startup and smooth transition in subsea permanent magnet synchronous motors driven by surface vessels via long umbilical cables, this paper proposes a novel sensorless control scheme based on motor-side rotor position estimation and a three-stage smooth transition strategy. First, a distributed architecture comprising "subsea observation and surface control" is constructed. This architecture effectively mitigates the distortion in current and voltage sampling, thereby enhancing the accuracy of rotor position estimation. Second, a three-stage startup method is designed, encompassing I-f open-loop acceleration, given reference frame pre-synchronization, and online compensation of closed-loop errors. Simulation results indicate that the proposed strategy significantly outperforms the conventional two-stage switching method during the handover from open-loop I-f control to closed-loop speed control. Specifically, under heavy load conditions, the peak speed oscillation and the q-axis current ripple amplitude are mitigated by approximately 83.7% and 88.5%, respectively. Consequently, the starting stability under long-cable and heavy-load conditions is markedly improved, further bolstering system reliability and robustness.

Research on Operational Effectiveness Evaluation of Underwater Vehicles Based on Bayesian Optimized BP Neural Network

Pu Shi, Wang Jiarong, Jia Yushan, Fu Xuezhi

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0039

[Abstract](91) [FullText HTML] (45) [PDF 2052KB](8)

Abstract:
Aiming at the problem of operational effectiveness evaluation of underwater vehicles, an evaluation indicator system for reconnaissance operational effectiveness of underwater vehicles has been established, consisting of 9 core indicators and 6 nonlinear interaction indicators. Traditional BP neural networks are prone to fall into local optima in operational effectiveness evaluation. Therefore, Bayesian Optimization (BO) is used to optimize the hyperparameters of the BP neural network, and an operational effectiveness evaluation model of underwater vehicles based on the BO-BP neural network is established. The simulation verification of the model is realized by Python.The operational effectiveness evaluation indicator system and engineering-oriented adaptation scheme constructed in this paper can more accurately reflect the actual operational effectiveness of underwater vehicles, and provide practically valuable references for their equipment demonstration, operational deployment, modification and upgrading. Compared with generalized evaluation schemes, the proposed approach exhibits significant improvements in scenario adaptability, evaluation accuracy and engineering practical value.

Bionic Fish Lateral-Line Triboelectric Underwater Multidirectional Flow Velocity Sensing Technology

LIU Jianhua, YU Aiqiang, WANG Siyuan, XU Peng, XU Minyi

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0065

[Abstract](74) [FullText HTML] (32) [PDF 3504KB](7)

Abstract:
Inspired by the superficial neuromast structure in the lateral line system of fish, this paper proposes a bionic fish lateral-line sensor(BFLS) based on a triboelectric nanogenerator. When an external flow field is applied to the sensor, its dome structure deflects and drives an internal spherical trigger, causing the conductive fabric and fluorinated ethylene propylene film in the sensing unit to undergo contact separation motion, thereby generating a corresponding electrical signal output. By optimizing structural parameters such as the carbon fiber connecting rod of the sensor and the back spring of the sensing unit, the sensing performance of BFLS has been significantly improved. Underwater experiments show that BFLS can effectively sense underwater flow speed and direction at intervals of 22.5°. The coefficient of determination for the linear fit between the BFLS’s output electrical signal and flow speed exceeds 0.9, and the flow speed detection sensitivity is approximately 3.2V/ms⁻¹, demonstrating the sensor's reliable capability to detect underwater flow speed. In addition, the BFLS also has advantages such as simple structure, low manufacturing cost, and self-powered. This sensor is expected to play an important role in underwater flow field monitoring in marine environments.

Calibration Method for SINS/USBL Integrating Sound Speed Correction and Time-Delay Estimation

QIN Hongmao, ZHAO Wenhao, CUI Fengyi, WANG Guangcai, YANG Fan

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0084

[Abstract](72) [FullText HTML] (32)

Abstract:
The integration of the strapdown inertial navigation system(SINS) and the ultra-short baseline(USBL) acoustic positioning system has become the mainstream solution for deep-sea navigation. High-precision sensor calibration is the core enabler for ensuring system performance. Existing calibration methods suffer from insufficient real-time capability during deep-sea acoustic ray bending correction. Moreover, traditional methods often ignore the strong coupling characteristics between random acoustic measurement time-delay jitter and spatial errors. This limitation restricts the calibration accuracy, which in turn limits the operational precision and adaptability of the navigation system in deep-sea scenarios. To address these issues, this paper proposes an online joint calibration method that integrates an effective sound speed table with time-delay estimation. First, a refined mapping table among the transponder depth, the initial grazing angle, and the effective sound speed is established offline. During the calibration process, real-time first-order bilinear interpolation is performed online to circumvent complex iterative computations of ray tracing, thereby achieving real-time correction of sound speed errors. Furthermore, the system calculates historical position and attitude matrices through time-delay backtracking. The residual time delay is modeled as a first-order Markov process to compensate for the time jitter introduced by information reconstruction, ensuring the dynamic update of acoustic measurements. Finally, the sound speed error correction method and the residual time-delay model are incorporated into a state estimation framework. Through joint closed-loop estimation of multi-source errors, the joint calibration of USBL range distortion and installation misalignment angles is realized. Simulation and experimental results demonstrate that the proposed method can effectively suppress deep-sea sound speed fluctuation errors and mitigate the negative impact of measurement latency. Compared with traditional methods without considering deep-sea errors, the proposed calibration approach improves the positioning accuracy by approximately 90.1%.

ncc-audit-test-e0bf2fa9

, Available online

[Abstract](70)

Abstract:

Compact Sensorless Control for Underwater Propulsion Using Single-Current-Sensor Reconstruction

WANG Yu, SUN Guo Rui, XU Lei, WANG Yuankui

, Available online , doi: 10.11993/j.issn.2096-3920.2025-0174

[Abstract](93) [FullText HTML] (42) [PDF 2877KB](6)

Abstract:
The propulsion system of underwater operating platform is typically integrated within narrow compartment, imposing severe constraints on spatial layout. Traditional vector control system for Permanent Magnet Synchronous Motor relies on multiple current sensor and a position sensor. This not only occupies the thermal dissipation space required for power device but also increases the risk of sensor failure in harsh deep-sea environment characterized by high pressure and corrosion.To address these challenges, this paper proposes a compact sensorless control scheme that integrates single current sensor technology with an Extended Kalman Filter (EKF) to simplify the hardware architecture. To overcome the measurement blind spot inherent in single current sensor sampling, the scheme optimizes Space Vector Pulse Width Modulation timing. Specifically, when the duration of the effective voltage vector is insufficient for sampling, a non-zero measurement vector is dynamically inserted to create a stable current sampling window. This effectively suppresses distortion in the reconstructed phase current waveform.Furthermore, to enhance the system response speed under dynamic load, this paper improves the EKF observer by incorporating the load torque as an extended state variable for real-time estimation. A feedforward compensation mechanism is employed to bolster the system ability to suppress load disturbance.Experimental results demonstrate that the reconstructed current waveform closely matches the actual current waveform. Compared to traditional EKF, sliding mode observer (SMO) and Luenberger sliding mode observer (L-SMO), the improved EKF exhibits superior performance in suppressing speed fluctuation and achieving faster recovery.The proposed solution successfully simplifies the hardware structure while achieving control performance with high precision and strong robustness, making it highly suitable for complex underwater operating environment.

Analysis of Damage Effectiveness for Underwater Vehicles Equipped with Terminal Guidance Mechanism

CAI Yi, MA Feng, ZHU Wei, JIA Xiyu, XUE Jianmin

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0037

[Abstract](121) [FullText HTML] (56)

Abstract:
Under the background of the development of underwater precise and efficient damage technology, the damage effectiveness assessment for mobile targets needs to more comprehensively consider the integrated effects of multiple subsystem parameters such as guidance, fuze, and warhead. This paper proposes a damage effectiveness assessment method for underwater mobile weapons that introduces a terminal guidance mechanism. This method retrospectively analyzes the terminal guidance mechanism, introduces two types of error models for control and guidance, and performs Monte Carlo sampling on random trajectories to obtain the explosion point locations. The paper adopts three typical guidance laws, establishes a terminal trajectory model, and thereby achieves damage effectiveness analysis under mobile conditions. Case study analysis shows that compared to the direct aiming method, the sliding mode guidance law has a higher probability of severe damage when striking mobile targets. Taking the proportional navigation method striking a serpentine target as an example, the influence of multiple coupled parameters on the final damage effectiveness is analyzed. The results indicate that to achieve stable and efficient damage, the ammunition needs to reach 2 to 3 times the target speed, a guidance coefficient above 1.5, a relatively ideal fuze activation distance of 2.5m to 4m, and a relatively ideal delay time below 0.4s. This method can comprehensively consider the parameters of multiple subsystems such as guidance, fuze, and warhead, and can more accurately and comprehensively reflect the influence of parameters on damage results under mobile conditions, providing references for the optimal design and operational application of underwater weapons.

Effects of Front-Side Angles of Floating Boxes on the Navigation Performance of an Amphibious Vehicle in Calm Water

ZHANG Zhewei, XU Xiaojun, FENG Yikun, ZHANG Guoqing, JIN Haobin, GE Qiqian, WANG Jiancheng, ZHANG Xin

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0092

[Abstract](92) [FullText HTML] (37) [PDF 5396KB](6)

Abstract:
To clarify the effects of the front-side angles of floating boxes on the resistance, attitude response, and local flow field of an amphibious vehicle in calm water, a specific amphibious vehicle was investigated based on the Reynolds-Averaged Navier–Stokes(RANS) equations. Within the framework of computational fluid dynamics(CFD), the shear stress transport(SST) k-ω turbulence model and the volume of fluid (VOF) free-surface model were employed to analyze the hydrodynamic effects of the front-side geometric angles of the floating boxes. By varying the front inclined angle α₁ and the outer guide-surface angle α₂, the resistance, pitch, heave, free-surface patterns, and pressure distributions were comparatively analyzed at a fixed displacement condition and a speed of 6m·s⁻¹. Towing-tank experiments were also conducted to validate the numerical method. The results show that, within the parameter ranges investigated, as α₁ increases from 15° to 35°, the normalized resistance R/∇ increases from 2.509 to 2.690, corresponding to an increase of approximately 7.2%. At larger α₁, the high-pressure region near the leading part of the floating box expands, accompanied by intensified free-surface disturbance and wave-making effects. The pitch angle first decreases slightly and then gradually increases, whereas the heave displacement first decreases and subsequently exhibits small fluctuations. Among the discrete cases investigated, the minimum heave displacement of 0.009 529 m occurs at α₁=25°, where R/∇ is 2.542, approximately 1.3% higher than that at α₁=15°, and the pitch angle is 4.255°. Therefore, the favorable values of α₁ differ among individual hydrodynamic performance indices, and an overall optimum cannot be identified without an explicitly defined multi-objective evaluation criterion. In comparison, variations in α₂ produce a smaller resistance response. At α₂=30°, R/∇ reaches the maximum value of 2.650 among the investigated cases, accompanied by relatively strong transverse-wave interference. The results indicate that the front-side angles of the floating boxes affect navigation resistance through changes in local pressure distribution, vehicle attitude response, and free-surface wave patterns, providing a reference for the selection and further optimization of the front-side geometric parameters of the floating boxes for the vehicle investigated in this study.

A Review of Key Technologies for Underwater Resident AUV Docking

WANG Yun, WANG Meng, ZHANG Xilin, GUO Lei, SUN Zhilei

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0051

[Abstract](116) [FullText HTML] (55) [PDF 19234KB](10)

Abstract:
Docking of autonomous underwater vehicles (AUVs) is a key link for underwater resident systems to achieve long-term continuous operation, energy replenishment, and data return, and its performance directly affects system operation efficiency and long-term reliability. In recent years, AUV docking technology for complex marine environments has evolved from early functional verification into a systematic technical framework covering docking station structures, positioning and navigation, planning and control, as well as energy and data management, exhibiting a development trend characterized by the deep coupling of structure, sensing, control, and energy. However, current studies still suffer from insufficient environmental adaptability, limited robustness under sensing degradation, inadequate evaluation of long-term operational reliability, and insufficient system-level collaborative design. This paper systematically reviews the recent progress in key technologies for AUV docking in underwater resident scenarios, including docking station structures, stage-based positioning and navigation, planning and control, and energy and data management, and analyzes their intrinsic relationships and the challenges they face. The results show that AUV docking technology is shifting from isolated performance improvement toward system-level fault tolerance and collaborative optimization under uncertainty.

Loss Evolution Characteristics of a Micro Axial-Flow Turbine Under Variable Operating Conditions

LI Xiangyu, YI Jinbao, MA Weifeng, WEN Menggang, HUANG qilong, YANG jun

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0077

[Abstract](99) [FullText HTML] (53) [PDF 3076KB](5)

Abstract:
A study on the evolution mechanism of off-design losses is carried out to address the efficiency issues of micro axial-flow turbines in micro underwater vehicles operating over a wide speed range. A thermo-mechanical-fluid multi-physics coupling model is developed using the Modelica language, incorporating micro-scale nonlinear loss corrections such as disk friction loss, partial admission loss, tip clearance leakage loss, and exit velocity loss. Through transient simulations under a typical mission profile and full-speed-range parameter sweeps, the stage-wise evolution of the loss mechanism under variable operating conditions is revealed: during low-speed cruise, exit velocity loss, disk friction loss, and partial admission loss share comparable contributions, forming a co-dominant regime, while leakage loss remains relatively stable; during high-speed dash, partial admission loss and disk friction loss become the dominant loss sources. Further investigation shows that system performance is significantly influenced by both operating and geometric parameters. Increasing inlet pressure and temperature can reduce the loss proportion, but proper matching between inlet parameters and aerodynamic characteristics must be ensured. Controlling tip clearance, appropriately increasing the number of nozzles to raise the partial admission ratio, and adopting a small rotor exit angle can effectively suppress kinetic energy dissipation and micro-scale penalties, thereby improving the turbine efficiency over the whole operating range. The findings provide theoretical support and simulation-based guidance for energy management, definition of high-efficiency operating intervals, and structural optimization of micro power systems.

Research on path planning of underwater inspection robot based on improved ant colony algorithm

Shen Zhaowen, Pan Xinfeng, Huang Mengshi, Yang Tianci, Zhang Wenze, Yi Zhengyao

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0066

[Abstract](148) [FullText HTML] (71)

Abstract:
Aiming at the problems of slow convergence speed, easy to fall into local optimum and poor environmental adaptability of traditional ant colony algorithm in path planning of underwater inspection robot, an improved ant colony algorithm suitable for path planning of underwater inspection robot is proposed. Firstly, the ant colony pheromone update strategy is optimized, and a comprehensive evaluation function including path length, safety, smoothness and water flow adaptability is constructed. The pheromone release coefficient is dynamically adjusted by the path fitness. Adjust the pheromone volatilization coefficient to adjust with the iterative process, and speed up the convergence speed in the later stage of the algorithm iteration; the water flow factor and obstacle distance factor are introduced into the heuristic function to guide the ants to choose a better path. Then the simulated annealing mechanism is integrated to avoid the algorithm falling into local optimum. Finally, the B-spline curve is used to smooth the path to reduce the energy consumption of the underwater vehicle. The simulation results show that the improved ant colony algorithm not only has faster convergence speed and fewer turning points, but also significantly shortens the optimal path length.

AUV Path Planning Method Based on RRT* and Improved Artificial Potential Field

LIU Yuqing, CHEN Mojiang, HAO Chengpeng

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0025

[Abstract](146) [FullText HTML] (72) [PDF 6989KB](10)

Abstract:
Aiming at the problem that the complex ocean current environment significantly affects the navigation efficiency of autonomous underwater vehicles (AUVs), this paper proposes an energy consumption-optimized path planning algorithm (FAPF-Bi-RRT*) that integrates bidirectional rapidly-exploring random tree with flow field-guided sampling and improved artificial potential field method. Firstly, a multi-strategy sampling probability correction method based on flow field information is designed. By calculating the consistency measure between the flow field and the target direction, the sampling weight is dynamically adjusted to guide the random tree to expand toward downstream and low-energy-consumption areas, thus solving the blind search problem of the traditional RRT*. Secondly, an adaptive flow field potential field model is established, which converts the pushing and resistance effects of ocean currents into potential field gradients. An adaptive adjustment mechanism of flow field gain is introduced into the artificial potential field to assist node expansion and reconnection in real time, thereby avoiding countercurrent traps. Finally, a composite cost function including path length and energy consumption is constructed for progressive optimization. Simulation results under real terrain and flow field data show that while ensuring probabilistic completeness and obstacle avoidance capability, the proposed algorithm reduces the path length by approximately 14% and the path energy consumption by approximately 10% compared with the traditional Bi-RRT* algorithm in strong flow field environments. This realizes efficient and energy-saving path planning in complex marine environments.

Hydrodynamic Characteristics of Underwater Vehicles Under the Flow Disturbance Condition of Submarine

HUANG Chuang, HAO Sihan, ZHANG Chenchen, QIN Kan, XU Haiyu

, Available online , doi: 10.11993/j.issn.2096-3920.2025-0165

[Abstract](422) [FullText HTML] (157) [PDF 4530KB](11)

Abstract:
A numerical model of an underwater vehicle in the flow field of the submarine is established by combining the overset grid technology and the moving reference frame. The rationality of the numerical method is verified by comparing the simulation results with the standard model test data reported in the literature. The effects of the flow around the submarine on the external flow field and hydrodynamic characteristics of the vehicle are investigated. The results show that the underwater vehicle is most significantly affected by the flow around the submarine when it is located at the front side of the submarine, and the drag coefficient, lateral force coefficient, and yaw moment coefficient of the vehicle exhibit nonlinear characteristics. As the distance between the vehicle and the submarine increases, the influence of the flow around the submarine on the hydrodynamic characteristics of the underwater vehicle gradually weakens, and the influence can be neglected when the distance is greater than 1.75 times the diameter of the submarine. With the increase in the speed of the submarine, the hydrodynamic parameters of the vehicle are more strongly affected by the flow around the submarine. The research results can provide a reference can provide a reference for the prediction of launching initial trajectory of underwater vehicles.

Entry-Exit Water Perception and Damage Monitoring of Rotors for Cross-Media Vehicle

Bo Fusen, Pan Deng, Fang Ziyi, Fu Bowei, Li Hanyang, Liu Zhihai

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0045

[Abstract](175) [FullText HTML] (80)

Abstract:
In the process of medium conversion, the rotor of cross-medium vehicle faces complex instantaneous strain problems due to the sudden change of medium physical parameters. Long-term alternating load can easily lead to the initiation, propagation and even fracture of blade microcracks, which seriously threatens the operational reliability and task continuity of the vehicle. In order to solve the core problems of rotor inlet and outlet water state perception and damage monitoring, this paper constructs a complete technical system of ' signal acquisition-processing-feature extraction-model training-state perception ': The experimental device controls the inlet and outlet water of the rotor drive structure through the guide rail, and uses the fiber grating demodulator to collect the whole process signal of the rotor inlet and outlet water in real time. The singular point elimination method is selected to identify and eliminate the abnormal data points, and the characteristics of the signal entering water, water and outlet water are completely retained. The timestamp and amplitude information of the time series are retained by image conversion, and the time correlation characteristics are enhanced and the redundancy is eliminated. Then the feature signal and the converted image are used for model training. The results show that the training model realizes the accurate identification and real-time damage warning of the rotor inlet and outlet water state of the cross-media vehicle, provides decision support for the route adjustment of the vehicle, effectively improves its intelligent perception and penetration ability, and promotes the in-depth application of optical monitoring technology in the field of new weapon equipment and national defense.

A UAV-USV Collaborative Water Depth Inversion Model Based on Multi-Modal Heterogeneous Sensors

Lin Zibo, Lai Yunshan, Zhang Runda

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0006

[Abstract](176) [FullText HTML] (81)

Abstract:
In response to the low accuracy of water depth inversion in complex water environments using a single remote sensing platform or homogeneous data, this paper proposes a collaborative water depth inversion model between unmanned aerial vehicles and unmanned ships based on multimodal heterogeneous sensors. The model first extracts the target water body through multi spectral index joint extraction, and then integrates dual platform multimodal sensor data. It uses gradient boosting decision tree (GBDT) to mine nonlinear mapping relationships and invert water depth. The experiment shows that the model has a coefficient of determination of 0.971, a root mean square error of 0.31 m, and the water depth fluctuation is controlled within ± 0.2 m. The accuracy is significantly better than traditional comparison methods.

Design and Application of Vehicle Mounted Industrial CT Nondestructive Testing System and Underwater Unmanned Equipment Defect Recognition Software

MA Junli, WANG Shuai, LIU Wenbin, GAO Haoyang, LI Chenhui, ZHANG Xitong, SU Chenghai

, Available online , doi: 10.11993/j.issn.2096-3920.2025-0171

[Abstract](570) [FullText HTML] (335) [PDF 4796KB](40)

Abstract:
To satisfy the inspection demands of unmanned undersea equipment that are stored in remote areas, unsuitable for frequent transportation and difficult to conduct centralized testing, this paper builds a vehicle-mounted industrial computed tomography(CT) nondestructive defect detection and identification system with hardware subsystems including power supply, X-ray source, detection, acquisition and transmission. Combined with deep neural networks, an image recognition software for defects in industrial CT inspection products is developed to process inspection images and realize the identification and marking of internal product defects. Practical test results demonstrate that the proposed system can effectively detect and identify pores, cracks, debonding and other internal defects in high-energy fillers of unmanned undersea equipment. The imaging accuracy exceeds 98%, and both the missed detection rate and false alarm rate of defect identification are no more than 5%. This system can adapt to different materials and shapes of the products, and can operate in various environments. It has a promising application prospect.

A TCN-Attention-Based Pseudo-Velocity Measurement Generation Method for Loosely Coupled SINS/DVL Integrated Navigation

WANG Guoxiang, HAN Xingcheng, GAO Shengwen, WANG Ling

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0047

[Abstract](630) [FullText HTML] (343) [PDF 3674KB](59)

Abstract:
DVL unavailability degrades the accuracy of loosely coupled SINS/DVL integrated navigation for autonomous underwater vehicles. To address this problem, a pseudo-DVL velocity measurement generation method based on a temporal convolutional network with an attention mechanism (TCN-Attention) is proposed. The method uses the angular velocity and specific force measured by the inertial measurement unit (IMU), together with the attitude, position, and velocity obtained from inertial navigation computation, as sequential inputs. During the DVL-available stage, supervised samples are constructed using DVL velocity for offline network training. During the DVL-unavailable stage, the trained model outputs pseudo-velocity measurements, which are incorporated into the extended Kalman filter (EKF) update to suppress inertial error accumulation. Causal dilated convolutions are adopted to extract temporal features, and an attention mechanism is introduced to enhance the representation of key dynamic segments such as turning, acceleration, and deceleration. Simulation results based on 16 trajectory datasets show that, compared with the temporal convolutional network (TCN) and the gated recurrent unit with attention model (GRU-Attention), the proposed method achieves better performance in east- and north-velocity errors as well as absolute trajectory error, and reconstructs trajectories closer to the ground truth, demonstrating its effectiveness and robustness under continuous DVL-outage conditions.

Research on Factors Affecting Leakage Current of Lithium Reserve Battery Packs

GAO Xinlong, JIA Bin, LIN Pei, CHENG Haichao, LI Xuehai

, Available online , doi: 10.11993/j.issn.2096-3920.2026-0083

[Abstract](550) [FullText HTML] (309) [PDF 995KB](76)

Abstract:
The distribution of leakage current of lithium reserve battery packs where the battery cells are in a common electrolyte state was studied, using the equivalent circuit simulation calculation method. The effects of the series cells quantity, electrolyte conductivity, and the structural characteristics of the injection tube on the leakage current were analyzed. A leakage current experimental device for battery modules was established. By comparing the measured results with the simulation results, the validity of the simulation calculation method was confirmed.

Latest Articles

Journal News> >

Download> >