Testing method and device for ship navigation decision system

Abstract

The invention relates to a testing method and device of a ship navigation decision system, and belongs to the technical field of navigation decision testing, and the testing method of the ship navigation decision system constructs a virtual-real fusion testing scene based on historical ship navigation data from a departure port piloting station to a destination port piloting station and physical parameters and motion parameters of a to-be-tested ship; a ship navigation decision-making system of the ship to be tested is tested in the virtual-real fusion test scene, key test indexes are extracted, and the ship navigation decision-making system is used for controlling the ship to be tested to navigate from the departure port piloting station to the destination port piloting station according to historical ship navigation data; perception, control and decision-making of the ship navigation decision-making system are evaluated according to key test indexes, and the test integrity and authenticity of the ship navigation decision-making system are improved.

Description

Technical Field

[0001] This invention relates to the field of navigation decision testing technology, and in particular to a testing method and apparatus for a ship navigation decision system. Background Technology

[0002] With the development of intelligent shipping technology, autonomous navigation has become an important research direction for intelligent ships. As a key functional module, the navigation decision system is directly related to the safety and reliability of ships in complex environments. However, there are still many constraints in the testing and verification process of this system.

[0003] Existing testing methods mostly focus on verification in fixed waters or ideal environments, making it difficult to cover the complete voyage process of a ship departing from a pilot station, traversing different types of waters, and finally arriving at the destination pilot station. In particular, during the transition phases such as a ship accelerating from full speed in port to sea speed or decelerating from high-speed sea travel to enter port, these typical operating conditions place higher demands on the timeliness, accuracy, and continuous control capabilities of the decision-making system. However, these conditions are often ignored or simplified in traditional testing processes, affecting the comprehensiveness and representativeness of the tests. On the other hand, although virtual simulation testing has the advantages of low cost and high efficiency, there are still significant gaps in the dynamic realism of scenario construction, the accuracy of ship physical property modeling, and consistency with actual data, resulting in limited credibility of test results. Summary of the Invention

[0004] In view of this, it is necessary to provide a testing method and apparatus for a ship navigation decision system to solve the technical problems of insufficient completeness and authenticity in the testing of navigation decision systems.

[0005] To address the aforementioned problems, in a first aspect, the present invention provides a testing method for a ship navigation decision-making system, comprising: A virtual-real fusion test scenario is constructed based on historical ship navigation data from the origin port pilot station to the destination port pilot station, as well as the physical and motion parameters of the ship to be tested. The ship navigation decision system of the vessel under test is tested in the virtual-real fusion test scenario, and key test indicators are extracted. The ship navigation decision system is used to control the vessel under test to navigate from the origin port pilot station to the destination port pilot station based on the historical ship navigation data. The perception, control, and decision-making of the ship navigation decision-making system are evaluated based on the key test indicators.

[0006] In one possible implementation, the construction of a virtual-real fusion test scenario based on historical vessel navigation data from the originating port pilot station to the destination port pilot station, and the physical and motion parameters of the vessel under test includes: A virtual digital ship model is constructed based on the physical and kinematic parameters of the ship under test; Based on historical vessel navigation data from the originating port pilot station to the destination port pilot station, key elements of multiple scenarios from the originating port pilot station to the destination port pilot station are determined. These multiple scenarios include complex waters, transitional waters, and open waters. The key elements include vessel-specific elements, target vessel-specific elements, and environmental elements. A virtual scene is constructed based on the virtual digital ship model and key elements of multiple scenarios. The historical ship navigation data is then input into the virtual scene to construct a virtual-real fusion test scenario.

[0007] In one possible implementation, the navigation decision-making system of the vessel under test is tested in the virtual-real fusion test scenario, and key test indicators are extracted, including: The environmental element identification test involves using the navigation decision system to identify key elements in multiple scenarios from the originating port pilot station to the destination port pilot station. The identified key elements are then compared with historical data to obtain the first key test index, which includes target detection accuracy, environmental update frequency, environmental update delay, and chart fusion accuracy.

[0008] In one possible implementation, the step of testing the navigation decision-making system of the vessel under test in the virtual-real fusion test scenario and extracting key test indicators further includes: Motion control testing involves setting a planned route from the originating port pilot station to the destination port pilot station, determining the motion control command for the vessel under test to navigate along the planned route, and obtaining a second key test index corresponding to the motion control command based on the motion control command. The second key test index includes heading error, speed stability, and control delay.

[0009] In one possible implementation, the step of setting a planned route from the originating port pilot station to the destination port pilot station and determining the motion control commands for the vessel under test to navigate along the planned route includes: The ship's course, speed, and trajectory are determined based on the planned route. Based on the ship's heading, speed, and trajectory, the test ship is controlled to navigate along the planned route and motion control commands are obtained, including ship heading control commands, ship speed control commands, and trajectory tracking commands.

[0010] In one possible implementation, the step of testing the navigation decision-making system of the vessel under test in the virtual-real fusion test scenario and extracting key test indicators further includes: The rule and good seamanship compliance test involves constructing a typical encounter virtual scenario in the virtual-real fusion test scenario, implementing collision avoidance decisions on the vessel under test in the typical encounter virtual scenario, and obtaining a third key test indicator corresponding to the collision avoidance decision. The third key test indicator includes rule compliance rate and seamanship rationality matching degree.

[0011] In one possible implementation, the step of testing the navigation decision-making system of the vessel under test in the virtual-real fusion test scenario and extracting key test indicators further includes: The automatic collision avoidance and return-to-course capability coupling test involves controlling the vessel under test to navigate from the origin port pilot station to the destination port pilot station in the virtual-real fusion test scenario, and conducting a coupling test on the automatic collision avoidance and return-to-course capability of the vessel under test to obtain a fourth key test index, which includes collision avoidance success rate, collision avoidance response time, and return-to-course trajectory deviation.

[0012] In one possible implementation, the step of testing the navigation decision system of the vessel under test in the virtual-real fusion test scenario and obtaining test results further includes: The navigation adaptability test involves setting up right-hand crossing and encounter scenarios in the virtual-real fusion test scenario. The ship under test is then subjected to navigation adaptability tests in the right-hand crossing and encounter scenarios to obtain the fifth key test index, which includes course adjustment sensitivity, autonomous recovery rate, and decision robustness.

[0013] In one possible implementation, the evaluation of the ship navigation decision-making system's perception, control, and decision-making based on the key test indicators includes: The perception of the ship navigation decision system is evaluated based on the first key test indicator. The control of the ship navigation decision-making system is evaluated based on the second key test indicator; The decision-making effectiveness and safety of the ship navigation decision-making system are evaluated based on the third, fourth, and fifth key test indicators.

[0014] Secondly, the present invention also provides a testing apparatus for a ship navigation decision-making system, comprising: The virtual-real test scenario construction module is used to construct virtual-real fusion test scenarios based on historical ship navigation data from the origin port pilot station to the destination port pilot station, as well as the physical and motion parameters of the ship under test. The testing module is used to test the ship navigation decision system of the vessel under test in the virtual-real fusion test scenario and extract key test indicators. The ship navigation decision system is used to control the vessel under test to navigate from the origin port pilot station to the destination port pilot station based on the historical ship navigation data. The evaluation module is used to evaluate the perception, control, and decision-making of the ship navigation decision-making system based on the key test indicators.

[0015] The beneficial effects of this invention are as follows: A virtual-real fusion test scenario is constructed based on historical ship navigation data from the originating port pilot station to the destination port pilot station, along with the physical and motion parameters of the ship under test. By fusing multi-source information, a highly realistic station-to-station virtual-real fusion test environment is built, covering the complete voyage task chain from the pilot station, traversing complex terrain, transitioning to open water, and finally reaching the target pilot station. The ship navigation decision-making system of the ship under test is tested within this virtual-real fusion test scenario, and key test indicators are extracted. The ship navigation decision-making system is used to control the ship under test from the originating port pilot station to the destination port pilot station based on historical ship navigation data. The perception, control, and decision-making of the ship navigation decision-making system are evaluated based on the key test indicators. In the station-to-station virtual-real fusion test scenario, the dynamic coupling relationship between speed changes and maneuvering response can be accurately presented, realistically reproducing the decision-making mechanism under multiple operating conditions, thus improving the test completeness and realism of the navigation decision-making system. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating an embodiment of the testing method for the ship navigation decision system provided by the present invention; Figure 2 A visual schematic diagram of the test scenario for the test method of the ship navigation decision system provided by the present invention; Figure 3 A schematic diagram of an elliptical ship domain model for the testing method of the ship navigation decision system provided by the present invention; Figure 4 A schematic diagram of the avoidance rules for the testing method of the ship navigation decision system provided by the present invention; Figure 5 This is a schematic diagram of an embodiment of the testing device for the ship navigation decision system provided by the present invention. Detailed Implementation

[0018] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0019] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0020] One specific embodiment of the present invention discloses a testing method for a ship navigation decision-making system, which can be executed by a computer, specifically by one or more processors of the computer. For example... Figure 1 As shown, the testing methods for ship navigation decision-making systems include: S101. Construct a virtual-real fusion test scenario based on historical ship navigation data from the origin port pilot station to the destination port pilot station, as well as the physical and motion parameters of the ship to be tested. It should be noted that by constructing typical station-to-station virtual-real fusion test scenarios covering complex waters, transitional waters, and open waters using historical vessel navigation data from the originating port pilot station to the destination port pilot station, the adaptability and reliability of the navigation decision-making system under various water conditions can be verified.

[0021] S102. Test the ship navigation decision system of the ship under test in a virtual-real fusion test scenario and extract key test indicators. The ship navigation decision system is used to control the ship under test to navigate from the origin port pilot station to the destination port pilot station based on historical ship navigation data. It should be noted that the testing of the ship navigation decision system includes environmental element identification testing, motion control testing, rule and good seamanship compliance testing, automatic collision avoidance and rerouting capability coupling testing, and navigation adaptability testing. The ship navigation decision system is tested in a station-to-station virtual-real fusion test scenario, which can accurately present the dynamic coupling relationship between speed changes and maneuvering response, realistically reproduce the decision-making mechanism under multiple working conditions, and significantly improve the fitting ability to real navigation environment and complex maneuvering behavior during the testing process.

[0022] S103. Evaluate the perception, control, and decision-making of the ship's navigation decision-making system based on key test indicators; It should be noted that the ship's navigation decision-making system is evaluated through key test indicators to ensure that the system has the ability to perform navigation safely and efficiently.

[0023] In some embodiments, in step S101, a virtual-real fusion test scenario is constructed based on historical ship navigation data from the originating port pilot station to the destination port pilot station, the physical parameters and motion parameters of the ship under test, and a virtual digital ship model is constructed based on the physical parameters and motion parameters of the ship under test. Key elements of multiple scenarios from the originating port pilot station to the destination port pilot station are determined based on historical ship navigation data from the originating port pilot station to the destination port pilot station. These multiple scenarios include complex waters, transitional waters, and open waters. Key elements include ship-specific elements, target ship elements, and environmental elements. A virtual scenario is constructed based on the virtual digital ship model and the key elements of multiple scenarios, integrating historical data... Ship navigation data is input into a virtual scene to construct a virtual-real fusion test scenario. Using digital twin technology, a virtual digital ship model is constructed based on the physical and motion parameters of the real ship, combining the ship's physical characteristics and maneuvering motion features. This model is then simulated in a simulation platform to achieve the mapping from the real ship to the virtual ship. Real historical data is extracted from historical ship navigation scenarios to determine the key elements that each scenario should include. With these elements included as a premise, a virtual scenario based on the real scenario is constructed in the test simulation platform using digital twin technology. Historical data is input into the virtual scenario to reenact the scenario. The fusion of the virtual scenario and the real data forms a virtual-real fusion test scenario. The constructed virtual-real fusion test scenario mainly relies on the organic integration of two types of elements: "real" and "virtual." The "real" part includes high-precision navigation history data extracted from actual voyages (such as AIS trajectory, hydrological and meteorological data, traffic density, etc.), the physical structural parameters of real ships (such as main dimensions, propulsion characteristics, and control system models), and the corresponding motion response characteristics of ships (such as turning radius, response time, acceleration and deceleration limits, etc.), forming the basis for the authenticity of the test scenario. The "virtual" part extends and reconstructs the "real" part, including a virtual navigation environment model built based on historical scenarios (reproducing key spatiotemporal elements). The system combines dynamic features and a digital twin model of the ship with real parameters to dynamically interact and respond to decision commands in the control closed loop. The resulting "virtual-real fusion navigation scenario" is a complete test platform built on the basis of real navigation background. It includes a dynamic interaction mechanism between the virtual ship and the environment and can be used to verify functional modules and extract indicators. The test platform can not only support independent or coupled testing of functional modules such as environmental element identification, avoidance decision, and adaptive control, but also realize the system comparative analysis of multiple repeated and controllable variables, providing a scientific and reliable basis for the performance evaluation of the intelligent navigation decision system. Key elements of the test scenario are the core basis for describing the test scenario and are the primary step in conducting virtual-real fusion testing. There are three main elements in a ship navigation scenario: the ship's own elements, the target ship's elements, and the environmental elements. A typical 76,000-ton fully loaded bulk carrier, "Huayang Ideal," was selected as the test ship, i.e., the vessel to be tested. The ship type of the vessel to be tested is highly representative and suitable for a complete navigation mission from the port waters, through various types of waters, and finally into the target port. It can fully reflect the various dynamic challenges faced by ships in the actual autonomous navigation environment. "Huayang Ideal" takes into account both ocean-going navigation capabilities and port berthing adaptability in its design and performance parameters. Its full-load displacement, main engine power, maneuvering characteristics, and other indicators are in line with the mainstream configuration of current medium and large dry bulk carriers. It has typicality and engineering practicality. Therefore, using it as the test object not only helps to demonstrate the adaptability of the test method in medium and large ship scenarios, but also provides a comparable reference for subsequent promotion to other ship types.

[0024] A high-precision modeling and simulation system is used to construct a complete motion model, propulsion system response model, and handling characteristic parameters of the vessel under test. This ensures that the virtual-real fusion testing environment can realistically reproduce the vessel's actual behavior in complex, transitional, and open waters. Furthermore, the actual operational limitations of the vessel under test, such as speed control lag, acceleration / deceleration restrictions, and steering response lag, are accurately reflected in the test model, thus guaranteeing the authenticity and reference value of the test data. For the vessel's key parameter information, please refer to Table 1. Table 1

[0025] As shown in Table 1, the main parameter information of the vessel to be tested includes the vessel's physical parameters and motion parameters; for the target vessel element data, AIS data from the Dalian-Incheon route between 00:00 on November 24th and 00:00 on November 25th, 2024, was selected. The AIS data for this period was interpolated and filtered using the cubic spline method. The processed data is shown in Table 2. Table 2

[0026] As shown in Table 2, the target vessel data includes the MMSI, latitude and longitude coordinates, heading, and speed of multiple vessels; the environmental data includes static and dynamic elements. Static environmental elements include navigation lanes, channel boundaries, medians, warning zones, anchorages, wharves, islands, shorelines, shoals, and buoys, which are geographically and facility-related and remain constant over time, providing a basic reference framework for navigation; dynamic environmental elements include the target vessel's heading, speed, position, and time-varying factors such as wind and current, which significantly affect real-time performance and navigation safety. For environmental element data for each waterway, please refer to Table 3. Table 3

[0027] As shown in Table 3, the environmental element data includes environmental elements of complex water areas, transitional water areas, and open water areas; real-time wind flow data at 08:00 on November 24, 2024, for Dalian Port, Incheon Port, and open water areas (122.031326°E, 38.69006°N) are shown in Table 4. Table 4

[0028] As shown in Table 4, the dynamic environmental data of the waterways between stations include the speed and direction of wind currents in each waterway. Classifying the environmental elements into models helps in accurate identification and modeling, providing effective support for collision risk assessment and decision-making during navigation. For the model classification of these environmental elements, please refer to Table 5. Table 5

[0029] As shown in Table 5, point-like and circular obstacles in the feature model can be considered as circles with different radii. , in, For the first The center of each region For the first The radius of each region; in the feature model, strip obstacles and polygonal obstacles can be represented by lines connecting multiple points. , in, For the first An irregular area For the points that make up the irregular region, The number of points; Quadrilateral isolation zone in the element model , in, Indicates the first An inaccessible area; Indicates the first The first region A boundary point; a linear boundary line in the feature model, with the two sides of the boundary line serving as the endpoints of the linear target. , in, , For line segments; , The number of points; The test station-to-station navigation decision-making system is applicable to complex waters, transitional waters, and open waters. The test scenario is the route from Dalian Pilot Station to Incheon Pilot Station, with the overall route running from northwest to southeast. The complex waters near Dalian Port include areas with fixed-line conditions. To meet navigation decision-making requirements, the traffic environment of the station-to-station waters is classified, deconstructed, and reorganized. Mathematical modeling is used to digitally represent static elements (such as channel boundaries and warning zones) and dynamic elements (such as target vessels, wind, and currents), constructing an environmental information database. This database supports real-time access and dynamic updates, providing accurate and efficient information support for intelligent and automated navigation decisions. The test scenario is constructed based on real historical AIS trajectory data and typical route characteristics, ensuring the authenticity and representativeness of the test environment. For a visual representation of the test scenario, please refer to [link / reference needed]. Figure 2 ,like Figure 2 As shown, the planned route is near 6 isolation zones marked in pink, 12 channel boundaries marked in pink dashed lines, 4 islands marked in yellow, 2 shallow water areas marked in light blue, 1 fishing area marked in blue, 2 pilot stations marked in red dashed lines, 2 anchorages, 12 target vessels marked in red dots, and 1 vessel marked in black dots.

[0030] In some embodiments, in step S102, the ship navigation decision system of the vessel under test is tested in a virtual-real fusion test scenario, and key test indicators are extracted. The ship navigation decision system is used to control the vessel under test to navigate from the origin port pilot station to the destination port pilot station based on historical ship navigation data. The "Specification" stipulates that in all navigation scenarios, the vessel should analyze and make decisions based on the perceived and acquired scenario information, control the propulsion and maneuvering systems according to the predetermined route, realize autonomous navigation and berthing and unberthing operations, and implement collision avoidance decisions and operations in accordance with the requirements of the "Rules". According to the requirements of the "Specification", the key technologies for autonomous navigation include situational awareness, motion control and navigation decision. The five functions of the three key technologies of the navigation decision system are tested respectively, namely, environmental element identification test, motion control test, rule and good seamanship compliance test, automatic collision avoidance and rerouting capability coupling test and navigation adaptability test. Environmental element identification testing involves using a navigation decision system to identify key elements across multiple scenarios from the originating port pilot station to the destination port pilot station. These key elements are then compared with historical data to obtain the first key test indicators, which include target detection accuracy, environmental update frequency, environmental update delay, and chart fusion accuracy. The test also examines the station-to-station route characteristics from Dalian Pilot Station to Incheon Pilot Station, encompassing multiple scenarios along the route. These scenarios include typical complex waters, transitional waters, and open waters within the route. Environmental element identification tests are designed for each of these scenarios. The objective of the tests in each waterway is to verify whether the navigation decision-making system can accurately, timely, and structurally identify typical elements that have a critical impact on navigation safety in the corresponding navigation scenarios. That is, to identify key elements, the typical environmental elements identified by the station-to-station navigation decision-making system in different waterways are integrated into the electronic nautical chart system, and the consistency and completeness with real historical data are verified by dynamically reproducing the traffic environment on the nautical chart. This determines the effectiveness and accuracy of the environmental identification module, and extracts the first key test indicators. The effectiveness and accuracy of the navigation decision-making system are then determined based on the first key test indicators.

[0031] Motion control testing involves setting a planned route from the originating port pilot station to the destination port pilot station, determining the motion control commands for the vessel under test to navigate along the planned route, and obtaining the corresponding second key test indicators based on the motion control commands. These second key test indicators include course holding error, speed stability, and control delay. Course holding error is the deviation between the set course and the actual course; speed stability is the range of actual speed fluctuations (change per unit time); and control delay is the time lag and degree of deviation in the execution of control commands. The planned navigation of the vessel under test is set along the route from the originating port pilot station to the destination port pilot station. Based on the set planned route, the vessel's course, speed, and trajectory are determined. Based on the vessel's course and speed... The speed and trajectory control system guides the test vessel along the planned route and obtains motion control commands, including vessel heading control commands, vessel speed control commands, and trajectory tracking commands. When the navigation decision system controls the test vessel to navigate along the planned route, it issues motion control commands, namely vessel heading control commands, vessel speed control commands, and trajectory tracking commands, and extracts a second key test index. Based on the second key test index, the navigation decision system is evaluated to verify the execution accuracy of vessel propulsion, steering, and other control commands, the vessel's response speed, and trajectory stability under the constraints of the planned path and collision avoidance strategy. This verifies the navigation decision system's ability to maintain a safe posture in variable waters. Motion control testing includes course control testing, speed control testing, and track tracking testing, all conducted under wind, wave, and current conditions. The purpose of course control is to ensure the vessel travels along a predetermined course, adjusting the course promptly according to different navigation conditions to maintain stability. The actual course must remain within the allowable deviation range of the target course. The purpose of speed control is to adjust the vessel's speed according to the navigation environment while ensuring safe navigation. To ensure safe navigation, the test model's accuracy must be as close as possible to or identical to the actual vessel. Adjusting the speed according to the navigation environment requires the vessel to be able to reach the target speed required by the ship's length in a timely manner through acceleration or deceleration (including completion within 30 minutes in transitional waters). Track tracking aims to guide the vessel along a predetermined route to its destination. This includes navigation along the planned route in complex waters. When navigating within a channel, maintain the center of the channel as much as possible and keep a certain distance from the channel boundary line; in transition waters and open waters, vessels should navigate along the planned route; determine and quantify test indicators, the course control test indicators are: initial course is 000°, course deviation should not exceed 1° after changing course to the left, right and left by 15°, 30° and 45° respectively; the speed control test indicators include speed accuracy (MMG model speed and actual ship speed accuracy are consistent), speed difference (the difference between the speed to a certain marker and the target speed should not exceed 0.5kn), and maintaining a constant speed within 30 minutes in transition waters (accelerating from full speed in port to sea speed (or decelerating from sea speed to full speed in port) within 30 minutes); the track tracking test indicators include: in complex waters, the distance from the vessel to both sides of the channel is consistent (if there is a channel), the distance to the planned route does not exceed 0.5 times the ship length, and in transition waters and open waters, the distance from the vessel to the planned route does not exceed 1 time the ship length.

[0032] The rule and good seamanship compliance test constructs typical encounter virtual scenarios in a virtual-real fusion test environment. In these virtual scenarios, the vessel under test implements collision avoidance decisions, obtaining a third key test indicator corresponding to the collision avoidance decisions. This third key test indicator includes rule compliance rate and seamanship rationality matching degree. Rule compliance rate refers to whether the collision avoidance behavior complies with the encounter type provisions in COLREGs, and seamanship rationality matching degree refers to the degree of matching with historical human maneuvering data trajectory. Article 8, "Collision Avoidance Actions," in the Rules stipulates that vessels should maintain a sufficient safe distance when taking collision avoidance measures. To improve the intelligence and computability of collision avoidance decisions, a ship domain (Ship) approach is proposed. The concept of a "domain" is used to quantify "safe distance" at the modeling level. A ship's domain refers to a dynamic safety space constructed around the ship, which should be protected from intrusion by other ships or obstacles to ensure navigational safety. This safety range is influenced not only by the ship's dimensions, speed, and maneuverability, but also by factors such as the captain's experience, navigation environment, and traffic density, thus allowing for some subjective adjustment. Common ship domain models include circular, semi-elliptical, elliptical, and irregular polygonal domains. Among these, the elliptical domain model is widely used in autonomous collision avoidance due to its simple structure, strong directional sensitivity, and adaptability to different speeds and collision avoidance direction changes. For a schematic diagram of an elliptical ship domain model, please refer to [link to schematic diagram]. Figure 3 ,use Figure 3 The elliptical vessel domain model shown serves as the basis for determining whether another vessel has entered the safe space of the vessel. The elliptical vessel domain model takes the bow direction of the vessel as the major axis of the ellipse, and the size of the domain is dynamically set according to the vessel type, speed and the complexity of the surrounding waters. It can accurately simulate the crew's perception of the safe distance boundary in actual collision avoidance, and provides a calculable basis for the judgment mechanism of "whether to trigger a collision avoidance command" in the navigation decision system. Considering the significant differences in traffic density, maneuverability, and collision avoidance strategies among vessels in different waters, corresponding avoidance rules have been formulated for complex waters, transitional waters, and open waters. In typical virtual encounter scenarios, the navigation decision system implements collision avoidance decisions for the vessel under test, which requires comprehensive consideration of the rules and the principles of good seamanship. A collision avoidance judgment mechanism adaptable to changes in multiple waters is constructed, introducing an encounter situation identification model based on hull angle comparison. By identifying the target vessel's bearing change trend relative to the vessel itself, the model determines whether there is a collision risk between vessels and the necessary avoidance measures. The navigation decision system's encounter situation identification model based on hull angle comparison uses the target vessel's hull angle changes and its relative motion trajectory as a basis, combined with speed difference, distance, and encounter type (face-to-face, crossing, overtaking) for situation identification. The identification results are then integrated with the traffic environment characteristics of different waters to construct a navigation-based... The collision avoidance rule base for various scenarios includes: in complex waters with high traffic density, deceleration, steering, and combined maneuvers are permitted for collision avoidance; in transitional waters where constant speed is required, steering is the primary method of collision avoidance; in open waters, due to limitations imposed by the main engine protection mechanism, significant speed changes are restricted, and course changes are the primary method of collision avoidance. Based on this navigation scenario-based collision avoidance rule base, the navigation decision system uses an encounter situation identification model based on hull angle comparison to determine various encounter situations and corresponding collision avoidance relationships in typical virtual encounter scenarios, and extracts a third key test indicator corresponding to the collision avoidance decision. For a schematic diagram of the collision avoidance rules, please refer to [link to relevant documentation]. Figure 4 ,like Figure 4 As shown, in a typical virtual encounter scenario between the vessel and the target vessel, determining the vessel's position involves first identifying whether it is in open, transitional, or complex waters; assessing the Collision Risk Index (CRI) between the vessel and the target vessel, and determining whether a collision risk exists based on the CRI; when a collision risk exists, identifying the encounter situation, such as a head-on encounter, crossing, or overtaking, and determining whether the vessel is the give-way vessel according to navigation rules; if the vessel is not the give-way vessel, it must maintain its course and speed while ensuring good seamanship, and wait for further action; determining whether the target vessel has taken evasive action, and executing evasive action, choosing to turn right, change speed, or combine both evasive actions depending on the encounter situation; and then confirming whether the selected evasive action is feasible. If not, immediate manual intervention is required.

[0033] The coupling test of automatic collision avoidance and return-to-course capability was conducted in a virtual-real fusion test scenario, controlling the test vessel to navigate from the origin port pilot station to the destination port pilot station. The automatic collision avoidance and return-to-course capability of the test vessel were coupled and tested to obtain the fourth key test index, which includes collision avoidance success rate, collision avoidance response time, and return-to-course trajectory deviation. On the route from the origin port pilot station to the destination port pilot station, the ship's decision-making system controlled the test vessel to depart from the starting point (i.e., the origin port pilot station) and navigate along the planned route to the destination. During the navigation, the ship's decision-making system controlled the test vessel to safely avoid the target vessel and return to the planned route, extracting the fourth key test index. The return-to-course accuracy of the navigation decision-making system was determined by the return-to-course trajectory deviation, and the automatic collision avoidance capability was determined by the collision avoidance success rate and collision avoidance response time. Collision avoidance safety; its return-to-course accuracy refers to the ability of a vessel to strictly follow the predetermined route after collision avoidance, involving precise control of course and speed to ensure that the vessel remains on the planned route after successful collision avoidance; high-precision return-to-course not only helps improve navigation safety but also optimizes navigation efficiency, reduces fuel consumption and travel time; after collision avoidance, the distance between the test vessel and the planned route is used as the quantitative criterion for judging return-to-course accuracy. When the distance from the complex waterway to the planned route does not exceed 0.5 times the vessel length, and the distance from the transitional waterway and open waterway to the planned route does not exceed 1 times the vessel length, it indicates high return-to-course accuracy; the territorial distance from the target vessel to the test vessel is used as the quantitative criterion for judging the safety of automatic collision avoidance, and the criterion is that the territorial distance from the target vessel to the test vessel is greater than or equal to 0.5 times the vessel length.

[0034] Ship adaptiveness testing involves pre-setting starboard crossing and head-on encounter scenarios in a virtual-real fusion test environment. Under these scenarios, the ship undergoes adaptive navigation tests to obtain the fifth key performance indicator (KPI), which includes course adjustment sensitivity, autonomous recovery rate, and decision robustness. The ship adaptiveness test verifies the navigation decision-making system's ability to handle non-ideal (extreme) situations and tests whether the route decision-making system can control the ship to navigate adaptively. In maritime practice, extreme situations may arise, such as target ships arbitrarily changing course or speed. In these cases, the navigator needs to promptly assess the encounter situation and maneuver the ship to avoid collisions within a very short time. Regarding the navigation decision system, when the target ship undergoes extreme changes, the adaptive navigation decision system can navigate adaptively and avoid collisions with the target ship. The adaptive navigation test scenarios are set as a right-hand crossing situation and a head-on encounter situation. In this case, the target ship is set as a virtual ship instead of being obtained from real AIS data, so as to achieve control over the target ship's speed and course. In the right-hand crossing situation and the head-on encounter situation, the navigation decision system is tested to ensure that the test ship can still safely avoid the target ship when the target ship changes speed and course significantly. The fifth key test indicator is extracted, and it is determined that the safety requirements are met when the distance between the test ship and the target ship is not less than 0.5 times the ship length.

[0035] In some embodiments, in step S104, the perception, control, and decision-making of the ship navigation decision-making system are evaluated based on key test indicators. Specifically, the perception of the ship navigation decision-making system is evaluated based on a first key test indicator; the control of the ship navigation decision-making system is evaluated based on a second key test indicator; the decision-making effectiveness and safety of the ship navigation decision-making system are evaluated based on a third, fourth, and fifth key test indicator; the ability of the navigation decision-making system to perceive key elements in different waters in a virtual-real fusion test scenario is evaluated based on target detection accuracy, environmental update frequency, environmental update delay, and chart fusion accuracy; the control capability of the navigation decision-making system of the ship under test is evaluated based on course maintenance error, speed stability, and control delay; and the decision-making effectiveness and safety of the navigation decision-making system are evaluated based on rule compliance rate, seamanship rationality matching degree, collision avoidance success rate, collision avoidance response time, retracing trajectory deviation, course adjustment sensitivity, autonomous recovery rate, and decision robustness.

[0036] In summary, the testing method for a ship navigation decision-making system provided by this invention constructs a virtual-real fusion test scenario based on historical ship navigation data from the originating port pilot station to the destination port pilot station, as well as the physical and motion parameters of the ship under test. The ship navigation decision-making system of the ship under test is tested within this virtual-real fusion test scenario, and key test indicators are extracted. The ship navigation decision-making system is used to control the ship under test to navigate from the originating port pilot station to the destination port pilot station based on historical ship navigation data. The perception, control, and decision-making of the ship navigation decision-making system are evaluated based on the key test indicators, thereby improving the completeness and realism of the navigation decision-making system test.

[0037] To better implement the testing method for the ship navigation decision system in this embodiment of the invention, based on the testing method for the ship navigation decision system, correspondingly, as follows: Figure 5 As shown, this embodiment of the invention also provides a testing device for a ship navigation decision-making system. The testing device 500 for the ship navigation decision-making system includes: The virtual-real test scenario construction module 501 is used to construct a virtual-real fusion test scenario based on historical ship navigation data from the origin port pilot station to the destination port pilot station, as well as the physical and motion parameters of the ship to be tested. Test module 502 is used to test the ship navigation decision system of the ship under test in a virtual-real fusion test scenario and extract key test indicators. The ship navigation decision system is used to control the ship under test to navigate from the origin port pilot station to the destination port pilot station based on historical ship navigation data. Evaluation module 503 is used to evaluate the perception, control, and decision-making of the ship's navigation decision-making system based on key test indicators.

[0038] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A test method for a ship navigation decision-making system, characterized in that, include: A virtual-real fusion test scenario is constructed based on historical ship navigation data from the origin port pilot station to the destination port pilot station, as well as the physical and motion parameters of the ship to be tested. The ship navigation decision system of the vessel under test is tested in the virtual-real fusion test scenario, and key test indicators are extracted. The ship navigation decision system is used to control the vessel under test to navigate from the origin port pilot station to the destination port pilot station based on the historical ship navigation data. The perception, control, and decision-making of the ship navigation decision-making system are evaluated based on the key test indicators.

2. The test method for the ship navigation decision system according to claim 1, characterized in that, The virtual-real fusion test scenario is constructed based on historical ship navigation data from the originating port pilot station to the destination port pilot station, as well as the physical and motion parameters of the ship under test. This includes: A virtual digital ship model is constructed based on the physical and kinematic parameters of the ship under test; Based on historical vessel navigation data from the originating port pilot station to the destination port pilot station, key elements of multiple scenarios from the originating port pilot station to the destination port pilot station are determined. These multiple scenarios include complex waters, transitional waters, and open waters. The key elements include vessel-specific elements, target vessel-specific elements, and environmental elements. A virtual scene is constructed based on the virtual digital ship model and key elements of multiple scenarios. The historical ship navigation data is then input into the virtual scene to construct a virtual-real fusion test scenario.

3. The test method for the ship navigation decision system according to claim 2, characterized in that, The navigation decision-making system of the vessel under test is tested in the virtual-real fusion test scenario, and key test indicators are extracted, including: The environmental element identification test involves using the navigation decision system to identify key elements in multiple scenarios from the originating port pilot station to the destination port pilot station. The identified key elements are then compared with historical data to obtain the first key test index, which includes target detection accuracy, environmental update frequency, environmental update delay, and chart fusion accuracy.

4. The test method for the ship navigation decision system according to claim 3, characterized in that, The process of testing the navigation decision-making system of the vessel under test in the virtual-real fusion test scenario and extracting key test indicators also includes: Motion control testing involves setting a planned route from the originating port pilot station to the destination port pilot station, determining the motion control command for the vessel under test to navigate along the planned route, and obtaining a second key test index corresponding to the motion control command based on the motion control command. The second key test index includes heading error, speed stability, and control delay.

5. The test method for the ship navigation decision system according to claim 4, characterized in that, The method of determining motion control commands for the vessel under test to navigate along the planned route, based on the planned route from the originating port pilot station to the destination port pilot station, includes: The ship's course, speed, and trajectory are determined based on the planned route. Based on the ship's heading, speed, and trajectory, the test ship is controlled to navigate along the planned route and motion control commands are obtained, including ship heading control commands, ship speed control commands, and trajectory tracking commands.

6. The test method for the ship navigation decision system according to claim 4, characterized in that, The process of testing the navigation decision-making system of the vessel under test in the virtual-real fusion test scenario and extracting key test indicators also includes: The rule and good seamanship compliance test involves constructing a typical encounter virtual scenario in the virtual-real fusion test scenario, implementing collision avoidance decisions on the vessel under test in the typical encounter virtual scenario, and obtaining a third key test indicator corresponding to the collision avoidance decision. The third key test indicator includes rule compliance rate and seamanship rationality matching degree.

7. The test method for the ship navigation decision system according to claim 5, characterized in that, The process of testing the navigation decision-making system of the vessel under test in the virtual-real fusion test scenario and extracting key test indicators also includes: The automatic collision avoidance and return-to-course capability coupling test involves controlling the vessel under test to navigate from the origin port pilot station to the destination port pilot station in the virtual-real fusion test scenario, and conducting a coupling test on the automatic collision avoidance and return-to-course capability of the vessel under test to obtain a fourth key test index, which includes collision avoidance success rate, collision avoidance response time, and return-to-course trajectory deviation.

8. The test method for the ship navigation decision system according to claim 6, characterized in that, The step of testing the navigation decision-making system of the vessel under test in the virtual-real fusion test scenario and obtaining test results also includes: The navigation adaptability test involves setting up right-hand crossing and encounter scenarios in the virtual-real fusion test scenario. The ship under test is then subjected to navigation adaptability tests in the right-hand crossing and encounter scenarios to obtain the fifth key test index, which includes course adjustment sensitivity, autonomous recovery rate, and decision robustness.

9. The test method for the ship navigation decision system according to claim 8, characterized in that, The evaluation of the ship navigation decision-making system's perception, control, and decision-making based on the key test indicators includes: The perception of the ship navigation decision system is evaluated based on the first key test indicator. The control of the ship navigation decision-making system is evaluated based on the second key test indicator; The decision-making effectiveness and safety of the ship navigation decision-making system are evaluated based on the third, fourth, and fifth key test indicators.

10. A testing device for a ship navigation decision-making system, characterized in that, include: The virtual-real test scenario construction module is used to construct virtual-real fusion test scenarios based on historical ship navigation data from the origin port pilot station to the destination port pilot station, as well as the physical and motion parameters of the ship under test. The testing module is used to test the ship navigation decision system of the vessel under test in the virtual-real fusion test scenario and extract key test indicators. The ship navigation decision system is used to control the vessel under test to navigate from the origin port pilot station to the destination port pilot station based on the historical ship navigation data. The evaluation module is used to evaluate the perception, control, and decision-making of the ship navigation decision-making system based on the key test indicators.

Images