High-simulation water surface autonomous ship virtual test simulation method

Abstract

The invention provides a high-simulation water surface autonomous ship virtual test simulation method, which comprises the following steps of: performing motion simulation on an offshore water surface autonomous ship: designing a high-fineness under-driven container ship cymbo twinbody through UE5, establishing a kinematics and dynamics model of a six-degree-of-freedom container ship by combining a separated MMG thought, and establishing a six-degree-of-freedom container ship on the basis of a PhysX engine; accurate stress simulation and collision detection are realized; radar perception is simulated by adopting a line tracking channel and a time axis, depth maps, segmentation maps and surface normal maps of monocular and binocular cameras are obtained through a Colosseum platform, and the perception ability based on a computer vision algorithm is verified; performing scene simulation: generating a sea wave dynamic effect based on a Phillips wave theory in combination with fast Fourier transform, and drawing a ship trace wave form through a mechanical wave model; dynamic sky and dynamic weather are combined to realize simulation time and complex weather change simulation; and importing a Google terrain height map through Cesium to construct a simulation scene with extremely high sense of reality.

Description

Technical Field

[0001] The present invention relates to the fields of navigation simulation, ship motion control, and autonomous navigation decision-making technology, and in particular to a highly realistic virtual test simulation method for autonomous surface ships. Background Art

[0002] With the rapid development of cutting-edge technologies such as artificial intelligence, machine learning, and embodied intelligence, the level of autonomy of Maritime Autonomous Surface Ships (MASS) is continuously improving. MASS are capable of autonomous navigation in complex marine environments and completing a variety of complex tasks, such as cargo transportation and ocean monitoring. Before MASS is officially put into commercial or operational use, it must undergo a series of scientific and systematic tests to identify potential issues and enhance its reliability and safety. The key technical aspects of these tests mainly cover intelligent perception, autonomous decision-making, and precise control.

[0003] Currently, autonomous navigation simulation testing is a key method for verifying the performance of MASS. Existing simulation testing methods often rely on heterogeneous simulators, which are capable of simulating a ship's motion, perception, and environmental scenarios to a certain extent. For example, some simulators use simple physical models to simulate the ship's motion, employ virtual sensors to simulate radar and camera perception, and incorporate pre-set scenarios to simulate different marine environments. These simulators have played a vital role in preliminary testing, providing fundamental support for the development and optimization of MASS.

[0004] However, existing simulation testing methods have many limitations. First, the modeling capabilities of these traditional simulators are limited, making it difficult to build large and complex scenarios, resulting in insufficient realism and fidelity in the simulation environment. Second, system integration is time-consuming and version compatibility is poor, which increases the difficulty of development and maintenance. In addition, resource management is decentralized, making efficient system integration and expansion difficult. These shortcomings have largely restricted the ability of simulators to simultaneously meet the requirements of full-process simulation of intelligent perception, autonomous decision-making, and precise control while ensuring high realism and high fidelity. Therefore, there is an urgent need for a testing method that has high realism and reliability and can realize full-process simulation of autonomous ships to meet the testing needs in complex and dynamic environments. Summary of the Invention

[0005] In response to the technical problems raised above, a highly realistic virtual testing simulation method for autonomous surface ships is provided. This invention effectively addresses the limitations of existing simulation testing methods and provides a more efficient, reliable, and realistic simulation environment for the development and testing of autonomous ships.

[0006] The technical means adopted in the present invention are as follows:

[0007] A highly realistic virtual testing simulation method for autonomous surface ships, comprising:

[0008] S1. Motion simulation of the maritime autonomous surface ship (MASS):

[0009] Using UE5 to design a highly detailed cyber twin of an underdriven container ship, combined with the concept of separated MMG, a six-degree-of-freedom kinematic and dynamic model of the container ship was established. Based on the PhysX engine, accurate force simulation and collision detection functions were achieved.

[0010] S2. Perception simulation of maritime autonomous surface ships (MASS):

[0011] Using line tracking channels and timelines to simulate radar perception, the Colosseum platform was used to acquire depth maps, segmentation maps, and surface normal maps from monocular and binocular cameras, verifying the perception capabilities of computer vision algorithms.

[0012] S3. Scenario simulation for Maritime Autonomous Surface Ship (MASS):

[0013] The dynamic effect of ocean waves is generated based on Phillips wave theory combined with fast Fourier transform, and the ship track wave morphology is drawn through the mechanical wave model; the dynamic sky and dynamic weather are combined to realize the simulation of simulation time and complex weather changes; the Google terrain height map is imported through Cesium to build a highly realistic simulation scene.

[0014] Furthermore, step S1 specifically includes:

[0015] S11. Design a cyber twin based on a real underdriven container ship using UE5. Since UE5 analyzes the stress conditions of the ship by calculating the stress conditions of triangular patches, the cyber twin is simplified to ensure the model's sophistication and reduce the simulator's computing power.

[0016] S12. The driving force of the container ship cyber twin comes from a single-propeller and single-rudder design. However, in order to fully consider the impact of waves on the ship, the present invention constructs a six-degree-of-freedom ship kinematic model that describes the motion state of the container ship, as follows:

[0017] Where, χ represents the ship's position vector; σ represents the ship's attitude angle vector; V represents the ship's linear velocity vector; W represents the ship's angular velocity vector; x, y, and z represent the ship's absolute position in the Earth coordinate system; θ and ψ represent the ship's heading, trim, and heel attitude; u, v, and w represent the longitudinal, transverse, and vertical velocities in the ship's coordinate system, respectively; p, q, and r correspond to the ship's roll, pitch, and yaw rates, respectively;

[0018] S13. Based on mathematical derivation, calculate the transformation matrix from the velocity in the ship coordinate system to the earth coordinate system. The calculation formula is as follows:

[0019]

[0020] S14. Based on mathematical derivation, calculate the conversion matrix from the angular velocity in the ship coordinate system to the earth coordinate system. The calculation formula is as follows:

[0021]

[0022] S15. Based on the calculated transformation matrix, the kinematic model of the ship is obtained, which is as follows:

[0023] χ t =χ t-1 +R ES1 (σ t-1 )V·Δt

[0024] σ t =σ t-1 +R ES2 (σ t-1 )W·Δt

[0025] Among them, χ t represents the ship position vector at time t; t-1 represents the ship position vector at time t-1; σ t represents the ship attitude angle vector at time t; σ t-1 represents the ship attitude angle vector at time t-1; R SE1 (σ t-1 ) represents the transformation matrix from the centerline velocity of the ship coordinate system to the earth coordinate system at time t-1; R SE2 (σ t-1 ) represents the conversion matrix from the angular velocity in the ship coordinate system to the earth coordinate system at time t-1; Δt represents the sampling time interval;

[0026] S16. Based on the basic principle of MMG separation modeling for ships, the external forces and moments acting on the hull are subdivided into the fluid force on the hull, the thrust generated by the propeller, the control force provided by the rudder, and the additional forces and moments caused by waves. Ignoring the effects of wind and ocean currents on the ship, a dynamic model is constructed as follows:

[0027]

[0028] Among them, I x , I y , I z Both represent the moment of inertia; XH 、X P 、X R 、X W They represent the external forces on the bare hull, propeller, rudder and wave in the x-axis direction respectively; H 、Y P 、Y R 、Y W Respectively represent the external forces on the bare hull, propeller, rudder, and wave in the y-axis direction; Z H , Z P , Z R , Z R They represent the external forces on the bare hull, propeller, rudder and wave in the z-axis direction respectively; K H , K P , K R , K W They represent the external moments of the bare hull, propeller, rudder and wave in the x-axis direction respectively; M H 、M P 、M R 、M W Represent the external moments of the bare hull, propeller, rudder and wave in the y-axis direction respectively; N H 、N P 、N R 、N R They represent the external moments of the bare hull, propeller, rudder and wave in the z-axis direction respectively.

[0029] Furthermore, in step S1, the hull force is updated once per frame through Update Physical in the event tick, and the ship's motion state is solved in real time, with an update period of 0.01S.

[0030] Furthermore, step S2 specifically includes:

[0031] S21, using the Line Trace by Channel function, by detecting the collision target, collision location, distance and collision depth parameters, to simulate the function of marine radar. The line tracking function is fixed to the center of the hull and rotated 360 degrees to complete a realistic simulation of the radar's operating characteristics and simulate the detection capability of the ship's surrounding environment in real time.

[0032] S22. Use the monocular and binocular cameras in the Colosseum plug-in to obtain object depth maps, segmentation maps, and surface normal maps to verify computer vision-based algorithms.

[0033] Furthermore, step S2 further includes:

[0034] By leveraging the rich API interfaces provided by Colosseum, Python and C++ language systems are introduced into the simulation environment. Through the API interfaces, different system function modules can be called, which improves the scalability of the simulator and provides flexible support for subsequent algorithm development and system function verification.

[0035] Furthermore, step S3 specifically includes:

[0036] S31 integrates Cesium for Unreal, a Google digital earth scanning plug-in, to obtain high-precision terrain height map data collected by Google satellites, realistically simulating terrain features at any location on Earth. This allows the construction of autonomous navigation simulation scenarios that highly reproduce actual terrain, providing a more realistic testing and verification environment for ship intelligent navigation systems.

[0037] S32, for the modeling of ocean wave motion, Phillips wave theory is used in combination with fast Fourier transform to simulate ocean waves;

[0038] S33. In the simulation of track waves, a cross-sectional image of a water surface object is first captured and converted into a water surface height map, which serves as an interference source for disturbing the water surface waves. Then, the interference source is input into the water propagation medium material ball. The water propagation medium material ball uses the current interference source and the previous frame of water surface data to calculate the current water surface height change. Based on the mechanical wave model, the propagation process of the water surface waves is further simulated. The calculation formula is as follows:

[0039]

[0040] in, represents the partial derivative; u(x,t) represents the displacement function of the water wave, which represents the amount by which the medium deviates from its equilibrium position at position x and time t; v represents the propagation speed of the wave in the medium;

[0041] S34. To verify the autonomous decision-making algorithm in the computer vision field, the Ultra Dynamic Sky plug-in was introduced to dynamically simulate the simulation time. At the same time, the Ultra Dynamic Weather plug-in was used to construct a variety of typical maritime navigation weather scenarios, including sunny days, rainy days, heavy fog, and snowy days. Through the realistic simulation of weather changes, a diverse and complex environment support was provided for algorithm testing.

[0042] Furthermore, in step S32, Phillips wave theory is used in combination with fast Fourier transform to simulate the ocean waves, specifically including:

[0043] S321. The fast Fourier transform expression formula of the sea surface statistical model is as follows:

[0044]

[0045] in, represents the coordinate vector on the horizontal plane; t represents time; complex number The real part represents the wave height, and the imaginary part represents the phase. By performing an inverse discrete Fourier transform, the height of each vertex in the time domain can be obtained.

[0046] S322. Wave rendering is achieved based on the Phillips spectrum model. The formula is as follows:

[0047]

[0048] Among them, A represents the Phillips constant; k represents the wave number, which is the magnitude of the wave vector; L represents the wind direction, ω s is the wind speed, g is the gravity constant; represents a two-dimensional vector, located in the horizontal plane; represents the wind direction vector;

[0049] S323, the formula in step S321 Rewrite as follows:

[0050]

[0051] Where * is a conjugate complex number; Represents the initial spectral model, which is expressed as follows:

[0052]

[0053] Among them, ε j and ε i represents two independent Gaussian random numbers, which are independent of each other, have a mean of 0 and a standard deviation of 1; P h (k) represents the spectral function;

[0054] S324. During the simulation, the wave height, real-time three-dimensional wave velocity, and wave normal of each grid vertex are calculated by the above formula, thereby solving the buoyancy of the maritime autonomous surface ship (MASS) and the interference force and torque of the waves on the maritime autonomous surface ship (MASS);

[0055] S323. Calculate the horizontal offset function. The calculation formula is as follows:

[0056]

[0057] Compared with the prior art, the present invention has the following advantages:

[0058] 1. The present invention provides a highly realistic virtual testing simulation method for autonomous surface ships. Through highly detailed cyber twins and precise kinematic and dynamic models, the simulation environment can highly restore the motion state of a real ship. Advanced radar and computer vision simulation technologies are used to ensure the high fidelity of the perception module, providing reliable data support for intelligent perception and autonomous decision-making.

[0059] 2. The present invention provides a highly realistic virtual test simulation method for autonomous surface ships, which can simultaneously meet the full-process simulation requirements of intelligent perception, autonomous decision-making and precise control, and cover all key technologies in the autonomous navigation process.

[0060] 3. The present invention provides a highly realistic virtual test simulation method for autonomous surface ships. By introducing advanced plug-ins and API interfaces, it achieves efficient system integration and flexible scalability, facilitating subsequent algorithm development and system function verification.

[0061] 4. The present invention provides a highly realistic virtual testing simulation method for autonomous surface ships. By constructing highly realistic simulation scenes and dynamic weather change simulations, it provides diversified and complex environmental support for the testing of autonomous ships in complex and dynamic environments, thereby enhancing the comprehensiveness and effectiveness of the tests.

[0062] Based on the above reasons, the present invention can be widely promoted in the fields of navigation simulation, ship motion control, autonomous navigation decision-making, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0064] Figure 1 This is a schematic diagram of the full-stack ship navigation simulation design framework based on Colosseum and UE5 in the present invention.

[0065] Figure 2 A static mesh diagram of a container ship provided by an embodiment of the present invention.

[0066] Figure 3 Schematic diagram of the conversion between the hull coordinate system and the world coordinate system in the method of the present invention.

[0067] Figure 4 This is a schematic diagram of simulating radar perception in the present invention.

[0068] Figure 5 Schematic diagram of track wave simulation of the present invention. DETAILED DESCRIPTION

[0069] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0070] It should be noted that the terms "including" and "having" and any variations thereof in the specification and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or are inherent to these processes, methods, products or apparatuses.

[0071] like Figure 1 As shown, the present invention provides a highly realistic virtual test simulation method for autonomous surface ships, comprising:

[0072] S1. Motion simulation of the maritime autonomous surface ship (MASS):

[0073] Using UE5 to design a highly detailed cyber twin of an underdriven container ship, combined with the concept of separated MMG, a six-degree-of-freedom kinematic and dynamic model of the container ship was established. Based on the PhysX engine, accurate force simulation and collision detection functions were achieved.

[0074] S2. Perception simulation of maritime autonomous surface ships (MASS):

[0075] Using line tracking channels and timelines to simulate radar perception, the Colosseum platform is used to acquire depth maps, segmentation maps, and surface normal maps from monocular and binocular cameras, verifying the perception capabilities based on computer vision algorithms. The Colosseum platform API is also used to improve platform scalability.

[0076] S3. Scenario simulation for Maritime Autonomous Surface Ship (MASS):

[0077] The dynamic effect of ocean waves is generated based on Phillips wave theory combined with fast Fourier transform, and the ship track wave morphology is drawn through the mechanical wave model; the dynamic sky and dynamic weather are combined to realize the simulation of simulation time and complex weather changes; the Google terrain height map is imported through Cesium to build a highly realistic simulation scene.

[0078] In specific implementation, as a preferred embodiment of the present invention, step S1 specifically includes:

[0079] S11. Based on the real underdriven container ship, the cyber twin is designed through UE5. Since UE5 analyzes the stress condition of the ship by calculating the stress condition of the triangular patch, the cyber twin is simplified to ensure the fineness of the model and reduce the computing power burden of the simulator. The static mesh model is shown in Figure 2 .

[0080] S12. The driving force of the container ship cyber twin comes from a single-propeller and single-rudder design. However, in order to fully consider the impact of waves on the ship, the present invention constructs a six-degree-of-freedom ship kinematic model that describes the motion state of the container ship, as follows:

[0081]

[0082] Where, χ represents the ship's position vector; σ represents the ship's attitude angle vector; V represents the ship's linear velocity vector; W represents the ship's angular velocity vector; x, y, and z represent the ship's absolute position in the Earth coordinate system; θ and ψ represent the heading, trim and heel attitude of the ship; u, v and w represent the longitudinal, transverse and vertical velocities in the hull coordinate system respectively; p, q and r correspond to the roll, pitch and bow rates of the ship respectively; the motion state of the container ship is shown in the above formula, and the conversion between the hull coordinate system and the world coordinate system is shown in Figure 3 .

[0083] S13. Based on mathematical derivation, calculate the transformation matrix from the velocity in the ship coordinate system to the earth coordinate system. The calculation formula is as follows:

[0084]

[0085] S14. Based on mathematical derivation, calculate the conversion matrix from the angular velocity in the ship coordinate system to the earth coordinate system. The calculation formula is as follows:

[0086]

[0087] S15. Based on the calculated transformation matrix, the kinematic model of the ship is obtained, which is as follows:

[0088] χ t =χ t-1 +R ES1 (σ t-1 )V·Δt

[0089] σ t =σ t-1 +R ES2 (σt-1 )W·Δt

[0090] Among them, χ t represents the ship position vector at time t; t-1 represents the ship position vector at time t-1; σ t represents the ship attitude angle vector at time t; σ t-1 represents the ship attitude angle vector at time t-1; R SE1 (σ t-1 ) represents the transformation matrix from the centerline velocity of the ship coordinate system to the earth coordinate system at time t-1; R SE2 (σ t-1 ) represents the conversion matrix from the angular velocity in the ship coordinate system to the earth coordinate system at time t-1; Δt represents the sampling time interval;

[0091] S16. Based on the basic principle of MMG separation modeling for ships, the external forces and moments acting on the hull are subdivided into the fluid force on the hull, the thrust generated by the propeller, the control force provided by the rudder, and the additional forces and moments caused by waves. Ignoring the effects of wind and ocean currents on the ship, a dynamic model is constructed as follows:

[0092]

[0093] Among them, I x , I y , I z Both represent the moment of inertia; X H 、X P 、X R 、X W They represent the external forces on the bare hull, propeller, rudder and wave in the x-axis direction respectively; H 、Y P 、Y R 、Y W Respectively represent the external forces on the bare hull, propeller, rudder, and wave in the y-axis direction; Z H , Z P , Z R , Z R They represent the external forces on the bare hull, propeller, rudder and wave in the z-axis direction respectively; K H , K P , K R , K W They represent the external moments of the bare hull, propeller, rudder and wave in the x-axis direction respectively; M H 、M P 、M R 、M W Represent the external moments of the bare hull, propeller, rudder and wave in the y-axis direction respectively; N H 、N P、N R 、N R They represent the external moments of the bare hull, propeller, rudder and wave in the z-axis direction respectively.

[0094] In specific implementation, as a preferred embodiment of the present invention, in step S1, the hull force is updated once per frame through Update Physical in the event tick, and the ship's motion state is solved in real time, with an update period of 0.01S.

[0095] In specific implementation, as a preferred embodiment of the present invention, step S2 specifically includes:

[0096] S21. Use the Line Trace by Channel function to simulate the function of ship radar by detecting the collision target, collision position, distance and collision depth parameters. Fix the line tracking function at the center of the hull and rotate it 360 degrees to complete the realistic simulation of the radar working characteristics and simulate the detection capability of the ship's surrounding environment in real time. See the schematic diagram of simulated radar perception for details. Figure 4 .

[0097] S22. Use the monocular and binocular cameras in the Colosseum plug-in to obtain object depth maps, segmentation maps, and surface normal maps to verify computer vision-based algorithms.

[0098] In specific implementation, as a preferred embodiment of the present invention, step S2 further includes:

[0099] By leveraging the rich API interfaces provided by Colosseum, Python and C++ language systems are introduced into the simulation environment. Through the API interfaces, different system function modules can be called, which improves the scalability of the simulator and provides flexible support for subsequent algorithm development and system function verification.

[0100] In specific implementation, as a preferred embodiment of the present invention, step S3 specifically includes:

[0101] S31 integrates Cesium for Unreal, a Google digital earth scanning plug-in, to obtain high-precision terrain height map data collected by Google satellites, realistically simulating terrain features at any location on Earth. This allows the construction of autonomous navigation simulation scenarios that highly reproduce actual terrain, providing a more realistic testing and verification environment for ship intelligent navigation systems.

[0102] S32. For modeling of ocean wave motion, Phillips wave theory is used in combination with fast Fourier transform to simulate ocean waves. Two main methods are usually used: statistical method and theoretical method. The statistical method describes the ocean wave motion by combining actual observation data with energy spectrum analysis of wave theory; the theoretical method studies the regular wave behavior based on fluid dynamics equations, assuming that seawater is an incompressible, irrotational fluid. In order to take into account both the real-time performance and realism of the simulation, the present invention uses Phillips wave theory in combination with fast Fourier transform to simulate ocean waves, specifically including:

[0103] S321. The fast Fourier transform expression formula of the sea surface statistical model is as follows:

[0104]

[0105] in, represents the coordinate vector on the horizontal plane; t represents time; complex number The real part represents the wave height, and the imaginary part represents the phase. By performing an inverse discrete Fourier transform, the height of each vertex in the time domain can be obtained.

[0106] S322. Wave rendering is achieved based on the Phillips spectrum model. The formula is as follows:

[0107]

[0108] Among them, A represents the Phillips constant; k represents the wave number, which is the magnitude of the wave vector; L represents the wind direction, ω s is the wind speed, g is the gravity constant; represents a two-dimensional vector, located in the horizontal plane; represents the wind direction vector;

[0109] S323, the formula in step S321 Rewrite as follows:

[0110]

[0111] Where * is a conjugate complex number; Represents the initial spectral model, which is expressed as follows:

[0112]

[0113] Among them, ε j and ε i represents two independent Gaussian random numbers, which are independent of each other, have a mean of 0 and a standard deviation of 1; P h (k) represents the spectral function;

[0114] S324. During the simulation, the wave height, real-time three-dimensional wave velocity, and wave normal of each grid vertex are calculated by the above formula, thereby solving the buoyancy of the maritime autonomous surface ship (MASS) and the interference force and torque of the waves on the maritime autonomous surface ship (MASS);

[0115] S323. Calculate the horizontal offset function. The calculation formula is as follows:

[0116]

[0117] S33. In the simulation of ship track waves, first capture the cross-sectional image of the water surface object and convert the cross-sectional image into a water surface height map as the interference source for disturbing the water surface waves. Then, input the interference source into the water propagation medium material ball. The water propagation medium material ball uses the current interference source and the previous frame of water surface data to calculate the current water surface height change. Based on the mechanical wave model, the propagation process of the water surface waves is further simulated. The schematic diagram of ship track wave simulation is shown in Figure 5 , the calculation formula is as follows:

[0118]

[0119] in, represents the partial derivative; u(x,t) represents the displacement function of the water wave, which represents the amount by which the medium deviates from its equilibrium position at position x and time t; v represents the propagation speed of the wave in the medium;

[0120] S34. To verify the autonomous decision-making algorithm in the computer vision field, the Ultra Dynamic Sky plug-in was introduced to dynamically simulate the simulation time. At the same time, the Ultra Dynamic Weather plug-in was used to construct a variety of typical maritime navigation weather scenarios, including sunny days, rainy days, heavy fog, and snowy days. Through the realistic simulation of weather changes, a diverse and complex environment support was provided for algorithm testing.

[0121] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A highly realistic virtual testing simulation method for autonomous surface ships, characterized in that: include: S1. Motion simulation of autonomous surface ships at sea: Using UE5 to design a highly detailed cyber twin of an underdriven container ship, combined with the concept of separated MMG, a six-degree-of-freedom kinematic and dynamic model of the container ship was established. Based on the PhysX engine, accurate force simulation and collision detection functions were achieved. S2. Perception simulation of autonomous surface ships at sea: Using line tracking channels and timelines to simulate radar perception, the Colosseum platform was used to acquire depth maps, segmentation maps, and surface normal maps from monocular and binocular cameras, verifying the perception capabilities of computer vision algorithms. S3. Scenario simulation for autonomous surface ships at sea: The dynamic effect of ocean waves is generated based on Phillips wave theory combined with fast Fourier transform, and the ship track wave morphology is drawn through the mechanical wave model; the dynamic sky and dynamic weather are combined to realize the simulation of simulation time and complex weather changes; the Google terrain height map is imported through Cesium to build a highly realistic simulation scene.

2. A highly realistic virtual test simulation method for autonomous surface ships according to claim 1, characterized in that: Step S1 specifically includes: S11. Design a cyber twin based on a real underdriven container ship using UE5. Since UE5 analyzes the stress conditions of the ship by calculating the stress conditions of triangular patches, the cyber twin is simplified to ensure the model's sophistication and reduce the simulator's computing power. S12. Construct a six-degree-of-freedom ship kinematic model that describes the motion state of the container ship, as follows: Where, χ represents the ship's position vector; σ represents the ship's attitude angle vector; V represents the ship's linear velocity vector; W represents the ship's angular velocity vector; x, y, and z represent the ship's absolute position in the Earth coordinate system; θ and ψ represent the ship's heading, trim, and heel attitude; u, v, and w represent the longitudinal, transverse, and vertical velocities in the ship's coordinate system, respectively; p, q, and r correspond to the ship's roll, pitch, and yaw rates, respectively; S13. Based on mathematical derivation, calculate the transformation matrix from the centerline velocity of the hull coordinate system to the earth coordinate system. The calculation formula is as follows: S14. Based on mathematical derivation, calculate the conversion matrix from the angular velocity in the ship coordinate system to the earth coordinate system. The calculation formula is as follows: S15. Based on the calculated transformation matrix, the kinematic model of the ship is obtained, which is as follows: x t =x t-1 +R SE1 (s t-1 )V·Δt s t =s t-1 +R SE2 (s t-1 )W·Δt Among them, χ t represents the ship position vector at time t; t-1 represents the ship position vector at time t-1; σ t represents the ship attitude angle vector at time t; σ t-1 represents the ship attitude angle vector at time t-1; R SE1 (σ t-1 ) represents the transformation matrix from the centerline velocity of the ship coordinate system to the earth coordinate system at time t-1; R SE2 (σ t-1 ) represents the conversion matrix from the angular velocity in the ship coordinate system to the earth coordinate system at time t-1; Δt represents the sampling time interval; S16. Based on the basic principle of MMG separation modeling for ships, the external forces and moments acting on the hull are subdivided into the fluid force on the hull, the thrust generated by the propeller, the control force provided by the rudder, and the additional forces and moments caused by waves. Ignoring the effects of wind and ocean currents on the ship, a dynamic model is constructed as follows: Among them, I x , I y , I z Both represent the moment of inertia; X H 、X P 、X R 、X W They represent the external forces on the bare hull, propeller, rudder and wave in the x-axis direction respectively; H 、Y P 、Y R 、Y W Respectively represent the external forces on the bare hull, propeller, rudder, and wave in the y-axis direction; Z H , Z P , Z R , Z R They represent the external forces on the bare hull, propeller, rudder and wave in the z-axis direction respectively; K H , K P , K R , K W They represent the external moments of the bare hull, propeller, rudder and wave in the x-axis direction respectively; M H 、M P 、M R 、M W Represent the external moments of the bare hull, propeller, rudder and wave in the y-axis direction respectively; N H 、N P 、N R 、N R They represent the external moments of the bare hull, propeller, rudder and wave in the z-axis direction respectively.

3. The highly realistic virtual testing simulation method for autonomous surface ships according to claim 1, characterized in that: In step S1, the hull forces are updated once per frame through the Update Physical event tick, solving the ship's motion state in real time with an update period of 0.01s.

4. The highly realistic virtual testing simulation method for autonomous surface ships according to claim 1, characterized in that: Step S2 specifically includes: S21, using the Line Trace by Channel function, by detecting the collision target, collision location, distance and collision depth parameters, to simulate the function of marine radar. The line tracking function is fixed to the center of the hull and rotated 360 degrees to complete a realistic simulation of the radar's operating characteristics and simulate the detection capability of the ship's surrounding environment in real time. S22. Use the monocular and binocular cameras in the Colosseum plug-in to obtain object depth maps, segmentation maps, and surface normal maps to verify computer vision-based algorithms.

5. The highly realistic virtual testing simulation method for autonomous surface ships according to claim 1, characterized in that: Step S2 further includes: By leveraging the rich API interfaces provided by Colosseum, Python and C++ language systems are introduced into the simulation environment. Through the API interfaces, different system function modules can be called, which improves the scalability of the simulator and provides flexible support for subsequent algorithm development and system function verification.

6. The highly realistic virtual testing simulation method for autonomous surface ships according to claim 1, characterized in that: Step S3 specifically includes: S31 integrates Cesium for Unreal, a Google digital earth scanning plug-in, to obtain high-precision terrain height map data collected by Google satellites, realistically simulating terrain features at any location on Earth. This allows the construction of autonomous navigation simulation scenarios that highly reproduce actual terrain, providing a more realistic testing and verification environment for ship intelligent navigation systems. S32, for the modeling of ocean wave motion, Phillips wave theory is used in combination with fast Fourier transform to simulate ocean waves; S33. In the simulation of track waves, a cross-sectional image of a water surface object is first captured and converted into a water surface height map, which serves as an interference source for disturbing the water surface waves. Then, the interference source is input into the water propagation medium material ball. The water propagation medium material ball uses the current interference source and the previous frame of water surface data to calculate the current water surface height change. Based on the mechanical wave model, the propagation process of the water surface waves is further simulated. The calculation formula is as follows: in, represents the partial derivative; u(x,t) represents the displacement function of the water wave, which represents the amount by which the medium deviates from its equilibrium position at position x and time t; v represents the propagation speed of the wave in the medium; S34. To verify the autonomous decision-making algorithm in the computer vision field, the Ultra Dynamic Sky plug-in was introduced to dynamically simulate the simulation time. At the same time, the Ultra Dynamic Weather plug-in was used to construct a variety of typical maritime navigation weather scenarios, including sunny days, rainy days, heavy fog, and snowy days. Through the realistic simulation of weather changes, a diverse and complex environment support was provided for algorithm testing.

7. A highly realistic virtual testing simulation method for autonomous surface ships according to claim 6, characterized in that: In step S32, Phillips wave theory is used in combination with fast Fourier transform to simulate ocean waves, specifically including: S321. The fast Fourier transform expression formula of the sea surface statistical model is as follows: in, represents the coordinate vector on the horizontal plane; t represents time; complex number The real part represents the wave height, and the imaginary part represents the phase. By performing an inverse discrete Fourier transform, the height of each vertex in the time domain can be obtained. S322. Wave rendering is achieved based on the Phillips spectrum model. The formula is as follows: Among them, A represents the Phillips constant; k represents the wave number, which is the magnitude of the wave vector; L represents the wind direction, ω s is the wind speed, g is the gravity constant; represents a two-dimensional vector, located in the horizontal plane; represents the wind direction vector; S323, the formula in step S321 Rewrite as follows: Where * is a conjugate complex number; represents the initial spectral model, which is expressed as follows: Among them, ε j and ε i Represents two independent Gaussian random numbers, which are independent of each other, have a mean of 0 and a standard deviation of 1; P h (k) represents the spectral function; S324. During the simulation, the wave height, real-time three-dimensional wave velocity, and wave normal of each grid vertex are calculated using the above formula, thereby solving the buoyancy of the autonomous surface ship at sea and the interference force and torque of the waves on the autonomous surface ship at sea; S323. Calculate the horizontal offset function. The calculation formula is as follows:

Images