水下无人系统学报

Progress of AUV Intelligent Swarm Collaborative Task

HU Qiao, ZHAO Zhenyi, FENG Haobo, JIANG Chuan

2023, 31(2): 189-200. doi: 10.11993/j.issn.2096-3920.2023-0002

[Abstract](242) [FullText HTML] (39) [PDF 2032KB](124)

Abstract:
With developments in the intelligent swarm technology, the cooperation of autonomous undersea vehicle (AUV) swarms in performing underwater tasks has become an inevitable trend for the future development of such tasks. However, owing to the particularity of underwater environments, collaborative tasks of underwater multi-vehicle swarms encounter major challenges. This paper provides an overview of the research progress on AUV swarms. From the perspective of cooperative hunting, path planning, formation control, and so on, multiple tasks and key technologies of intelligent swarms developed worldwide are introduced, and some recent results obtained by the authors based on cooperative hunting and multi-path underwater planning tasks are summarized. Additionally, through a review and analysis of existing research results, some references for the future development of underwater multi-vehicle swarm tasks are provided.

CPG Motion Control for a Bionic Manta Ray Robot Fish Propelled by Flexible Pectoral Fin

NAN Kaigang, JIANG Sheng, ZHANG Jinhua, CHENG Haiyan

2023, 31(2): 201-210. doi: 10.11993/j.issn.2096-3920.202203009

[Abstract](168) [FullText HTML] (19) [PDF 4780KB](45)

Abstract:
Considering a manta ray as the bionic object, this paper proposes a design scheme for a bionic manta-ray robot fish with flexible pectoral fin propulsion. A motion control strategy is designed based on a central pattern generator(CPG), the influence of control parameters is analyzed, and the performance of multiple motion modes is tested. First, the movement characteristics of the manta ray are analyzed, and an overall design scheme for the bionic manta ray robotic fish is proposed under the actual bionic function requirements. Then, to realize continuous and stable swimming of the bionic manta-ray robot fish in different swimming modes, a CPG motion control method is designed, and the influence of the control parameters is analyzed for the output signals of the designed CPG topology network structure. Finally, a further swimming test is conducted on the maneuverability of the bionic manta-ray robotic fish; the results verified that the bionic manta-ray robotic fish can flexibly realize various motion modes such as linear cruise, steering, floating, and diving, and therefore, it has excellent motion performance and application prospects.

Effect of Material Density on the Tail-slapping Characteristics ofSupercavitating Projectiles

HUANG Baozhu, LI Daijin, HUANG Chuang, GU Jianxiao, LUO Kai

2023, 31(2): 211-220. doi: 10.11993/j.issn.2096-3920.202204014

[Abstract](151) [FullText HTML] (11) [PDF 3052KB](29)

Abstract:
Generally, the tail-slapping motion of underwater supercavitating projectiles has a significant impact on their ballistic stability and operational performance. In this study, a numerical model of such supercavitating projectiles integrating a dynamic mesh is established to investigate the impacts of the tail-slapping characteristics of supercavitating projectiles under different material densities. Moreover, the ballistic and hydrodynamic characteristics of projectiles using aluminum alloys, structural steels, and tungsten alloys are scrutinized. Consequently, the attack angle, pitch angular velocity, and hydrodynamic coefficient of the tail-slapping motion of the projectile are found to demonstrate periodic changes with the material density under a given export kinetic energy. Additionally, the greater the material density, the longer the period of tail slapping; the slower the speed decay, the smaller the impact on the vertical speed of the projectile. Interestingly, the structural steel projectile demonstrates the best ballistic performance among the three candidate materials.

Adaptive Flocking Control for Crowded UUV Swarm with Time-Delay Constraint

LIANG Hongtao, KANG Fengju

2023, 31(2): 221-228, 258. doi: 10.11993/j.issn.2096-3920.202112012

[Abstract](88) [FullText HTML] (15) [PDF 2848KB](32)

Abstract:
To address the flocking control problem of a crowded unmanned undersea vehicle (UUV) swarm under time-delay constraints, an adaptive flocking control approach was investigated using a multiscale bio-inspired mechanism. First, an adaptive flocking interaction model with a bio-inspired optimal neighbor selection strategy was established to robustly switch between single neighbor following and multiple neighbors following, which ensures the minimum quantity and optimal distribution. Second, considering time-delay constraints, a flocking controller was developed by incorporating a consensus protocol, potential field function model, and disturbance observer into the proposed interaction model, thereby guaranteeing collision avoidance and connectivity maintenance. Finally, by virtue of the Lyapunov theorem, state consensus under the time-delay condition is proved. Simulation results verify the effectiveness and superiority of the proposed control method.

Inversion Estimation of AUV Attitude Based on Pressure Sensor Array

WANG Chongyang, YU Huapeng, LI Ziyuan, ZHAO Dexin

2023, 31(2): 229-236. doi: 10.11993/j.issn.2096-3920.202202011

[Abstract](164) [FullText HTML] (12) [PDF 1936KB](53)

Abstract:
Attitude is an essential parameter in autonomous navigation. Existing high-precision navigation sensors are expensive and large. Therefore, miniaturized and low-cost autonomous navigation technologies have become the research focus. Inspired by biomimetics, an inversion estimation model of autonomous undersea vehicle(AUV) attitude based on a pressure sensor array was constructed. The model takes known pressure and attitude data as input and trains the model parameters to use the pressure data to invert attitude information. The effectiveness of the proposed method was verified through multiple experiments with different trajectories, depths, speeds, and other factors. The experimental results show that using the proposed method, the estimated pitch angle error is below 2.069 9°, and the estimated roll angle error is below 2.990 8°. Furthermore, there are no cumulative attitude errors; this, implies that the proposed method has strong potential for applications in autonomous navigation systems.

Energy-optimal Path Planning Algorithm for Unmanned Surface Vessel Based on Reinforcement Learning

LI Peijuan, YAN Tingwu, YANG Shutao, LI Rui, DU Junfeng, QIAN Fufu, LIU Yiting

2023, 31(2): 237-243. doi: 10.11993/j.issn.2096-3920.202203002

[Abstract](184) [FullText HTML] (106) [PDF 1349KB](88)

Abstract:
To address path planning for unmanned surface vessels(USVs) in ocean environments affected by physical disturbances, such as ocean currents and obstacles, an energy-optimal algorithm based on improved reinforcement learning is proposed. First, a two-dimensional ocean-current model comprising multiple random vortices and a plane kinematic model of the USV is established. Then, the study determined whether a waypoint is reachable using the relative velocity relationship between the USV and the ocean current. An improved reward function, an action set, and a state set are used to obtain a global optimal path, and the B-spline method is applied to smooth the plan. Finally, numerical simulations are performed in two typical environments. The simulation results show that a path with optimal energy consumption and smoothness can be planned based on the proposed algorithm.

Route Optimization of on Call Submarine Search Based on Genetic Algorithm

ZHANG Ning, KOU Xiaoming, LI Bin, LI Qian, ZHOU Jingjun

2023, 31(2): 244-251. doi: 10.11993/j.issn.2096-3920.2022-0002

[Abstract](160) [FullText HTML] (118) [PDF 1414KB](55)

Abstract:
To address the situation wherein an enemy submarine maneuvers in an unknown course when an anti-submarine surface ship is called, a genetic algorithm-based submarine search method is proposed. The method combines the ship sonar detection model, enemy submarine target motion model, ship search motion model, and search path discovery probability calculation model. It introduces information confidence into the discovery probability calculation model, which enhances its reliability. Subsequently, a genetic algorithm is used to solve the optimal heading angle and speed of each section of a single ship and double ships in the on-call search process, and the optimal paths of single and double ships on call spiral search are determined. Finally, the variation law of the discovery probability of searching the target is formulated under the conditions of changing only the heading angle and changing both the angle and speed of single and double ships. The results indicate that, compared to the traditional spiral algorithm, the ship search mechanism with increasing change speed is more flexible and can improve the discovery probability. When the search force is sufficient, using a multiship formation search can significantly improve the discovery probability. The results provide a tactical reference for surface ship searches and submarine attacks.

Pressure Sensor Error Compensation Algorithm Based on MEA-BP Neural Network

SHI Hao, FAN Hui, LI Jianchen, ZHAO Runhui, LI Ya

2023, 31(2): 252-258. doi: 10.11993/j.issn.2096-3920.202205002

[Abstract](109) [FullText HTML] (60) [PDF 2122KB](26)

Abstract:
Piezoresistive pressure sensors are sensitive to environmental changes. Ambient temperature changes would produce thermal drift, which affects sensor performance. This study entailed the use of the mind evolutionary algorithm(MEA)-back propagation(BP) neural network algorithm to establish an error compensation model for piezoresistive pressure sensors. The model uses the MEA algorithm to optimize the initial weight and threshold of the neural network, which reduces the possibility of the training falling into local optimization owing to the uncertainty of the initial value. The Levenberg–Marquardt algorithm replaces the gradient descent method to accelerate the convergence speed of the neural network and increase the reliability of the compensation algorithm. The results of the simulation show that, compared with those of the BP neural network compensation algorithm and genetic algorithm(GA)-BP neural network, the expectation of the root mean square error of the MEA-BP algorithm is lower by 48.7% and 8.29%, respectively. The standard deviations are reduced to 5% and 4% of those of the BP and GA-BP neural networks, respectively. This demonstrates that the BP neural network compensation method optimized by the MEA algorithm can compensate for the influence of temperature more accurately and that the compensation result is more reliable.

Treatment Method for Multi-AUV Cooperative Positioning Underwater Acoustic Propagation Delay Based on IMM Algorithm

CHEN Shijie, LIU Xixiang, HUANG Yongjiang, ZHANG Caixia, TAO Yujie, TONG Jinwu

2023, 31(2): 259-268, 277. doi: 10.11993/j.issn.2096-3920.202204011

[Abstract](212) [FullText HTML] (11) [PDF 3507KB](52)

Abstract:
To address the delay problem of underwater acoustic detection and communication in a master-slave autonomous undersea vehicle(AUV) cooperative positioning system, a time-delay processing method based on an interacting multiple model(IMM) algorithm is proposed. First, a cooperative positioning calculation model for a multi-AUV system is established. Aiming at the nonlinear motion equation and nonlinear measurement of the system, the positioning error caused by the propagation delay of underwater acoustic signal in the cooperative positioning result of extended Kalman filter(EKF) is analyzed. Second, the problem that delay EKF(DEKF) cannot accurately track the motion state of the maneuverable target AUV in handling time delay is described. Finally, the IMM-DEKF algorithm is designed, the appropriate motion models are selected as the sub filters, the model probability is updated via innovation, the motion states of the main AUVs are accurately tracked, the estimation errors of the state value of the main AUVs from the slaver AUV’s filter are reduced, and the positioning accuracy of the overall cooperative system is improved. The simulation results verify that the proposed algorithm effectively improves the prediction accuracy of the main track of the AUVs from the slave AUV filter based on the conventional EKF and improves the overall positioning accuracy of the cooperative positioning system.

A Detection Method of Magnetic Anomaly Signal in Offshore Waters

DU Defeng, CHEN Shuai, WANG Lei, MENG Fankai

2023, 31(2): 269-277. doi: 10.11993/j.issn.2096-3920.202203008

[Abstract](251) [FullText HTML] (32) [PDF 4294KB](72)

Abstract:
Magnetic anomaly signal detection has broad application prospects in offshore defense, which is an urgent problem for realizing remote and long-term monitoring. In this paper, an underwater magnetic anomaly detection model is proposed, and wavelet domain analysis and orthogonal basis decomposition detection methods are introduced. Through the analysis of magnetic anomaly signal characteristics and wavelet decomposition frequency, adaptive layer determination method is proposed, and independent processing flow of denoising orthogonal basis decomposition energy detection was performed. To address the problem of wavelet domain denoising under different decomposition scales, different shrinkage coefficients are used to independently deal with detail coefficients to improve the effect of high-frequency noise suppression. The results show that compared with the traditional wavelet denoising, the signal-to-noise ratio of this method is improved by approximately 27%, and the false recognition rate is reduced by 39%.

Velocity Estimation of Moving Acoustic Source Based on Acoustic Intensity Interference with Dual Hydrophones in Shallow Water

YAO Yuan, SUN Chao, XIE Lei, LIU Xionghou

2023, 31(2): 278-284. doi: 10.11993/j.issn.2096-3920.2022-0068

[Abstract](150) [FullText HTML] (24) [PDF 2529KB](46)

Abstract:
Considering the difficulty in estimating the acoustic source motion velocity using an interference striation pattern of a broadband continuum spectrum in a low-frequency analysis and recording(LOFAR) spectrum at a low signal-to-noise ratio(SNR), a moving sound source velocity estimation method based on the interference fluctuation of the line spectrum acoustic intensity with dual hydrophones is proposed. This method utilizes the motion parameters of the acoustic source to resample the temporal interference fluctuation of the line spectrum acoustic intensity and obtains the interference fluctuation of the line spectrum acoustic intensity at different frequencies that satisfy a proportional relationship. The time of the closest position of approach(CPA) and the range-to-velocity ratio of each hydrophone are determined by establishing the correlation coefficient cost function of the acoustic intensity interference fluctuation between the two frequency line spectra. By combining the position relationship between the acoustic source and the dual hydrophones, the velocity of the moving acoustic source can be calculated. The simulation results demonstrate that the proposed method can effectively estimate the velocity of a moving acoustic source and achieve good performance for a low-speed acoustic source in background noise.

Subsea Broadband Reverberation Modeling and Simulation of High-speed Motion Sonar

YANG Jiayi, YANG Yunchuan, LI Yongsheng, SHI Lei, YANG Xiangfeng

2023, 31(2): 285-290. doi: 10.11993/j.issn.2096-3920.2022-0003

[Abstract](205) [FullText HTML] (26) [PDF 1610KB](41)

Abstract:
Subsea reverberation is the main source of interference when torpedoes detect targets in shallow waters where the sound velocity approximately satisfies the negative gradient distribution. In order to obtain the reverberation signal accurately, conveniently and reproducibly, and study its statistical characteristics, this paper proposes a high-speed motion sonar subsea broadband reverberation simulation method based on the Bellhop model. In this method, the seabed area that contributes to the reverberation is divided into several non-uniform scattering units, and the scattering characteristic function of each unit is calculated. The propagation loss is calculated by the Bellhop model. Subsea reverberation analog signal with main influencing factors such as doppler shift. Finally, the time-frequency domain analysis and similarity comparison between the simulated signal and the actual flight signal are carried out. The results show that using the dot product ratio of the probability density distribution of the instantaneous value of the simulated signal and the measured data as the evaluation standard, the similarity between the method in this paper and the actual flight data is greatly improved, which verifies the validity of the model.

Single-loop Robust Control of Permanent Magnet Synchronous Motors Based on Generalized Predictive Control

YANG Xi, LIU Guohai, LIU Yabing, CUI Jialun

2023, 31(2): 291-297. doi: 10.11993/j.issn.2096-3920.202203007

[Abstract](111) [FullText HTML] (168) [PDF 1750KB](26)

Abstract:
To improve the dynamic performance and robustness of the control system of permanent magnet synchronous motors (PMSMs) for unmanned undersea vehicles (UUVs) and simplify the structure of the control system, a single-loop control strategy based on generalized predictive control (GPC) is proposed. A novel objective function is optimized to design a single-loop speed-current PMSM controller based on the continuous-time PMSM model. Furthermore, a current-limiting strategy is employed to avoid any damage to the motor and inverter controller caused by extremely high currents. The controller has a simple structure. Further, robust PMSM control can be realized by only adjusting the prediction time domain; this is easy to implement in engineering applications. Simulation results show that the control strategy can effectively suppress the influence of external disturbance and internal parameter perturbation, thus realizing a fast dynamic response and strong robustness of the PMSM for UUVs.

Simulation Analysis of Fuel Pump Piston Oil Film Friction Heat Generation Based on CFD Method

CHEN Wenjie, LI Yongdong, BAI Changqing

2023, 31(2): 298-304, 315. doi: 10.11993/j.issn.2096-3920.202111003

[Abstract](117) [FullText HTML] (17) [PDF 3202KB](14)

Abstract:
Aiming at the friction heat generation of fuel pump piston oil films for underwater gas turbines, combined with the computational fluid dynamics dynamic grid and slip grid methods, user-defined function programming is performed according to the piston motion equation. Furthermore, a fuel-pump piston oil-film simulation calculation model is established considering the viscosity-temperature characteristics of the oil. A modeling analysis method for the frictional heat generation of the piston oil film is presented. Using the proposed analysis method, the frictional heat generation of the piston oil film was simulated and analyzed, and the influences of the outlet pressure, wall temperature, and rotational speed on the temperature change caused by the oil film frictional heat generation were studied. The following conclusions are obtained. 1) When the inlet pressure was 0.5 MPa, the changes in the outlet pressure had little effect on the temperature rise of the oil film. Moreover, the temperature rise at the top of the oil film was the largest, and the temperature rise could reach up to approximately 4 K at the rate of 2 250 r/min. 2) In the range of 300 K to 373 K, the temperature increase at the top of the oil film decreased by approximately 50% for every 20 K increase in wall temperature. Moreover, the temperature increase at the top of the oil film at 373 K was only 9.2% of that at 300 K. 3) The temperature increase of the oil film and the rotation speed have an approximately linear relationship.

Water-jet Propulsion Characteristics of Vehicle Planing

LIU Fuqiang, ZHOU Linyi, SUN Yuan, YAN Kao

2023, 31(2): 305-315. doi: 10.11993/j.issn.2096-3920.202201007

[Abstract](122) [FullText HTML] (25) [PDF 3140KB](37)

Abstract:
This study employed the shear stress transport(SST) k-ω turbulence model, multiple reference frame(MRF) model, and ideal pump model to construct a numerical simulation model of water-jet propulsion for vehicle planing. The model was implemented in STAR-CCM+ software, and its feasibility was verified. On the basis of the ideal pump model, the influent characteristics of the inner flow channel were simulated for one side with the characteristic pressure difference of the ideal pump and water absorption capacity of the axial flow pump. The flow field and hydrodynamic characteristics of vehicle planing with different pressure differences and immersion depths were studied using numerical simulation. The results show that the ideal pump model can simulate the water absorption of the axial flow pump well. A comparison of the hydrodynamic characteristics of the inner passage of the vehicle under different pressure differences reveals that with an increase in pressure, the inner passage drag increases significantly, and the inner passage lift force remains nearly unchanged. The numerical simulations of planing at different immersion depths reveal that for immersion depths of the inlet greater than 20 mm, the influent effect of the inner flow passage would no longer be impacted by immersion depth. These results provide a reference for water-jet propulsion engineering applications of vehicle planing.

Quality Control of Ocean Observation Data Based on Wave Glider

ZHOU Ying, YU Peiyuan, SUN Xiujun, SANG Hongqiang

2023, 31(2): 316-322, 328. doi: 10.11993/j.issn.2096-3920.202202003

[Abstract](255) [FullText HTML] (111) [PDF 2851KB](46)

Abstract:
The accuracy and reliability of observation data form the core of data quality control for wave gliders. An effective data quality control method is essential to promote the popularization and application of wave glider observation data. To improve the data quality of a wave glider, a new marine observation data quality control method with data inspection and data correction algorithms was developed, considering air temperature and pressure data as examples. Data inspection includes range and peak inspection, and abnormal values of the observation data are eliminated. A backpropagation(BP) neural network algorithm was adopted to correct the inspected observation data and improve the overall accuracy. In the early stage, sea trials were conducted with the “Black Pearl” wave glider-integrated AIRMAR-BP200 and GILL-GMX600 meteorological sensors, and a large number of data samples were obtained for BP neural network model training. Meanwhile, to verify the effectiveness of the proposed data quality control method, a sea trial was conducted, and the observation data obtained from the “Black Pearl” wave glider were analyzed. The experimental results show that the proposed data quality control method can effectively improve the accuracy of the observation data.

Demand and Method Discussions for Intelligent Decisions of Torpedo Attacking by Unmanned Undersea Vehicle

MA Liang, GUO Liqiang, ZHANG Hui, YANG Jing, LIU Jian

2023, 31(2): 323-328. doi: 10.11993/j.issn.2096-3920.202201003

[Abstract](150) [FullText HTML] (23) [PDF 1695KB](46)

Abstract:
The autonomous technology of unmanned equipment is currently the most dynamic frontier technology field. Therefore, improving the intelligence of decision-making is an inevitable trend in the development of future unmanned undersea vehicles (UUVs). Torpedo attack decision-making is a critical aspect of UUVs on operational missions and also the basis and prerequisite for the formation of self-organized cross-domain cooperation, autonomous swarm confrontation, and other operational capabilities. Beginning with a summary of the characteristics of their operational use and typical operational mission styles, the advantages and shortcomings of UUVs with manned platforms in operational applications were compared, and their decision-making content differed from that of traditional torpedo attacks was analyzed. The key problems that need to be solved to realize the decision-making function were expounded. Based on the development of machine learning technology, it is proposed that an intelligent decision-making method is suitable for solving the problems of large uncertainty of observation data, difficulty in guaranteeing real-time decision-making, and weak interaction ability of model perception. This study can provide a reference for future research in the field of unmanned equipment development and intelligent decision-making.

Initial Trajectory of Wire-guided Torpedo Launched byBattery-Activated Outside Tube

DING Hao, SUN Naiwei, HUANG Bo

2023, 31(2): 329-332, 348. doi: 10.11993/j.issn.2096-3920.202204002

[Abstract](76) [FullText HTML] (28) [PDF 2597KB](17)

Abstract:
To study the initial trajectory of a wire-guided torpedo launched by a battery-activated outside tube, hydrodynamic theory was used to analyze the dynamics of the torpedo when sailing outside the tube. Subsequently, a dynamic model of the torpedo movement in the vertical plane was built based on the dynamic equations of the torpedo and characteristics of a wire-guided torpedo. The initial trajectory of the longitudinal plane of the torpedo at different values of rail control rudder angle and tube speed was simulated. The horizontal distance and depth difference between the torpedo and launch boat, torpedo pitch angle, and speed were all obtained before power battery activation; thereafter, the simulation data were analyzed. The results show that the above two parameters influence the initial trajectory and navigation attitude of the wire-guided torpedo. A safe initial trajectory and good attitude control were achieved by setting the rail control rudder angle, improving the tube speed, and verifying the feasibility of the launch method.

Coupling Effect of Environmental Stress and Working Stress for Torpedo

SHE Yang, WANG Douhui, XIE Zhangyong, LU Jiale, LI Jin, CHEN Huan

2023, 31(2): 333-338. doi: 10.11993/j.issn.2096-3920.2022-0042

[Abstract](131) [FullText HTML] (11) [PDF 1088KB](25)

Abstract:
The coupling effect of environmental stress and working stress directly affects the reliability, working life, and operational effectiveness of torpedo equipment. The research on the coupling effect for torpedoes is still in its infancy, and no systematic research has been conducted on the failure mode and influence resulting from the coupling effect. Accordingly, this study considered typical failure cases by analyzing historical data. The failure modes of key parts of the torpedo under the coupling of environmental stress and working stress, including the spring, fastener, welded part, shaft part, bearing, gear, rotor blade, hydraulic system, and polymer material, were systematically summarized. The failure modes include loosening fracture, seal failure, corrosion, and mechanical deformation. We also explored the effects of these failure modes on the torpedo. The results show that the torpedo is subjected to the simultaneous coupling of environmental stress and working stress. Environmental stress, as an external stress, has a certain effect on the components of the torpedo, causing defects or weak parts of the components to deteriorate further, whereas working stress further deepens the cracks, wear, and aging of defects and weak links, ultimately leading to failure or malfunction.

Underwater Target Detection Based on Sonar Image

HAO Zixiao, WANG Qi

2023, 31(2): 339-348. doi: 10.11993/j.issn.2096-3920.202205004

[Abstract](320) [FullText HTML] (20) [PDF 1850KB](134)

Abstract:
Underwater target detection by processing sonar images is of great military and civil significance. This paper comprehensively describes the principles, methods, algorithms, and development trends in underwater target detection based on sonar images. Initially, we divide the underwater target detection task based on sonar images into traditional, deep learning-based, and combined deep learning- and transfer learning-based underwater target detection. Traditional target detection is divided into underwater target detection based on mathematical statistics, mathematical morphology, and pixels. Deep learning-based target detection methods are primarily divided into one-stage, two-stage, and detection transformer(DETR) methods. Combined deep learning- and transfer learning-based target detection is primarily divided into target detection based on simple deep neural network model transfer and complex deep learning model transfer. Finally, the advantages and disadvantages of the existing technology are summarized, and the future development direction of this field is discussed.

2023 Vol. 31, No. 2

Journal News> >

Download> >