Unmanned ship simulation test method and system, electronic equipment and readable storage medium

Abstract

The invention provides an unmanned ship simulation testing method and system, electronic equipment and a readable storage medium, and relates to the technical field of simulation testing, and the method comprises the steps: constructing an adversarial generation model based on a task scene library and an environment knowledge library; constructing a test environment generation model and a digital boat model based on the adversarial generation model; obtaining a test case model library; determining a parameter threshold value of a test case model library according to the test demand information, and inputting the parameter threshold value into a test environment generation model and a digital boat model; and performing simulation test on the test environment generation model and the digital boat model according to the test plan information, and outputting a test result. According to the technical scheme, an adversarial generation model is constructed based on a task scene library and an environment knowledge base, and a test environment generation model and a digital boat model are constructed. By carrying out simulation testing on the models, the efficiency of carrying out scene testing and evaluation on the unmanned ship and the accuracy of a testing result can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of simulation testing, and in particular, to a method, system, electronic device, and readable storage medium for unmanned ship simulation testing. Background Art

[0002] With the rapid development of technology, unmanned ships have shown great application potential in many fields, such as ocean monitoring, maritime transportation, and military reconnaissance. In related technologies, due to the large number of scenarios and complex tasks in the construction of the unmanned ship test environment, the efficiency of scenario-based testing and evaluation of unmanned ships is very low, and the accuracy of test results is not high. How to provide a method or system for unmanned ship simulation testing to improve the efficiency of scenario-based testing and evaluation of unmanned ships and the accuracy of test results is an urgent problem to be solved at present. Summary of the Invention

[0003] To solve or improve at least one of the above technical problems, an object of the present invention is to provide a method for unmanned ship simulation testing.

[0004] Another object of the present invention is to provide a system for unmanned ship simulation testing.

[0005] Another object of the present invention is to provide an electronic device.

[0006] Another object of the present invention is to provide a readable storage medium.

[0007] To achieve the above object, a first aspect of the present invention provides a method for unmanned ship simulation testing, including:

[0008] First, construct an adversarial generation model based on a task scenario library and an environment knowledge library.

[0009] Second, construct a test environment generation model based on the adversarial generation model.

[0010] Third, construct a plurality of digital prototype ship models based on the adversarial generation model, and information interaction can be carried out between the plurality of digital prototype ship models.

[0011] Fourth, obtain a test case model library based on the test environment generation model and the digital prototype ship models.

[0012] Fifth, generate test requirement information and test plan information according to the test outline information.

[0013] Sixth, determine the parameter thresholds of the test case model library according to the test requirement information, and input the parameter thresholds into the test environment generation model and the digital prototype ship models.

[0014] Step 7: Perform simulation tests on the test environment generation model and the digital prototype ship model according to the test plan information, and output the test results.

[0015] The present invention aims to provide a method for simulating and testing unmanned ships. An adversarial generation model is constructed based on a task scenario library and an environmental knowledge base, and a test environment generation model and multiple digital prototype ship models are constructed based on the adversarial generation model. By performing simulation tests on the test environment generation model and the digital prototype ship models, it is beneficial to improve the efficiency of scenario-based testing and evaluation of unmanned ships and the accuracy of test results.

[0016] In addition, the above technical solution provided by the present invention may also have the following additional technical features:

[0017] In some technical solutions, optionally, determine the parameter thresholds of the test case model library according to the test requirement information, and input the parameter thresholds into the test environment generation model and the digital prototype ship model. The steps include: determining the first parameter threshold of the test case model library according to the test requirement information, inputting the first parameter threshold into the test environment generation model, and simulating the typical ocean environment based on the test environment generation model; determining the second parameter threshold of the test case model library according to the test requirement information, inputting the second parameter threshold into the digital prototype ship model, and simulating the hull of the unmanned ship based on the digital prototype ship model.

[0018] In this technical solution, the first parameter threshold includes, but is not limited to, the wave height threshold, the wind speed threshold, and the sea condition duration threshold. These thresholds (the first parameter threshold) will provide a precise quantitative standard for the subsequent simulation of the test environment generation model to ensure that the test environment can truly reflect the challenges faced by unmanned ships in various ocean environments.

[0019] The second parameter threshold includes, but is not limited to, the hull size, the speed threshold, the hull stress threshold, and the load weight threshold. These thresholds will define the physical limits that the hull of an unmanned ship can withstand during operation, so as to accurately evaluate its structural safety in the simulation; it can also simulate various load conditions that an unmanned ship may encounter during actual operation, and observe the performance of the hull in terms of draft depth change, buoyancy distribution, and sailing stability under different load distributions.

[0020] In some technical solutions, optionally, determine the parameter thresholds of the test case model library according to the test requirement information, and input the parameter thresholds into the test environment generation model and the digital prototype model. The steps further include: determine the second parameter threshold of the test case model library according to the test requirement information, input the second parameter threshold into the digital prototype model, after simulating the hull of the unmanned ship based on the digital prototype model, determine the third parameter threshold of the test case model library according to the test requirement information, input the third parameter threshold into the digital prototype model, and simulate the inertial navigation load of the unmanned ship based on the digital prototype model to simulate the electromagnetic wave physical field from the perspective of the navigation radar.

[0021] In this technical solution, the third parameter threshold includes, but is not limited to, the navigation radar transmission power threshold and the navigation radar operating frequency band threshold. The third parameter threshold is used to study the influence of the characteristics of electromagnetic waves at different frequencies during propagation and reflection on the inertial navigation load.

[0022] After receiving the third parameter threshold, the digital prototype model will construct a simulation of the electromagnetic wave physical field from the perspective of the navigation radar.

[0023] In some technical solutions, optionally, determine the parameter thresholds of the test case model library according to the test requirement information, and input the parameter thresholds into the test environment generation model and the digital prototype model. The steps further include: determine the second parameter threshold of the test case model library according to the test requirement information, input the second parameter threshold into the digital prototype model, after simulating the hull of the unmanned ship based on the digital prototype model, determine the fourth parameter threshold of the test case model library according to the test requirement information, input the fourth parameter threshold into the digital prototype model, and simulate the navigation and positioning load of the unmanned ship based on the digital prototype model to simulate the underwater acoustic physical field from the perspective of the sonar.

[0024] In this technical solution, the fourth parameter threshold includes, but is not limited to, the sonar operating frequency range, the sonar transmission power threshold, and the acoustic reflection coefficient threshold of the target object. The sonar operating frequency range and the sonar transmission power threshold are used to simulate the operating states under different detection intensity requirements. The acoustic reflection coefficient of the target object is used to simulate the difference in the ability of targets with different materials and structures to reflect sound waves.

[0025] In some technical solutions, optionally, determine the parameter thresholds of the test case model library according to the test requirement information, and input the parameter thresholds into the test environment generation model and the digital prototype model. The steps further include: determining the second parameter threshold of the test case model library according to the test requirement information, inputting the second parameter threshold into the digital prototype model, and after simulating the hull of the unmanned ship based on the digital prototype model, determining the fifth parameter threshold of the test case model library according to the test requirement information, inputting the fifth parameter threshold into the digital prototype model, and simulating the optoelectronic tracking payload of the unmanned ship based on the digital prototype model to simulate the visible light physical field based on the optoelectronic visual angle.

[0026] In this technical solution, the fifth parameter threshold includes but is not limited to the visible light reflectivity threshold, the visibility threshold, and the resolution threshold. The visible light reflectivity threshold is used to evaluate the detection sensitivity and recognition accuracy of the optoelectronic tracking payload for targets of different materials. Visibility affects the visible distance and clarity of the optoelectronic tracking payload for targets, and different visibility thresholds can simulate various actual marine atmospheric environment scenarios. The resolution threshold is used to evaluate the accuracy of the optoelectronic tracking payload in extracting and recognizing target features at different resolutions.

[0027] In some technical solutions, optionally, determine the parameter thresholds of the test case model library according to the test requirement information, and input the parameter thresholds into the test environment generation model and the digital prototype model. The steps further include: determining the first parameter threshold of the test case model library according to the test requirement information, inputting the first parameter threshold into the test environment generation model, and after simulating the typical marine environment based on the test environment generation model, determining the sixth parameter threshold of the test case model library according to the test requirement information, inputting the sixth parameter threshold into the test environment generation model, and simulating the typical obstacles based on the test environment generation model.

[0028] In this technical solution, the sixth parameter threshold includes but is not limited to the size threshold of the typical obstacle, the position range of the typical obstacle, the hardness threshold of the typical obstacle, and the surface roughness threshold of the typical obstacle.

[0029] The size threshold of the typical obstacle can reflect the size of various typical obstacles that the unmanned ship may encounter during actual navigation, so that it can cover small buoys to large sunken ships.

[0030] The position range of the typical obstacle is used to evaluate the distribution of the typical obstacle at different positions in the waterway, including various situations such as being close to the waterway edge and being located in the center of the waterway, so as to comprehensively test the obstacle avoidance strategy and navigation accuracy of the unmanned ship.

[0031] The hardness threshold of the typical obstacle is used to evaluate the hardness of the typical obstacle.

[0032] The surface roughness threshold of typical obstacles is used to evaluate the surface roughness size of typical obstacles.

[0033] In some technical solutions, optionally, the typical obstacles include one or a combination of the following: static targets and dynamic targets; the static targets include one or a combination of the following: buoys and reefs; the dynamic targets include one or a combination of the following: fishing boats and ships.

[0034] In this technical solution, by dividing the typical obstacles into static targets and dynamic targets, it is convenient to track and lock the typical obstacles in subsequent steps, so as to comprehensively test the obstacle avoidance strategy and navigation accuracy of the unmanned ship.

[0035] By dividing the static targets into buoys and reefs, during the simulation test process, it can be determined whether the unmanned ship needs to avoid the static targets and the timing of avoidance, so as to comprehensively test the obstacle avoidance strategy and navigation accuracy of the unmanned ship.

[0036] By dividing the dynamic targets into fishing boats and ships, different avoidance strategies are implemented according to the dynamic targets with different speeds, so as to comprehensively test the obstacle avoidance strategy and navigation accuracy of the unmanned ship.

[0037] The second aspect of the present invention provides an unmanned ship simulation test system, including a first model construction module, a second model construction module, a third model construction module, a model library construction module, an information generation module, a parameter threshold determination module, and a simulation test module.

[0038] The first model construction module is used to construct an adversarial generation model based on the task scenario library and the environment knowledge library.

[0039] The second model construction module is used to construct a test environment generation model based on the adversarial generation model.

[0040] The third model construction module is used to construct a plurality of digital sample boat models based on the adversarial generation model, and information interaction can be carried out between the plurality of digital sample boat models.

[0041] The model library construction module is used to obtain a test case model library based on the test environment generation model and the digital sample boat model.

[0042] The information generation module is used to generate test requirement information and test plan information according to the test outline information.

[0043] The parameter threshold determination module is used to determine the parameter threshold of the test case model library according to the test requirement information, and input the parameter threshold into the test environment generation model and the digital sample boat model.

[0044] The simulation test module is used to perform a simulation test on the test environment generation model and the digital sample boat model according to the test plan information, and output the test results.

[0045] The present invention aims to provide an unmanned ship simulation test system, which constructs an adversarial generation model based on a task scenario library and an environment knowledge base, and constructs a test environment generation model and multiple digital ship models based on the adversarial generation model. By performing simulation tests on the test environment generation model and the digital ship models, it is beneficial to improve the efficiency of scenario-based testing and evaluation of unmanned ships and the accuracy of test results.

[0046] A third aspect of the present invention provides an electronic device, including a memory and a processor. Among them, a program or instruction that can run on the processor is stored on the memory, and when the processor executes the program or instruction, the steps of the unmanned ship simulation test method in any of the above technical solutions are implemented. Therefore, the electronic device has the beneficial effects of any of the above technical solutions, which will not be elaborated here.

[0047] A fourth aspect of the present invention provides a readable storage medium, which stores a program or instruction, and when the program or instruction is executed by a processor, the steps of the unmanned ship simulation test method in any of the above technical solutions are implemented. Therefore, the readable storage medium has the beneficial effects of any of the above technical solutions, which will not be elaborated here.

[0048] The additional aspects and advantages of the technical solutions of the present invention will become apparent in the following description section or be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 Shows a flowchart of an unmanned ship simulation test method according to an embodiment of the present invention;

[0050] Figure 2 Shows a flowchart of an unmanned ship simulation test method according to another embodiment of the present invention;

[0051] Figure 3 Shows a flowchart of an unmanned ship simulation test method according to another embodiment of the present invention;

[0052] Figure 4 Shows a flowchart of an unmanned ship simulation test method according to another embodiment of the present invention;

[0053] Figure 5 Shows a flowchart of an unmanned ship simulation test method according to another embodiment of the present invention;

[0054] Figure 6 Shows a flowchart of an unmanned ship simulation test method according to another embodiment of the present invention;

[0055] Figure 7The block diagram of an unmanned ship simulation test system according to an embodiment of the present invention is shown;

[0056] Figure 8 The block diagram of an electronic device according to an embodiment of the present invention is shown.

[0057] Among them, Figures 1 to 8 The corresponding relationship between the reference numerals and the component names in the figure is as follows:

[0058] 200: Unmanned ship simulation test system; 210: First model construction module; 220: Second model construction module; 230: Third model construction module; 240: Model library construction module; 250: Information generation module; 260: Parameter threshold determination module; 270: Simulation test module; 300: Electronic device; 310: Memory; 320: Processor. Specific embodiments

[0059] In order to be able to more clearly understand the above objects, features, and advantages of the embodiments of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, without conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other.

[0060] Many specific details are set forth in the following description in order to fully understand the present invention. However, the embodiments of the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the limitations of the specific embodiments disclosed below.

[0061] Next, refer to Figures 1 to 8 Describe an unmanned ship simulation test method, system, electronic device, and readable storage medium provided according to some embodiments of the present invention.

[0062] In an embodiment according to the present invention, as Figure 1 shown, the steps of the unmanned ship simulation test method include:

[0063] S102, constructing an adversarial generation model based on a task scenario library and an environment knowledge base.

[0064] The task scenario library is a collection storing various task-related scenario information of the unmanned ship. The task scenario library includes, but is not limited to, task type information and scenario description information. It should be noted that the scenario description information here corresponds to the task type information.

[0065] Among them, the task type information includes but is not limited to patrol task information, search and rescue task information, material transportation task information, and ocean monitoring task information. Taking the patrol task information as an example, the patrol task information includes the patrol area range, the patrol route planning, and the key attention targets during the patrol process.

[0066] The scene description information includes but is not limited to the target location or path, weather information, and sea condition information.

[0067] For example, in the search and rescue task scenario, the scene description information includes the location where the distress signal is sent (target location), the attributes of the distress target (such as a person falling into the water or a malfunctioning ship), weather information, and sea condition information.

[0068] It should be emphasized that the task scenario library provides rich task scenario materials for the unmanned ship simulation test, enabling the test to cover various actual possible situations. By calling and combining different scenarios in the task scenario library, the operating states of unmanned ships under different task requirements can be simulated, providing data support for evaluating the performance of unmanned ships in various task scenarios.

[0069] The environmental knowledge base is a repository storing knowledge related to the operating environment of unmanned ships. The environmental knowledge base includes but is not limited to marine environmental data, meteorological environmental data, and geographical environmental data.

[0070] Among them, the marine environmental data includes but is not limited to the water flow velocity, water flow direction, wave height data, wavelength data, and the distribution of seawater temperature and salinity.

[0071] The meteorological environmental data includes but is not limited to the wind direction, wind speed, air pressure (the influence of high and low pressure systems on the sea surface conditions), and visibility (the visible distance affected by weather conditions such as fog, rain, snow, etc.).

[0072] The geographical environmental data includes but is not limited to the location and shape of reefs, shoals, and harbors, as well as the distribution of underwater mountains and trenches.

[0073] It should be emphasized that the environmental knowledge base provides a reference basis for the environmental factors in the unmanned ship simulation, enabling the simulation environment to more realistically simulate the operating conditions of unmanned ships in the actual marine environment. When constructing the test environment and analyzing the performance of unmanned ships, this environmental knowledge can help consider the influence of environmental factors on various aspects of the unmanned ship's power system, navigation system, sensor system, etc.

[0074] The adversarial generation model is a neural network model structure composed of a generator and a discriminator, which generates data meeting specific requirements through the adversarial training between the generator and the discriminator. In the context of unmanned ship simulation, the adversarial generation model uses the data in the task scenario library and the environmental knowledge base to generate data related to the unmanned ship task scenario and environmental conditions.

[0075] Optionally, the adversarial generation model generates test data according to the data in the task scenario library and the environment knowledge library. A test environment generation model and a digital sample boat model are constructed based on the test data.

[0076] S104. Construct a test environment generation model based on the adversarial generation model.

[0077] The adversarial generation model generates first test data according to the task scenario library and the environment knowledge library. A test environment generation model is constructed based on the first test data.

[0078] Optionally, the adversarial generation model generates first test data according to the task type information, scenario description information, ocean environment data, meteorological environment data, and geographical environment data.

[0079] The first test data includes, but is not limited to, sea condition test data, meteorological test data, and geographical feature test data.

[0080] The sea condition test data includes, but is not limited to, the height, period, and direction of sea surface waves. The generator of the adversarial generation model generates detailed sea condition data such as the height, period, and direction of sea surface waves according to the input specific encoding or random vector. These data do not exist in isolation, but are wave form data with physical rationality generated based on the learning of a large amount of real sea condition data and following the principles of ocean dynamics.

[0081] The meteorological test data includes multi-dimensional meteorological parameters such as wind direction, wind speed, air pressure, and precipitation probability in different regions, different seasons, and different time points. These multi-dimensional meteorological parameters are interrelated, and the gradient relationship between wind speed and air pressure, the association between wind direction and sea-land distribution and seasonal changes, etc. can all be reflected in the generated data.

[0082] The geographical feature test data includes, but is not limited to, the undulation of the seabed topography, the distribution of reefs and shoals, and the shape characteristics of the coastline.

[0083] The sea condition test data, meteorological test data, and geographical feature test data are organically integrated to construct a test environment generation model.

[0084] S106. Construct multiple digital sample boat models based on the adversarial generation model, and information interaction can be carried out between the multiple digital sample boat models.

[0085] The adversarial generation model generates second test data according to the task scenario library and the environment knowledge library. A digital sample boat model is constructed based on the second test data.

[0086] The second test data includes, but is not limited to, initial route planning, sensor detection range and accuracy setting, and energy consumption mode of the power system.

[0087] It should be emphasized that the functional characteristics of the digital prototype boat model include simulating real characteristics and dynamic adaptability.

[0088] Regarding simulating real characteristics: Each digital prototype boat model aims to highly simulate various functions and characteristics of real unmanned ships. Simulate the navigation system (such as the accuracy and error characteristics of functions like GPS positioning and inertial navigation), the detection capabilities of the optoelectronic system (the recognition and tracking effects of different distances and different types of targets), and the underwater detection range and resolution of the sonar system in different ocean environments, etc. The parameters and performances of these simulation functions are all based on the learning and induction of a large amount of real data by the adversarial generative model.

[0089] Regarding dynamic adaptability: The digital prototype boat model can dynamically adjust its own state according to the simulated environmental changes.

[0090] Optionally, information interaction is achieved among multiple digital prototype boat models through a virtual network established by software-defined networking (SDN) technology.

[0091] Multiple digital prototype boat models form an unmanned ship model. Information interaction among multiple digital prototype boat models is the key to simulating the cooperative operation of an unmanned ship cluster. The types of information for interaction are rich and diverse, including position information, enabling each digital prototype boat to know the relative positions of each other in the virtual environment, so as to avoid collisions and achieve formation sailing or cooperative task execution; task-related information, such as when a digital prototype boat discovers a suspicious target, it transmits information such as the position and type of the target to other digital prototype boats, so that the entire cluster can jointly formulate countermeasures; environmental perception information, when a digital prototype boat detects special sea conditions or meteorological changes in a certain area, it shares this information with other digital prototype boats, enabling them to make preparations in advance.

[0092] Regarding the design of the prototype system: The simulation environment is mainly realized in the virtual environment, which can build test scenarios according to task requirements, simulate the main payload equipment (including the navigation system, optoelectronic system, and sonar system), and finally realize the test and evaluation of typical algorithms and task modules of the unmanned ship cluster. According to the above analysis, the simulation environment is mainly composed of software related to the test system, supporting tool software, etc., and has an interaction relationship with the command and control system of the unmanned ship cluster.

[0093] Regarding the construction of digital prototype boats (digital prototype boat models): A digital prototype boat is a logically independent computing environment created by a simulation engine using virtualization technology. It has the same computing resources and network environment as a real boat and can simulate basic loads such as navigation and optoelectronics, as well as typical mission loads such as underwater sonar. The digital prototype boat model supports real-motion boat-end public services and boat-end application services. Multiple digital prototype boats may run on the same or different host servers, and are connected by virtual network connections established by software-defined network technology. The upper layer of the digital prototype boat is the "display control" terminal of the shore-based data center, which is responsible for displaying the real-time status of the unmanned ship and issuing relevant control instructions. In some test scenarios, the shore-based data center, as a special case of a digital prototype boat, is also incorporated into the unmanned ship cluster in a virtualized form.

[0094] S108, generate model and digital prototype boat model based on the test environment to obtain a test case model library.

[0095] The test environment generation model can generate diverse unmanned ship operation environment scenarios according to different requirements and parameter settings. These scenarios cover various sea conditions (such as calm sea surface, strong winds and huge waves, etc.), meteorological conditions (clear weather, strong winds and heavy rains, thick fog, etc.), and geographical environments (open sea areas, narrow channels, areas close to islands or reefs, etc.).

[0096] The digital prototype boat model represents individual unmanned ships of different types, configurations, and mission settings. Each digital prototype boat model has its unique performance parameters, such as speed range, endurance, payload capacity, sensor accuracy, etc.

[0097] When constructing the test case model library, each environment scenario generated by the test environment generation model is combined and matched with each digital prototype boat model in an all-round and multi-level manner.

[0098] By systematically sorting, classifying, and storing a large number of test cases, a test case model library rich in content and comprehensive in coverage is finally formed, providing a complete, scientific, and highly targeted test plan system for unmanned ship simulation testing, ensuring the test quality and efficiency in the process of unmanned ship research and development and optimization.

[0099] Form typical mission test plans according to different mission types. According to requirements, conduct test modeling on typical missions to obtain a test case model library.

[0100] S110, generate test requirement information and test plan information according to the test outline information.

[0101] The test outline is the overall framework and guiding principle for the entire unmanned ship simulation test. It clarifies the purpose and scope of the test.

[0102] Based on the definition of the test purpose in the test outline, determine the test requirement information. Optionally, the test requirement information includes, but is not limited to, test item information and test scenario information.

[0103] Determine the test items according to the test item information, and determine the test scenarios according to the test scenario information.

[0104] For example, when the test item is the performance of the navigation system, the test requirement information also includes the positioning accuracy requirements of the navigation system in different environments (such as the positioning error range in open sea areas, nearshore areas, areas with electromagnetic interference, etc.) and the stability requirements for the navigation system (whether the system can continue to work properly during long-term operation or environmental mutations).

[0105] When the test scenario is a storm scenario, the test requirement information also includes parameters such as wind speed range, wave height range, and rainfall intensity.

[0106] According to the test outline, generate the test requirement information and test plan information for the task integration joint debugging simulation test based on the public service infrastructure.

[0107] Optionally, the test requirement information includes, but is not limited to, test item information and test scenario information. Optionally, the test plan information includes, but is not limited to, test arrangement information.

[0108] Extract the test case template list corresponding to the test item information from the test case template library according to the test requirement information, and specify the parameter thresholds through human-computer interaction to instantiate the test cases.

[0109] According to the test scenario information, provide input to the task test guidance and adjustment software, and drive the test environment configuration of the test load equipment simulator to be completed. According to the test arrangement information, automatically execute the test, record the test data, give the test results for the quantifiable test content, and provide input to the data analysis and evaluation software.

[0110] S112. Determine the parameter thresholds of the test case model library according to the test requirement information, and input the parameter thresholds into the test environment generation model and the digital prototype model of the boat.

[0111] Extract the test case template list corresponding to the test item information from the test case template library according to the test requirement information, and specify the parameter thresholds through human-computer interaction to instantiate the test cases.

[0112] If the test requirement is to evaluate the navigation stability of an unmanned ship in severe sea conditions, then the relevant parameter thresholds will be set around sea condition parameters (such as wave height, wind speed, water flow velocity), the attitude parameters of the unmanned ship itself (such as the maximum allowable values of roll angle, pitch angle, yaw angle), and the output parameters of the power system (such as the engine power adjustment range, the propeller speed fluctuation range), etc. These thresholds will define under what circumstances the navigation stability of the unmanned ship is considered qualified or unqualified.

[0113] The test environment generation model constructs a specific test environment based on the input parameter thresholds. When the parameter thresholds set the wave height to 5 meters, the wind speed to 30 knots, and the water flow velocity to 2 knots, the test environment generation model will use its algorithms and data resources to generate an ocean environment scene with the corresponding wave height, wind speed, and water flow velocity. This scene must not only meet the numerical requirements of the parameter thresholds but also ensure the physical rationality and dynamic variability of the environment.

[0114] The digital prototype model receives the parameter thresholds to adjust its own operating parameters and performance. Taking the power system of an unmanned ship as an example, when the parameter thresholds set the maximum power limit of the engine to 80% of the rated power, the power system in the digital prototype model will limit its power output according to this threshold, and during the simulation operation, adjust internal parameters such as fuel supply and intake air volume according to this threshold to achieve power output control.

[0115] S114, perform a simulation test on the test environment generation model and the digital prototype model according to the test plan information, and output the test results.

[0116] The test plan information includes but is not limited to the time series information of the test, the task process information, and the resource allocation information.

[0117] Determine when to start what test scenarios according to the test plan information, and the specific operation requirements for the test environment generation model and the digital prototype model under each scenario.

[0118] Determine the sequence and connection relationship of each test stage according to the test plan information. For example, after completing the basic navigation test, enter the complex sea condition test stage according to the plan, which requires the test environment generation model to adjust the sea condition parameters to complex situations such as windy waves and water flow changes, and the digital prototype model to correspondingly switch to control strategies and task modes for dealing with complex environments, such as starting the stability system, adjusting the navigation speed, and the course planning algorithm, etc.

[0119] During the test, verify the accuracy and stability of the environment generated by the test environment generation model. For example, for the generated sea condition data, check whether parameters such as wave height, wave length, and wave period conform to the expected probability distribution by comparing with the statistical characteristics of real ocean environment data.

[0120] Verify that when the environmental parameters change (such as the sea condition freezes due to a sudden drop in temperature), whether the test environment generation model can update the environmental state in a timely and accurate manner, and whether this update has a reasonable impact on the operation of the digital prototype boat model (such as the increased navigation resistance of the digital prototype boat caused by the ice surface, the interference of the detection system by the ice layer, etc.).

[0121] In the navigation task, check whether the positioning accuracy of the digital prototype boat model meets the threshold requirements set in the test plan. For example, during long-distance navigation, by comparing with the real geographical coordinates, evaluate the error range of the GPS or other navigation systems of the digital prototype boat model. Under different sea conditions and meteorological conditions, observe the navigation stability of the digital prototype boat model, such as whether the roll and pitch angles are within an acceptable range.

[0122] For the load system test of the digital prototype boat model, if the digital prototype boat model is equipped with detection equipment, check its detection range, detection accuracy, and target recognition ability during the simulation test.

[0123] Optionally, the test plan information includes but is not limited to the test arrangement information.

[0124] Automatically execute the test according to the test arrangement information, record the test data, give the test results for the quantifiable test content, and provide input for the data analysis and evaluation software.

[0125] Optionally, the output results include the performance index data of the digital prototype boat model and the relevant data of the test environment. Optionally, the performance index data of the digital prototype boat model includes but is not limited to quantifiable data such as navigation speed, navigation trajectory, energy consumption, and system failure rate. These data can intuitively reflect the operation efficiency and reliability of the digital prototype boat model in different test scenarios.

[0126] The relevant data of the test environment includes but is not limited to the statistical results of the parameter accuracy of the generated environment and the response time of the environmental change. For example, during the simulation of a storm, the time delay from when the test environment generation model receives the storm generation instruction to when it completes the construction of the storm environment, and the error statistical data of various parameters (such as wind speed, wave height, rainfall intensity, etc.) in the storm environment compared with the preset target values to evaluate the performance and quality of the test environment generation model.

[0127] Optionally, the output results also include the behavior records and analysis results of the digital prototype boat model during the test. The behavior records and analysis results of the digital prototype boat model during the test are provided in the form of a detailed log file, including the decision tree diagram display of the decision-making system, clearly presenting its decision-making logic in different scenarios; the detection data time series diagram of the sensor system, facilitating the analysis of the detection performance change law in different environments, etc.

[0128] The present invention aims to provide an unmanned ship simulation test method, which constructs an adversarial generation model based on a task scenario library and an environment knowledge base, and further constructs a test environment generation model and multiple digital prototype ship models. By performing simulation tests on the test environment generation model and the digital prototype ship models, it is beneficial to improve the efficiency of scenario-based testing and evaluation of unmanned ships and the accuracy of test results.

[0129] In the technical solution defined by the present invention, an adversarial generation model is constructed based on a task scenario library and an environment knowledge base, and a test environment generation model and digital prototype ship models are further constructed. Methods such as sample generation, network construction, model training, and model optimization of generative models (adversarial generation models) in semi-supervised and unsupervised modes are studied, and high-level semantic representations are embedded into static diverse data generation models (test environment generation models and digital prototype ship models) to generate critical samples for testing the ability boundaries of intelligent models. The dynamic foreground and background separation modeling technology based on scene three-dimensional reconstruction technology and the dynamic scene generation technology driven by real ship data are studied to complete the sensor model modeling of the visible light physical field based on the optical television view, the electromagnetic wave physical field based on the navigation radar view, and the underwater acoustic physical field based on the sonar view, and generate multi-modal simulation data for the three-dimensional scene.

[0130] It should be noted that in the context of unmanned ship simulation, the semi-supervised mode is a data processing and model training method that lies between supervised learning and unsupervised learning. It combines a small amount of labeled data (data with clear categories or target values) and a large amount of unlabeled data for learning.

[0131] The unsupervised mode means that in the absence of clear labeling information (such as class labels and target values), the model (adversarial generation model) automatically discovers potential structures, patterns, and regularities from the data. In the relevant data processing of unmanned ships, it mainly enables the model to learn the internal representations of task scenarios and environmental data without relying on predefined data categories or target outputs.

[0132] The unmanned ship simulation test method provided by the present invention can meet the software test requirements of unmanned ships at different levels and scales. By selecting appropriate mathematical models and combining with actual sea test data, a stable and reliable single unmanned ship platform model (digital prototype model) is constructed to meet the requirements of kinematics, dynamics and general mass characteristics; for the characteristics of communication, navigation radar, inertial navigation, optoelectronic and underwater acoustic mission payloads commonly carried on the unmanned ship platform, appropriate mathematical description methods are selected and combined with actual test data to construct stable and reliable mathematical models for each payload; using software-defined networking (SDN) technology, communication connections between multiple ships (multiple digital prototype models) are established, so as to support the simulation of multiple ships in the simulation test tool.

[0133] In some embodiments, optionally, as Figure 2 shown, the steps of S112 (determining the parameter thresholds of the test case model library according to the test requirement information and inputting the parameter thresholds into the test environment to generate models and digital prototype models) include:

[0134] S1122, determining the first parameter thresholds of the test case model library according to the test requirement information, inputting the first parameter thresholds into the test environment to generate models, and simulating the typical ocean environment based on the models generated by the test environment.

[0135] The first parameter thresholds include, but are not limited to, wave height thresholds, wind speed thresholds, and sea condition duration thresholds. These thresholds (the first parameter thresholds) will provide accurate quantitative criteria for the simulation of the subsequent test environment generation models, ensuring that the test environment can truly reflect the challenges faced by unmanned ships in various ocean environments.

[0136] The ocean environment is intricate and the situation changes rapidly, which has a very high impact on the performance and stability of unmanned ships. The purpose of this step is to conduct simulation tests on unmanned ships in complex environments by constructing a highly simulated virtual ocean environment. This method is beneficial to reducing the test cost.

[0137] Precise virtual ocean data (virtual ocean data including temperature field, salinity field, free surface height, and flow field) is generated by the coupled operation of the regional ocean model system and the weather forecast model.

[0138] According to the requirements of the test task for the virtual ocean environment, support inputting parameters such as the spatio-temporal range and spatio-temporal resolution of the input mode, and be able to automatically perform spatio-temporal splitting according to the test task parameters; automatically call the preprocessing module to complete the construction of the geographical environment, initial field, forcing field, and boundary condition models of the air-sea coupling model (coupled operation of the regional ocean model system and the weather forecasting model); automatically configure the input files, and automatically call the coupling model operation script to control the coupled operation of the two models; display the current operation status of the air-sea coupling model and view the generation progress of the current data; after the data generation is completed, form the required virtual ocean environment data.

[0139] The ocean environment data generation simulator based on the air-sea coupling model is also divided into multiple modules internally. The first module is the preprocessing module; the second module is the spatio-temporal splitting module for test tasks, and its function is the same as that of this module in the model configuration and operation control software; the third module is the preprocessing and operation control module, which can call the preprocessing module to generate the initial model, and at the same time can automatically control the coupled calculation operation to generate the final required data file.

[0140] In summary, the environment and target simulator simulates the ocean environment for the input parameters through a mathematical model program and outputs the simulation results in the form of virtual ocean environment data. The input parameters include typical sea state levels and various sea conditions and environmental parameters corresponding to typical mission scenarios, etc.

[0141] S1124, determine the second parameter threshold of the test case model library according to the test requirement information, input the second parameter threshold into the digital prototype model of the boat, and simulate the hull of the unmanned ship based on the digital prototype model of the boat.

[0142] The second parameter threshold includes but is not limited to hull size, speed threshold, hull force threshold, and load weight threshold. These thresholds will define the physical limits that the hull of the unmanned ship can withstand during operation, so as to accurately evaluate its structural safety in the simulation; it can also simulate various load conditions that the unmanned ship may encounter in actual operation, and observe the performance of the hull in terms of draft depth change, buoyancy distribution, and navigation stability under different load distributions.

[0143] In some embodiments, optionally, as Figure 3 shown, after S1124 (determine the second parameter threshold of the test case model library according to the test requirement information, input the second parameter threshold into the digital prototype model of the boat, and simulate the hull of the unmanned ship based on the digital prototype model of the boat), the steps of S112 (determine the parameter threshold of the test case model library according to the test requirement information, and input the parameter threshold into the test environment generation model and the digital prototype model of the boat) further include:

[0144] S1125. Determine the third parameter threshold of the test case model library according to the test requirement information, input the third parameter threshold into the digital prototype model, and simulate the inertial navigation payload of the unmanned ship based on the digital prototype model to simulate the electromagnetic wave physical field from the perspective of the navigation radar.

[0145] The third parameter threshold includes, but is not limited to, the transmitting power threshold of the navigation radar and the working frequency band threshold of the navigation radar. The third parameter threshold is used to study the influence of the characteristics of electromagnetic waves at different frequencies during propagation and reflection on the inertial navigation payload.

[0146] After receiving the third parameter threshold, the digital prototype model will construct a simulation of the electromagnetic wave physical field from the perspective of the navigation radar.

[0147] The digital prototype model determines the position, attitude, and transmission parameters of the radar according to the third parameter threshold. The digital prototype model determines the emission source position and emission direction of the electromagnetic wave according to the position, attitude, and transmission parameters of the radar, and uses the electromagnetic wave propagation theory to calculate the propagation path of the electromagnetic wave in space.

[0148] In some embodiments, optionally, as Figure 4 shown, after S1124 (determine the second parameter threshold of the test case model library according to the test requirement information, input the second parameter threshold into the digital prototype model, and simulate the hull of the unmanned ship based on the digital prototype model), the steps of S112 (determine the parameter threshold of the test case model library according to the test requirement information, and input the parameter threshold into the test environment generation model and the digital prototype model) further include:

[0149] S1126. Determine the fourth parameter threshold of the test case model library according to the test requirement information, input the fourth parameter threshold into the digital prototype model, and simulate the navigation and positioning payload of the unmanned ship based on the digital prototype model to simulate the underwater acoustic physical field from the perspective of the sonar.

[0150] The fourth parameter threshold includes, but is not limited to, the sonar working frequency range, the sonar transmitting power threshold, and the acoustic reflection coefficient threshold of the target object. The sonar working frequency range and the sonar transmitting power threshold are used to simulate the working states under different detection intensity requirements. The acoustic reflection coefficient of the target object is used to simulate the difference in the ability of targets with different materials and structures to reflect sound waves.

[0151] After receiving the fourth parameter threshold, the digital prototype model constructs a simulation of the underwater acoustic physical field from the perspective of the sonar. Determine the position and orientation of the sonar system on the hull through the digital prototype model, and calculate the emission direction and initial beam shape of the sound wave from this starting point. According to the sonar working frequency, transmitting power, and ocean environment parameters (temperature, salinity, seabed topography, etc.), use the acoustic propagation theory to simulate the propagation process of sound waves in seawater.

[0152] In some embodiments, optionally, as Figure 5 shown, after S1124 (determining the second parameter threshold of the test case model library according to the test requirement information, inputting the second parameter threshold into the digital prototype model, and simulating the hull of the unmanned ship based on the digital prototype model), the steps of S112 (determining the parameter threshold of the test case model library according to the test requirement information, and inputting the parameter threshold into the test environment generation model and the digital prototype model) further include:

[0153] S1127, determining the fifth parameter threshold of the test case model library according to the test requirement information, inputting the fifth parameter threshold into the digital prototype model, and simulating the optoelectronic tracking payload of the unmanned ship based on the digital prototype model to simulate the visible light physical field based on the optoelectronic viewing angle.

[0154] The fifth parameter threshold includes but is not limited to the visible light reflectivity threshold, the visibility threshold, and the resolution threshold. The visible light reflectivity threshold is used to evaluate the detection sensitivity and recognition accuracy of the optoelectronic tracking payload for targets of different materials. Visibility affects the visible distance and clarity of the target by the optoelectronic tracking payload, and different visibility thresholds can simulate various actual marine atmospheric environment scenarios. The resolution threshold is used to evaluate the accuracy of the optoelectronic tracking payload in extracting and recognizing target features at different resolutions.

[0155] After receiving the fifth parameter threshold, the digital prototype model constructs a simulation of the visible light physical field based on the optoelectronic viewing angle. Determine the installation position and orientation of the optoelectronic tracking payload according to the fifth parameter threshold, and based on this, determine the optical axis direction and the initial field of view range. Simulate the propagation process of visible light in the atmosphere according to parameters such as light intensity, atmospheric visibility, and target reflection characteristics. During the propagation process, consider the attenuation, scattering of light, and the light distribution after target reflection.

[0156] In some embodiments, optionally, as Figure 6 shown, after S1122 (determining the first parameter threshold of the test case model library according to the test requirement information, inputting the first parameter threshold into the test environment generation model, and simulating the typical marine environment based on the test environment generation model), the steps of S112 (determining the parameter threshold of the test case model library according to the test requirement information, and inputting the parameter threshold into the test environment generation model and the digital prototype model) further include:

[0157] S1123, determining the sixth parameter threshold of the test case model library according to the test requirement information, inputting the sixth parameter threshold into the test environment generation model, and simulating the typical obstacle based on the test environment generation model.

[0158] The sixth parameter threshold includes, but is not limited to, the size threshold of typical obstacles, the position range of typical obstacles, the hardness threshold of typical obstacles, and the surface roughness threshold of typical obstacles.

[0159] Optionally, the size threshold of typical obstacles includes one or a combination of the following: length threshold, width threshold, and height threshold.

[0160] The size threshold of typical obstacles can reflect the sizes of various typical obstacles that an unmanned ship may encounter during actual navigation, so that smaller buoys to large sunken ships can be covered.

[0161] The position range of typical obstacles is used to evaluate the distribution of typical obstacles at different positions in the waterway, including various situations such as being close to the edge of the waterway and being located in the center of the waterway, so as to comprehensively test the obstacle avoidance strategy and navigation accuracy of the unmanned ship.

[0162] The hardness threshold of typical obstacles is used to evaluate the hardness of typical obstacles. The setting of the hardness threshold helps to determine the interaction effects that may occur when the unmanned ship collides with typical obstacles, such as the degree of damage to the ship's hull and the deformation of the obstacle after the collision.

[0163] The surface roughness threshold of typical obstacles is used to evaluate the surface roughness of typical obstacles.

[0164] In some embodiments, optionally, typical obstacles include one or a combination of the following: static targets and dynamic targets.

[0165] When the type of typical obstacle is one, the typical obstacle includes only any one of static targets and dynamic targets; when the type of typical obstacle is two, the typical obstacle includes static targets and dynamic targets.

[0166] By classifying typical obstacles into static targets and dynamic targets, it is convenient to track and lock typical obstacles in subsequent steps, so as to comprehensively test the obstacle avoidance strategy and navigation accuracy of the unmanned ship.

[0167] In some embodiments, optionally, static targets include one or a combination of the following: buoys and reefs.

[0168] When the type of static target is one, the static target includes any one of buoys and reefs; when the type of static target is two, the static target includes buoys and reefs.

[0169] By classifying static targets into buoys and reefs, during the simulation test process, it can be determined whether the unmanned ship needs to avoid static targets and the timing of avoidance, so as to comprehensively test the obstacle avoidance strategy and navigation accuracy of the unmanned ship.

[0170] In some embodiments, optionally, the dynamic targets include one or a combination of the following: fishing boats and ships.

[0171] When there is one type of dynamic target, the dynamic target includes any one of fishing boats and ships; when there are two types of dynamic targets, the dynamic targets include fishing boats and ships.

[0172] By classifying dynamic targets into fishing boats and ships, different avoidance strategies are implemented for dynamic targets with different speeds, thereby comprehensively testing the obstacle avoidance strategy and navigation accuracy of unmanned ships.

[0173] In some embodiments, optionally, typical obstacles are simulated based on the test environment generation model.

[0174] Typical obstacles (typical barriers) can be set in the simulation environment, including static targets such as buoys and reefs, and dynamic targets such as fishing boats and ships. Based on the constructed map, typical obstacles are deployed and their attributes are set. The constructed typical obstacles can set their own attributes, such as threat radius, movement speed, color, etc. (depending on the type of obstacle, the attribute items it possesses are different). The constructed static targets mainly set their initial positions and can be set in batches. The constructed dynamic targets mainly set their initial positions and movement trends, and specific movement trends can be set, including sine-cosine movement, circular movement, and rectangular movement.

[0175] The scenario guidance software can generate the movement trajectories of typical obstacles during the mission. The movement trajectories are based on time steps to generate information such as positions and speeds for each time period. The movement trajectory data of typical obstacles can support the movement simulation of virtual targets during the test process.

[0176] Optionally, the meteorological and hydrological environment can be set in the simulation environment. In the test area, meteorological and hydrological environment sampling points can be set. The meteorological and hydrological environment includes water flow speed, water flow direction, wave height, wave length, etc. The guidance software calculates all tile coordinates included in the test area according to the test area and the OSM (Open Street Map) zoom level of the map, and determines the tile coordinates where the sampling points are located. The meteorological and hydrological environment data on all tile coordinates is deduced through interpolation algorithms. The meteorological and hydrological environment can support the use of hydrodynamic models and static target drift during the test process.

[0177] In the simulation environment, using virtualized payload devices and simulation virtual environments, a real mission scenario is simulated to assist unmanned ships in carrying out simulation tests for typical mission scenarios, verifying the completeness of the software process and the accuracy of various algorithms, and verifying through actual ship sea tests, so as to ensure that the unmanned ship cluster can successfully perform tasks at sea.

[0178] Develop simulation testing and analysis tools to achieve automated simulation testing by establishing a digital model of an unmanned ship (digital ship model).

[0179] The unmanned ship simulation testing tool can meet the software testing requirements of unmanned ships at different levels and scales. Select appropriate mathematical models, combine with actual sea test data, and construct a stable and reliable single-boat platform model of the unmanned ship to meet the requirements of kinematics, dynamics, and general mass characteristics; for the characteristics of communication, navigation radar, inertial navigation, optoelectronic, and underwater acoustic mission payloads commonly carried by the unmanned ship platform, select appropriate mathematical description methods, combine with actual test data, and construct stable and reliable mathematical models for each payload; use software-defined network technology to construct communication connections between multiple boats, so as to support the simulation of multiple boats in the simulation testing tool.

[0180] The unmanned ship test data analysis tool can provide the playback and analysis of the test data of a single unmanned ship and a cluster, and provide functions such as navigation accuracy analysis, payload working correctness analysis, platform working status analysis, and command correctness analysis; it can also check the correctness and completeness of each device and function module.

[0181] By using the unmanned ship simulation testing tool and the unmanned ship test data analysis tool, the automated testing of the unmanned ship software can be fully realized, which is beneficial to improving the efficiency of scenario-based testing and evaluation of the unmanned ship and the accuracy of the test results.

[0182] In some embodiments, optionally, the communication environment is simulated based on the test environment generation model and the digital prototype model.

[0183] The path that various signals pass through after being sent from the transmitter and before reaching the receiver is collectively called the channel. The influence of the channel on the transmitted signal is a key consideration in the design of various communication systems. Among them, if the transmitted signal is a radio signal, the path that the electromagnetic wave propagates through is called the wireless channel. The wireless channel may be a very simple straight-line propagation, or it may be interfered by many different factors, such as the multipath effect caused by the reflection of the signal passing through the sea level, islands, etc. The multipath effect will cause the signal to be amplified or attenuated, and the maximum and minimum can differ by 30 to 40 dB. In addition, the relative motion between the transmitter and the receiver will cause the Doppler effect on the signal. The Doppler effect will cause the characteristics of the channel to change with time, increasing the uncertainty of the signal quality. For wireless communication systems, due to the diversity and time-variability of the propagation path, the characteristics of the wireless channel play a key role in the design of the communication signal simulator.

[0184] After being sent out from the transmitting end, radio waves spread out in all directions in space, just like an expanding sphere. Based on the principle of energy conservation, no matter what the radius is, the energy distributed on the surface area of the entire sphere must be conserved. And the surface area of the sphere is proportional to the square of the distance. Therefore, the received power is inversely proportional to the square of the propagation distance. This model is used to estimate when there are no obstacles between the transmitting end and the receiving end, and at this time, the distance between the transmitting end and the receiving end is also the shortest. Basically, in this propagation mode, the signal strength received by the receiver is inversely proportional to the square of the distance.

[0185] In some embodiments, optionally, the ship system of the unmanned ship is simulated based on the digital ship model.

[0186] The ship system simulation is mainly used for digital model management, and standard interfaces are adopted to realize the exchange and drive of digital models.

[0187] Study the typical payload simulation data generation model and the data access method of the unmanned ship simulation system, and design the aggregation and storage scheme.

[0188] The data access method mainly includes three parts: data access, data aggregation, and data storage.

[0189] Regarding data access: The typical payload simulation data generation model sends the task operation data to the simulation environment; the simulation environment connects to the simulation interface, and the simulation interface obtains the simulation task environment data.

[0190] Regarding data aggregation: The simulation environment aggregates the obtained simulation task environment data.

[0191] Regarding data storage: The simulation environment stores the obtained simulation task environment data.

[0192] The steps of the typical payload simulation data generation model and the simulation system data access process include: sending the task operation data to the simulation environment through the typical payload simulation data generation model; the simulation environment connects to the simulation interface, the simulation interface obtains the simulation task environment data, and the simulation environment aggregates and stores the obtained simulation task environment data.

[0193] In some embodiments, optionally, regarding the exchange format of digital models, various models will be packaged into FMU (Functional Mock-up Unit) files through the ZIP (a data compression format) compression algorithm in directory format to complete the exchange of digital models.

[0194] Among them, the binaries folder (i.e., the binary folder) contains the binary files of the algorithm source code compiled under the linux system (an operating system). The sources folder (i.e., the source code folder) contains the algorithm source code. The modelDescription.xml (a file format) file saves the specific formats and types of parameters such as the input and output of the model.

[0195] Specifically, in the working process of the acoustic sensor payload model, it involves the navigation positioning model outputting the position of the boat (digital prototype boat), the target generation model generating the target position, and the sonar model filtering the target. The algorithms of the above three models will be mapped into a system of hybrid differential equations and written in the source code of the FMU.

[0196] Taking the sonar model as an example, the model formula is Among them, "sonar" represents the sonar. "T sonar " represents the physical quantities, parameters, or data sets related to the sonar. "Filter sonar " represents the filtering operation. "Blur sonar " represents the blurring operation. represents the output data (physical quantities, parameters, or data sets related to the sonar) after data processing (filtering and blurring). Both interference addition and coordinated filtering are differential equations with events (i.e., the geometric relationships established through sonar parameters during filtering. Here, the difference from ordinary differential equations is that the interference addition part involves random errors).

[0197] The source code part may also involve the implementation of the solver, and whether to add the solver determines the subsequent form of model driving. The above models will all write the solver. After writing the algorithm source code of the system of hybrid differential equations of the above three models, the algorithm is packaged, and the modelDescription.xml file and the binaries folder can be automatically generated.

[0198] In the example of the sonar model, the modelDescription.xml will contain all the information of the parameters and variables, that is, it records in xml (a file format) that the input of the model is the target position and the position of the boat (and these two variables are of two-dimensional coordinate type, and the same applies to the subsequent variables and parameters), the maximum operating distance of the sonar, the observation blind area, the maximum azimuth angle, the maximum elevation angle, etc. as model parameters, and the filtered target set as the output.

[0199] In some embodiments, optionally, regarding the driving method of the digital model, an implementation package fmpy (a functional model unit processing library for the Python language) in the Python language (a high-level programming language) is used, and the digital model stored in the FMU file is loaded by calling the implementation of fmpy. In the definition of the FMI interface (a standardized interface), the implementation includes at least two modes defined in FMI: co-simulation and model exchange. The former can be directly used to drive the cooperation between FMUs containing solvers, and the latter requires additional use of the integration algorithm provided by the interface to complete the model solution of FMUs without solvers.

[0200] In some embodiments, optionally, regarding the simulation of the basic ship loads: study the methods for generating typical load simulation data and the techniques for monitoring the working states of typical loads, analyze the application scenarios and working modes of typical loads such as inertial navigation systems, radars, and optoelectronic trackers, and study the methods for modeling typical loads and simulating the generation of working data according to their characteristics.

[0201] Regarding the simulation of inertial navigation loads: Inertial navigation uses the previous position, the acceleration and angular velocity measured by the inertial measurement unit to determine its current position. Since the carrier is affected by the state, power, and sea condition information at the previous moment during navigation, the model formula of the inertial navigation simulator is:

[0202] Where State M is the motion state information of the ship at the current moment, State D is the power information of the unmanned ship at the current moment, State S is the sea condition information of the environment where the ship is located at the current moment, and Kenetics() is an operation for calculating the motion state by calling the hydrodynamic model.

[0203] The steps for simulating inertial navigation loads include:

[0204] S1, input of the inertial navigation model.

[0205] The input data of the inertial navigation model consists of three parts, namely motion state information, power information, and sea condition environment information.

[0206] Specifically, the motion state information includes the position, motion speed, motion direction, acceleration, acceleration direction, and angular velocity information in three degrees of freedom at the current moment; the power information includes the power magnitude, gear position, rudder, speed, and direction information; the sea condition environment information includes the wind direction, wind speed, water depth, water flow direction, water speed, wavelength, wave height, and wave speed information. According to the motion state information, power information, and sea condition environment information, the specific motion information of the unmanned ship at the next moment is determined.

[0207] S2, Hydrodynamic model calculation.

[0208] The hydrodynamic model used by the inertial navigation model calculates the specific motion information of the unmanned ship at the next moment according to the kinematic and dynamic laws, based on the motion information, power information, and sea condition information of the unmanned ship. When in use, the hydrodynamic model should first be initialized according to the ship type, and then the resultant force should be calculated based on the input power information. Based on Newton's second law of motion, the magnitude of the acceleration of the unmanned ship is obtained, and then the motion information such as the speed, position, pitch, yaw, and roll of the unmanned ship is calculated one by one through kinematic formulas. The navigation parameters are transformed from the inertial coordinate system to the navigation coordinate system using the attitude information.

[0209] S3, Generation of inertial navigation payload data.

[0210] The inertial navigation system obtains the motion information at the next moment such as the position and attitude by specifying the input motion information, power information, and sea condition information, and then calling the hydrodynamic model for calculation. According to the data generation and processing rules, it is converted into a data type that the software can receive. Finally, in the nautical chart and messages (such as status messages) of the control platform, the data in the standard data form is displayed. The data output frequency of the inertial navigation simulator is designed to be consistent with that of the real inertial navigation payload.

[0211] In some embodiments, optionally, based on the power conversion model of the boat platform, the input throttle and rudder angle are converted into the main engine jet pump power parameters. The model formula is: power = converter(gear, rudder).

[0212] Where, power represents the main engine jet pump power, gear represents the throttle, and rudder represents the rudder angle.

[0213] Based on the power model, the input main engine jet pump power, the boat's own state, and the configured sea condition information are used to perform a force analysis on the unmanned ship, and the inertial parameters (acceleration, angular acceleration) of the unmanned ship itself are calculated and output.

[0214] According to the inertial parameters of the unmanned ship itself, under the condition of a given initial motion state, an integration operation is performed to output the state information of the unmanned ship at the next moment.

[0215] It should be noted that the input initial state parameters include position, attitude, and speed parameters, and the inertial parameters include acceleration and angular acceleration parameters. The output state parameters include position, attitude, speed, acceleration, and angular acceleration parameters.

[0216] In some embodiments, optionally, the optoelectronic tracking payload of the unmanned ship is simulated based on the digital prototype ship model, and the model formula is T^_optics = 〖Filter〗_optics(T_optics).

[0217] Wherein, T is the set of targets to be processed obtained through the data generation algorithm, and T^ is the output of the target set of the model. "T_optics" represents the input data; "T^_optics" represents the output data. "〖Filter〗_optics" represents a specific transformation or filtering mechanism acting on optical-related quantities. Filter() is to coordinate the filtering operation.

[0218] In some embodiments, optionally, the navigation control of the unmanned ship is simulated based on the digital prototype ship model.

[0219] The navigation control simulation is based on the relevant research and design work of the simulation of the working data of typical load equipment, the simulation program of typical loads and the access scheme and interface design of the simulation system, the load simulation data instances, the digital model management, and the digital model rapid configuration tool.

[0220] The methods for generating typical load simulation data and the monitoring and analysis of the working states of typical loads. Study the application scenarios and working modes of typical loads such as inertial navigation, radar, sonar, and optoelectronic trackers, and research the methods for modeling typical loads and simulating the generation of working data according to their characteristics.

[0221] In some embodiments, optionally, the unmanned ship perceives or obtains the surrounding environment information through various types of loads or sensors. The loads or sensors for environmental perception include but are not limited to attitude equipment, navigation radar, three-light sphere cameras, meteorographs, AIS (Automatic Identification System), bottom control boxes, lidar, and communication subsystems.

[0222] These loads or sensors all have their own independent services for obtaining data.

[0223] Deploy the loads or sensors; obtain the surrounding environment perception data through the service interface protocols of each load; connect the middleware, send the obtained perception data to the middleware, and the middleware publishes the received perception data.

[0224] It should be noted that the middleware is a software component located between different levels or different applications in a software system.

[0225] Regarding the navigation control simulation, for the ad-hoc adjustment of the navigation mission plan, the functions of inputting the ad-hoc plan, setting relevant parameters, and issuing and executing the ad-hoc plan are realized.

[0226] Suspend or stop the ongoing navigation mission, then temporarily adjust the mission plan, including the mission segments, routes, execution strategies, and related parameters. Finally, send the adjusted mission to the on-board payloads (payloads or sensors on unmanned ships) to execute the adjusted mission.

[0227] Navigation control supports the operator in sending tasks, executing tasks, pausing tasks, resuming tasks, and stopping tasks. It supports monitoring the task execution process, including viewing the basic information of the task, viewing the task details, and task classification.

[0228] The expected speed and heading of the ship (unmanned ship) are displayed through relevant software. The speed and heading start commands are sent through the centralized control center. The software receives the commands and sends them to execute the speed and heading control. If a target is found during the execution process, the target position is reported, and the centralized control sends the speed and heading stop command, and the software stops executing the speed and heading control.

[0229] In some embodiments, optionally, for the simulation of unmanned ship task execution, it plays a role in the three stages of mission planning, mission execution, and ad-hoc mission adjustment during the process of unmanned ship and cluster tracking.

[0230] In the mission planning stage, complete the configuration of mission-related parameters. The parameter configuration mainly includes tracking mission configuration, tracking scenario configuration, unmanned ship payload configuration, cluster quantity configuration, target status configuration, and requirement configuration. The parameter configuration requires multiple configuration modules to execute.

[0231] The unmanned ship control simulation is based on the configuration of various payloads in the mission planning stage, completes the data generation of the mission, and controls the unmanned ship to execute the mission according to the specific mission plan.

[0232] The data aggregation during the mission process mainly includes defining various algorithm data patterns and storage structures, the injection method of intelligent algorithm parameters, and the configuration of massive data stream processing algorithms. In the mission execution stage, the data aggregation modules of each payload aggregate the key data generated by intelligent algorithms, virtual or physical payload devices during mission execution to support comprehensive analysis.

[0233] In the ad-hoc mission adjustment stage, the unmanned ship and cluster platform control simulation supports functions such as setting ad-hoc mission trigger conditions, mission plan adjustment, and mission issuance.

[0234] In some embodiments, optionally, the unmanned ship target simulation function module conducts target simulation by establishing a mathematical model for static target and environmental information. For dynamic targets, through the dynamic target simulator, the dynamic target information is real-time converted into static target information as input parameters. In summary, this function module uses typical environmental data and static target data as input parameters, calculates the input parameters through the unmanned ship target simulation algorithm model, and outputs control information.

[0235] Optionally, the dynamic targets include but are not limited to surface ships and fishing boats.

[0236] Simulate the maneuvering speed, speed direction, turning radius, latitude, accuracy, acceleration, angular velocity, angular acceleration, etc. of the target according to the attributes of the targets set in the typical mission target database. During the simulation process, use the technology of the navigation grid to find the path from one walkable position to another in the virtual space scene. During the movement towards the destination, control the unmanned ship to avoid static targets (stationary obstacles) or dynamic targets (moving obstacles).

[0237] In one embodiment according to the present invention, as Figure 7 shown, the unmanned ship simulation test system 200 includes a first model construction module 210, a second model construction module 220, a third model construction module 230, a model library construction module 240, an information generation module 250, a parameter threshold determination module 260, and a simulation test module 270.

[0238] The first model construction module 210 is used to construct an adversarial generation model based on the mission scenario library and the environment knowledge library.

[0239] The mission scenario library is a collection that stores various mission-related scenario information of the unmanned ship. The mission scenario library includes but is not limited to mission type information and scenario description information. It should be noted that the scenario description information here corresponds to the mission type information.

[0240] Among them, the mission type information includes but is not limited to patrol mission information, search and rescue mission information, material transportation mission information, and ocean monitoring mission information. Taking the patrol mission information as an example, the patrol mission information includes the patrol area range, patrol route planning, and key attention targets during the patrol.

[0241] The scenario description information includes but is not limited to target positions or paths, weather information, and sea condition information.

[0242] For example, in the search and rescue mission scenario, the scenario description information includes the position where the distress signal is sent (target position), the attributes of the distress target (such as a person falling into the water or a malfunctioning ship), weather information, and sea condition information.

[0243] It should be emphasized that the mission scenario library provides rich mission scenario materials for the unmanned ship simulation test, enabling the test to cover various actual possible situations. By calling and combining different scenarios in the mission scenario library, the operating states of the unmanned ship under different mission requirements can be simulated, providing data support for evaluating the performance of the unmanned ship in various mission scenarios.

[0244] The environmental knowledge base is a repository of knowledge related to the operating environment of unmanned ships. The environmental knowledge base includes, but is not limited to, marine environmental data, meteorological environmental data, and geographical environmental data.

[0245] Among them, the marine environmental data includes, but is not limited to, water flow velocity, water flow direction, wave height data, wavelength data, and the distribution of seawater temperature and salinity.

[0246] The meteorological environmental data includes, but is not limited to, wind direction, wind speed, air pressure (the influence of high and low pressure systems on the sea surface conditions), and visibility (the visible distance affected by weather conditions such as fog, rain, snow, etc.).

[0247] The geographical environmental data includes, but is not limited to, the location and shape of reefs, shoals, and harbors, as well as the distribution of underwater mountains and the distribution of trenches.

[0248] It should be emphasized that the environmental knowledge base provides a reference basis for environmental factors in the simulation of unmanned ships, enabling the simulation environment to more realistically simulate the operation of unmanned ships in the actual marine environment. When constructing the test environment and analyzing the performance of unmanned ships, this environmental knowledge can help consider the impact of environmental factors on various aspects of the power system, navigation system, sensor system, etc. of unmanned ships.

[0249] The adversarial generation model is a neural network model structure composed of a generator and a discriminator, which generates data that meets specific requirements through adversarial training between the generator and the discriminator. In the context of unmanned ship simulation, the adversarial generation model uses the data in the task scenario library and the environmental knowledge base to generate data related to the task scenarios and environmental conditions of unmanned ships.

[0250] Optionally, the adversarial generation model generates test data based on the data in the task scenario library and the environmental knowledge base. A test environment generation model and a digital prototype model of the ship are constructed based on the test data.

[0251] The second model construction module 220 is used to construct a test environment generation model based on the adversarial generation model.

[0252] The adversarial generation model generates first test data according to the task scenario library and the environmental knowledge base. A test environment generation model is constructed based on the first test data.

[0253] Optionally, the adversarial generation model generates first test data according to the task type information, scenario description information, marine environmental data, meteorological environmental data, and geographical environmental data.

[0254] The first test data includes, but is not limited to, sea condition test data, meteorological test data, and geographical feature test data.

[0255] The sea condition test data includes, but is not limited to, the height, period, and direction of sea surface waves. The generator of the adversarial generation model generates detailed sea condition data such as the height, period, and direction of sea surface waves based on the input specific encoding or random vector. These data do not exist in isolation but are physically reasonable wave form data generated based on the learning of a large amount of real sea condition data and following the principles of ocean dynamics.

[0256] The meteorological test data includes multi-dimensional meteorological parameters such as wind direction, wind speed, air pressure, and precipitation probability at different regions, different seasons, and different time points. These multi-dimensional meteorological parameters are interrelated, and relationships such as the gradient relationship between wind speed and air pressure, and the correlation between wind direction and sea-land distribution and seasonal changes can all be reflected in the generated data.

[0257] The geographical feature test data includes, but is not limited to, the undulation of the seabed topography, the distribution of reefs and shoals, and the shape characteristics of the coastline.

[0258] Organically integrate the sea condition test data, meteorological test data, and geographical feature test data to construct a test environment generation model.

[0259] The third model construction module 230 is used to construct multiple digital prototype boat models based on the adversarial generation model, and information interaction can be carried out between the multiple digital prototype boat models.

[0260] The adversarial generation model generates second test data based on the task scenario library and the environment knowledge base. Construct digital prototype boat models based on the second test data.

[0261] The second test data includes, but is not limited to, initial route planning, sensor detection range and accuracy settings, and energy consumption patterns of the power system.

[0262] It should be emphasized that the functional characteristics of the digital prototype boat models include simulating real characteristics and dynamic adaptability.

[0263] Regarding simulating real characteristics: Each digital prototype boat model aims to highly simulate various functions and characteristics of real unmanned ships. Simulate the navigation system (such as the accuracy and error characteristics of functions like GPS positioning and inertial navigation), simulate the detection capabilities of optoelectronic systems (the recognition and tracking effects of different distances and different types of targets), and simulate the underwater detection range and resolution of sonar systems in different ocean environments. The parameters and performances of these simulation functions are all based on the learning and induction of a large amount of real data by the adversarial generation model.

[0264] Regarding dynamic adaptability: The digital prototype boat models can dynamically adjust their own states according to the simulated environmental changes.

[0265] Optionally, information interaction between multiple digital prototype boat models is achieved through a virtual network established by software-defined networking (SDN) technology.

[0266] Multiple digital prototype boat models form an unmanned ship model. Information interaction between multiple digital prototype boat models is the key to simulating the collaborative operation of an unmanned ship cluster. The types of information for interaction are rich and diverse, including position information, enabling each digital prototype boat to know the relative positions of others in the virtual environment, so as to avoid collisions and achieve formation sailing or collaborative task execution; task-related information. For example, when a digital prototype boat discovers a suspicious target, it transmits information such as the position and type of the target to other digital prototype boats, so that the entire cluster can jointly formulate countermeasures; environmental perception information. When a digital prototype boat detects special sea conditions or meteorological changes in a certain area, it shares this information with other digital prototype boats, enabling them to make preparations in advance.

[0267] Regarding the design of the prototype system: The simulation environment mainly realizes the construction of test scenarios according to task requirements in a virtual environment, simulates the main payload devices (including navigation systems, optoelectronic systems, and sonar systems), and finally realizes the test and evaluation of typical algorithms and task modules of the unmanned ship cluster. According to the above analysis, the simulation environment mainly consists of software related to the test system, supporting tool software, etc., and has an interaction relationship with the command and control system of the unmanned ship cluster.

[0268] Regarding the construction of digital prototype boats (digital prototype boat models): A digital prototype boat is a logically independent computing environment created by a simulation engine using virtualization technology. It has the same computing resources and network environment as a real boat and can simulate basic payloads such as navigation and optoelectronics, as well as typical task payloads such as underwater sonar. The digital prototype boat model supports boat-end public services and boat-end application services with realistic motion. Multiple digital prototype boats may run on the same or different host servers and are connected to each other by a virtual network established by software-defined networking technology. The upper layer of the digital prototype boat is the "display control" terminal of the shore-based data center, which is responsible for displaying the real-time status of the unmanned ship and issuing relevant control instructions. In some test scenarios, the shore-based data center, as a special case of a digital prototype boat, is also incorporated into the unmanned ship cluster in a virtualized form.

[0269] The model library construction module 240 is used to generate models and digital prototype boat models based on the test environment to obtain a test case model library.

[0270] The test environment generation model can generate diverse unmanned ship operating environment scenarios according to different requirements and parameter settings. These scenarios cover various sea conditions (such as calm sea surface, rough waves, etc.), meteorological conditions (clear weather, heavy storms, foggy weather, etc.), and geographical environments (open sea areas, narrow channels, areas close to islands or reefs, etc.).

[0271] The digital prototype boat models represent individual unmanned ships of different types, configurations, and mission settings. Each digital prototype boat model has its unique performance parameters, such as speed range, endurance, payload capacity, sensor accuracy, etc.

[0272] When constructing the test case model library, each environmental scenario generated by the test environment generation model is combined and matched with each digital prototype boat model in an all-round and multi-level manner.

[0273] By systematically organizing, classifying, and storing a large number of test cases, a test case model library rich in content and covering comprehensively is finally formed, providing a complete, scientific, and highly targeted test plan system for the unmanned ship simulation test, ensuring the test quality and efficiency in the process of unmanned ship research and development and optimization.

[0274] According to different mission types, typical mission test plans are formed. According to requirements, test modeling is carried out for typical missions to obtain the test case model library.

[0275] The information generation module 250 is used to generate test requirement information and test plan information according to the test outline information.

[0276] The test outline is the overall framework and guiding principle for the entire unmanned ship simulation test. It clarifies the purpose and scope of the test.

[0277] Based on the definition of the test purpose in the test outline, the test requirement information is determined. Optionally, the test requirement information includes, but is not limited to, test item information and test scenario information.

[0278] The test items are determined according to the test item information, and the test scenarios are determined according to the test scenario information.

[0279] For example, when the test item is the performance of the navigation system, the test requirement information also includes the positioning accuracy requirements of the navigation system in different environments (such as the positioning error range in open sea areas, nearshore areas, areas with electromagnetic interference, etc.) and the stability requirements for the navigation system (whether the system can continue to work normally during long-term operation or environmental mutations).

[0280] When the test scenario is a storm scenario, the test requirement information also includes parameters such as wind speed range, wave height range, and rainfall intensity.

[0281] According to the test outline, generate the test requirement information and test plan information for the task integration joint debugging simulation test based on the public service infrastructure.

[0282] Optionally, the test requirement information includes, but is not limited to, test item information and test scenario information. Optionally, the test plan information includes, but is not limited to, test arrangement information.

[0283] Extract a list of test case templates corresponding to the test item information from the test case template library according to the test requirement information, and specify parameter thresholds through human-computer interaction to instantiate the test cases.

[0284] According to the test scenario information, provide input for the mission experiment guidance software to drive the test environment configuration of the test load equipment simulator. According to the test arrangement information, automatically execute the test and record the test data, give test results for quantifiable test content, and provide input for the data analysis and evaluation software.

[0285] The parameter threshold determination module 260 is used to determine the parameter thresholds of the test case model library according to the test requirement information, and input the parameter thresholds into the test environment generation model and the digital prototype model of the ship.

[0286] Extract a list of test case templates corresponding to the test item information from the test case template library according to the test requirement information, and specify parameter thresholds through human-computer interaction to instantiate the test cases.

[0287] If the test requirement is to evaluate the navigation stability of an unmanned ship in severe sea conditions, then the relevant parameter thresholds will be set around sea condition parameters (such as wave height, wind speed, water flow speed), the attitude parameters of the unmanned ship itself (such as the maximum allowable values of roll angle, pitch angle, and yaw angle), and the output parameters of the power system (such as the engine power adjustment range, the propeller speed fluctuation range), etc. These thresholds will define under what circumstances the navigation stability of the unmanned ship is considered qualified or unqualified.

[0288] The test environment generation model constructs a specific test environment based on the input parameter thresholds. When the parameter thresholds set the wave height to 5 meters, the wind speed to 30 knots, and the water flow speed to 2 knots, the test environment generation model will use its algorithms and data resources to generate an ocean environment scene with the corresponding wave height, wind speed, and water flow speed. This scene must not only meet the numerical requirements of the parameter thresholds but also ensure the physical rationality and dynamic variability of the environment.

[0289] The digital prototype model of the ship receives the parameter thresholds to adjust its own operating parameters and performance. Taking the power system of an unmanned ship as an example, when the parameter thresholds set the maximum power limit of the engine to 80% of the rated power, the power system in the digital prototype model of the ship will limit its power output according to this threshold, and during the simulation operation, adjust internal parameters such as fuel supply and intake air volume according to this threshold to achieve power output control.

[0290] The simulation test module 270 is used to perform simulation tests on the generated model of the test environment and the digital prototype boat model according to the test plan information, and output the test results.

[0291] The test plan information includes, but is not limited to, the time series information of the test, the task flow information, and the resource allocation information.

[0292] Determine when to start what test scenarios according to the test plan information, and the specific operation requirements for the generated model of the test environment and the digital prototype boat model under each scenario.

[0293] Determine the sequence and connection relationship of each test stage according to the test plan information. For example, after completing the basic navigation test, enter the complex sea condition test stage according to the plan. This requires the generated model of the test environment to adjust the sea condition parameters to complex situations such as windy waves and water flow changes, and the digital prototype boat model to switch to the control strategy and task mode for dealing with complex environments accordingly, such as starting the stabilization system, adjusting the navigation speed, and the course planning algorithm.

[0294] During the test, verify the accuracy and stability of the environment generated by the generated model of the test environment. For example, for the generated sea condition data, check whether the parameters such as wave height, wavelength, and wave period conform to the expected probability distribution by comparing with the statistical characteristics of the real ocean environment data.

[0295] Verify whether the generated model of the test environment can update the environmental state in a timely and accurate manner when the environmental parameters change (such as the sea condition freezes due to a sudden drop in temperature), and whether this update has a reasonable impact on the operation of the digital prototype boat model (such as the increased navigation resistance of the digital prototype boat caused by the ice surface, and the interference of the detection system by the ice layer).

[0296] In the navigation task, check whether the positioning accuracy of the digital prototype boat model meets the threshold requirements set by the test plan. For example, during a long-distance navigation, evaluate the error range of the GPS or other navigation systems of the digital prototype boat model by comparing with the real geographical coordinates. Observe the navigation stability of the digital prototype boat model under different sea conditions and meteorological conditions, such as whether the roll and pitch angles are within the acceptable range.

[0297] For the test of the payload system of the digital prototype boat model, if the digital prototype boat model is equipped with detection equipment, check its detection range, detection accuracy, and target recognition ability during the simulation test.

[0298] Optionally, the test plan information includes, but is not limited to, the test arrangement information.

[0299] Automatically execute the test according to the test arrangement information, record the test data, give the test results for the quantifiable test content, and provide input for the data analysis and evaluation software.

[0300] Optionally, the output results include the performance index data of the digital prototype boat model and the relevant data of the test environment. Optionally, the performance index data of the digital prototype boat model includes, but is not limited to, quantified data such as sailing speed, sailing trajectory, energy consumption, and system failure rate. These data can intuitively reflect the operation efficiency and reliability of the digital prototype boat model in different test scenarios.

[0301] The relevant data of the test environment includes, but is not limited to, the statistical results of the parameter accuracy of the generated environment and the response time of environmental changes. For example, during the simulation of a storm, the time delay from when the test environment generation model receives the storm generation instruction to when it completes the construction of the storm environment, and the error statistical data of various parameters (such as wind speed, wave height, rainfall intensity, etc.) in the storm environment compared with the preset target values are used to evaluate the performance and quality of the test environment generation model.

[0302] Optionally, the output results also include the behavior records and analysis results of the digital prototype boat model during the test. The behavior records and analysis results of the digital prototype boat model during the test are provided in the form of detailed log files, including the decision tree diagram display of the decision-making system, clearly presenting its decision-making logic in different scenarios; the time series diagram of the detection data of the sensor system, facilitating the analysis of the variation law of its detection performance in different environments, etc.

[0303] The present invention aims to provide an unmanned ship simulation test system 200, which constructs an adversarial generation model based on a task scenario library and an environment knowledge base, and constructs a test environment generation model and multiple digital prototype boat models based on the adversarial generation model. By performing simulation tests on the test environment generation model and the digital prototype boat models, it is beneficial to improve the efficiency of scenario-based testing and evaluation of unmanned ships and the accuracy of test results.

[0304] The present invention aims to provide an unmanned ship simulation test system 200, which constructs an adversarial generation model based on a task scenario library and an environment knowledge base, and constructs a test environment generation model and multiple digital prototype boat models based on the adversarial generation model. By performing simulation tests on the test environment generation model and the digital prototype boat models, it is beneficial to improve the efficiency of scenario-based testing and evaluation of unmanned ships and the accuracy of test results.

[0305] In the technical solution defined by the present invention, an adversarial generation model is constructed based on a task scenario library and an environment knowledge base, and a test environment generation model and a digital prototype ship model are further constructed. Methods such as sample generation, network construction, model training, and model optimization of generative models (adversarial generation models) in semi-supervised and unsupervised modes are studied, and high-level semantic representations are embedded into static diversified data generation models (test environment generation models and digital prototype ship models) to generate critical samples for testing the ability boundaries of intelligent models. The dynamic foreground and background separation modeling technology based on scene three-dimensional reconstruction technology and the dynamic scene generation technology based on real ship data-driven are studied, and the sensor models of visible light physical fields based on optoelectronic viewpoints, electromagnetic wave physical fields based on navigation radar viewpoints, and underwater acoustic physical fields based on sonar viewpoints are modeled to generate multi-modal simulation data of three-dimensional scenes.

[0306] It should be noted that in the context of unmanned ship simulation, the semi-supervised mode is a data processing and model training method that lies between supervised learning and unsupervised learning. It combines a small amount of labeled data (data with clear categories or target values) and a large amount of unlabeled data for learning.

[0307] The unsupervised mode means that in the absence of clear labeled information (such as class labels and target values), the model (adversarial generation model) automatically discovers potential structures, patterns, and rules from the data. In the relevant data processing of unmanned ships, it mainly enables the model to learn the internal representations of task scenarios and environmental data without relying on predefined data categories or target outputs.

[0308] The unmanned ship simulation test system 200 provided by the present invention can meet the software test requirements of unmanned ships at different levels and scales. Select appropriate mathematical models, combine with actual sea test data, and construct a stable and reliable single-ship platform model (digital prototype ship model) of unmanned ships to meet the requirements of kinematics, dynamics, and general mass characteristics; for the characteristics of communication, navigation radar, inertial navigation, optoelectronic, and underwater acoustic mission payloads commonly carried by unmanned ship platforms, select appropriate mathematical description methods, combine with actual test data, and construct stable and reliable mathematical models for each payload; use software-defined network (SDN, Software-Defined Networking) technology to construct communication connections between multiple ships (multiple digital prototype ship models), so as to support the simulation of multiple ships in the simulation test tool.

[0309] In an embodiment according to the present invention, as Figure 8As shown, the electronic device 300 includes a memory 310 and a processor 320. Among them, programs or instructions that can run on the processor 320 are stored on the memory 310. When the processor 320 executes the programs or instructions, the steps of the unmanned ship simulation test method in any of the above embodiments are implemented. Therefore, the electronic device 300 has the beneficial effects of any of the above embodiments, which will not be elaborated here.

[0310] In an embodiment of the present invention, a readable storage medium stores programs or instructions. When the programs or instructions are executed by a processor, the steps of the unmanned ship simulation test method in any of the above embodiments are implemented. Therefore, the readable storage medium has the beneficial effects of any of the above embodiments, which will not be elaborated here.

[0311] In the present invention, the terms "first", "second", and "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance; the term "plural" refers to two or more, unless otherwise clearly defined. Terms such as "installed", "connected", "connected to", and "fixed" should all be understood in a broad sense. For example, "connected" can be a fixed connection, a detachable connection, or an integral connection; "connected" can be a direct connection or an indirect connection through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0312] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by terms such as "up", "down", "left", "right", "front", "rear", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or unit referred to must have a specific direction, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention.

[0313] In the description of this specification, the description of terms such as "one embodiment", "some embodiments", "specific embodiments", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0314] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for testing an unmanned ship simulation, characterized in that: include: Construct an adversarial generation model based on the task scenario library and environment knowledge library; Constructing a test environment generation model based on the adversarial generation model; Constructing multiple digital boat models based on the adversarial generation model, wherein the multiple digital boat models can exchange information with each other; Based on the test environment generation model and the digital prototype boat model, a test case model library is obtained; Generate test requirement information and test plan information based on test outline information; Determine the parameter threshold of the test case model library according to the test requirement information, and input the parameter threshold into the test environment generation model and the digital prototype boat model; The test environment generation model and the digital prototype boat model are simulated and tested according to the test plan information, and the test results are output.

2. The unmanned ship simulation test method according to claim 1, characterized in that: The step of determining the parameter threshold of the test case model library according to the test requirement information and inputting the parameter threshold into the test environment generation model and the digital prototype boat model comprises: Determine a first parameter threshold of the test case model library according to the test requirement information, input the first parameter threshold into the test environment generation model, and simulate a typical marine environment based on the test environment generation model; A second parameter threshold of the test case model library is determined according to the test requirement information, the second parameter threshold is input into the digital sample boat model, and the hull of the unmanned ship is simulated based on the digital sample boat model.

3. The unmanned ship simulation test method according to claim 2, characterized in that: The step of determining the parameter threshold of the test case model library according to the test requirement information, and inputting the parameter threshold into the test environment generation model and the digital prototype boat model, further comprises: After determining the second parameter threshold of the test case model library according to the test requirement information, inputting the second parameter threshold into the digital sample boat model, and simulating the hull of the unmanned ship based on the digital sample boat model, determining the third parameter threshold of the test case model library according to the test requirement information, inputting the third parameter threshold into the digital sample boat model, and simulating the inertial navigation load of the unmanned ship based on the digital sample boat model to simulate the electromagnetic wave physical field based on the navigation radar perspective.

4. The unmanned ship simulation test method according to claim 2, characterized in that: The step of determining the parameter threshold of the test case model library according to the test requirement information, and inputting the parameter threshold into the test environment generation model and the digital prototype boat model, further comprises: After determining the second parameter threshold of the test case model library according to the test requirement information, inputting the second parameter threshold into the digital sample boat model, and simulating the hull of the unmanned ship based on the digital sample boat model, determining the fourth parameter threshold of the test case model library according to the test requirement information, inputting the fourth parameter threshold into the digital sample boat model, and simulating the navigation positioning load of the unmanned ship based on the digital sample boat model to simulate the hydroacoustic physical field based on the sonar perspective.

5. The unmanned ship simulation test method according to claim 2, characterized in that: The step of determining the parameter threshold of the test case model library according to the test requirement information, and inputting the parameter threshold into the test environment generation model and the digital prototype boat model, further comprises: After determining the second parameter threshold of the test case model library according to the test requirement information, inputting the second parameter threshold into the digital sample boat model, and simulating the hull of the unmanned ship based on the digital sample boat model, determining the fifth parameter threshold of the test case model library according to the test requirement information, inputting the fifth parameter threshold into the digital sample boat model, and simulating the optoelectronic tracking load of the unmanned ship based on the digital sample boat model to simulate the visible light physical field based on the optoelectronic perspective.

6. The unmanned ship simulation test method according to claim 2, characterized in that: The step of determining the parameter threshold of the test case model library according to the test requirement information, and inputting the parameter threshold into the test environment generation model and the digital prototype boat model, further comprises: After determining a first parameter threshold of the test case model library according to the test requirement information, inputting the first parameter threshold into the test environment generation model, and simulating a typical marine environment based on the test environment generation model, determining a sixth parameter threshold of the test case model library according to the test requirement information, inputting the sixth parameter threshold into the test environment generation model, and simulating a typical obstacle based on the test environment generation model.

7. The unmanned ship simulation test method according to claim 6, characterized in that: The typical obstacles include one or a combination of the following: static targets and dynamic targets; The static targets include one or a combination of the following: buoys and islands and reefs; The dynamic targets include one or a combination of the following: fishing boats and naval vessels.

8. An unmanned ship simulation test system, characterized in that: include: A first model building module (210) is used to build an adversarial generation model based on a task scenario library and an environment knowledge library; A second model building module (220), configured to build a test environment generation model based on the adversarial generation model; A third model building module (230) is used to build a plurality of digital sample boat models based on the adversarial generation model, wherein the plurality of digital sample boat models can exchange information with each other; A model library construction module (240) is used to generate a model based on the test environment and the digital prototype boat model to obtain a test case model library; An information generation module (250), used to generate test requirement information and test plan information according to the test outline information; A parameter threshold determination module (260), used to determine the parameter threshold of the test case model library according to the test requirement information, and input the parameter threshold into the test environment generation model and the digital sample boat model; A simulation test module (270) is used to perform simulation tests on the test environment generation model and the digital prototype boat model according to the test plan information, and output test results.

9. An electronic device, characterized in that: include: A memory (310) and a processor (320), wherein the memory (310) stores a program or instruction that can be run on the processor (320), and when the processor (320) executes the program or the instruction, the steps of the unmanned ship simulation test method according to any one of claims 1 to 7 are implemented.

10. A readable storage medium, characterized in that: The readable storage medium stores a program or an instruction, and when the program or the instruction is executed by a processor, the steps of the unmanned ship simulation test method according to any one of claims 1 to 7 are implemented.

Images