Multi-ship collision avoidance method and system based on belief rule base clustering and speed obstacle method

Abstract

The invention provides a multi-ship collision avoidance method based on belief rule base clustering and a speed obstacle method, and relates to the technical field of ship navigation, and the method comprises the steps: determining the core characteristics of each to-be-clustered group obtained after grouping each ship according to the navigation data corresponding to each ship in a target ship setting range collected in real time, and according to the core features, adopting a confidence rule base to cluster each ship to obtain each ship group. And according to the average speed vector of each ship group, determining the relative speed between the target ship and each ship group, and according to the relative speed, adopting a speed obstacle method to determine a speed obstacle area. And according to each speed obstacle area, determining a collision avoidance speed vector, generating a collision avoidance path, and controlling the target ship to execute the collision avoidance path. The multiple ships near the target ship are clustered through the belief rule base, and the calculation amount is reduced. The speed obstacle area is determined based on the speed obstacle method, the dynamic collision avoidance path is generated, and the collision avoidance accuracy and efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the field of ship navigation technology, and in particular to a multi-ship collision avoidance method and system based on confidence rule base clustering and speed barrier method. Background Art

[0002] With the rapid development of the global shipping industry and the increase in maritime shipping traffic, especially in densely populated waters such as ports and waterways, the problem of ship encounter and collision avoidance has become a major factor affecting navigation safety.

[0003] Currently, in multi-ship encounter scenarios, ship-by-ship collision avoidance methods based on distance thresholds, dynamic programming, and rule-based collision avoidance methods can be used to achieve multi-ship collision avoidance. The ship-by-ship collision avoidance method based on distance thresholds sets a distance threshold for each ship, triggering a collision avoidance decision when other ships enter the range corresponding to the set distance threshold. The dynamic programming method generates a collision avoidance path based on the ship's real-time position, speed, heading, and other information. The rule-based collision avoidance method uses a preset rule base to determine whether a ship needs to yield.

[0004] However, in multi-ship encounter scenarios, both the ship-by-ship collision avoidance method based on distance thresholds and the dynamic programming method require ship-by-ship calculations, resulting in high computational complexity, low decision-making efficiency, and difficulty in achieving real-time response. The ship-by-ship collision avoidance method based on distance thresholds avoids collisions based on a pre-set distance threshold, while the rule-based collision avoidance method avoids collisions based on pre-set fixed rules. Both methods have poor adaptability in dynamic environments and struggle to quickly adjust decisions based on environmental changes. Therefore, how to quickly and accurately perform multi-ship collision avoidance based on confidence rule-based clustering and speed barrier methods is a crucial issue.

[0005] Based on this, this specification provides a multi-ship collision avoidance method based on confidence rule base clustering and speed barrier method. Summary of the Invention

[0006] In order to solve the problems of poor real-time performance, low accuracy and high computational complexity in multi-ship collision avoidance, this specification provides a multi-ship collision avoidance method and system based on confidence rule base clustering and speed barrier method, so as to cluster multiple ships near the target ship through the confidence rule base to reduce the number of ships that need to be avoided one by one, and generate a dynamic collision avoidance path for the target ship based on the speed barrier area between the target ship and each group of clustered ships determined by the speed barrier method, so as to avoid collision in real time and improve the accuracy and efficiency of collision avoidance.

[0007] This manual adopts the following technical solutions:

[0008] This specification provides a multi-ship collision avoidance method based on confidence rule base clustering and speed barrier method, including:

[0009] S1: Real-time collection of navigation data of all ships within the target ship’s set range;

[0010] S2: Grouping the ships to determine groups to be clustered, and determining core features corresponding to the groups to be clustered based on the navigation data corresponding to the groups to be clustered; wherein the group to be clustered includes two ships, and the core features include the distance, speed difference, heading difference, and average speed between the two ships in the group to be clustered;

[0011] S3: Clustering the ships according to the core features corresponding to the groups to be clustered using a pre-built confidence rule base to obtain ship groups; wherein the correlation between any two ships in each ship group is not less than a preset threshold, and the correlation is obtained based on the core features between the any two ships using the confidence rule base;

[0012] S4: determining an average speed vector corresponding to each of the ship groups, and determining a relative speed between the target ship and each of the ship groups based on the average speed vector of each of the ship groups, and determining a speed obstacle area between the target ship and each of the ship groups using a speed obstacle method based on the relative speeds;

[0013] S5: Determine the collision avoidance speed vector of the target ship according to each speed obstacle area;

[0014] S6: Generate a collision avoidance path for the target ship according to the collision avoidance velocity vector, and control the target ship to execute the collision avoidance path.

[0015] Optionally, the S3 specifically includes:

[0016] Determining the activation weight of each rule of each to-be-clustered group according to the core features of each to-be-clustered group and the set values corresponding to the respective preconditions in each rule in the pre-built confidence rule library;

[0017] Determining the degree of association between two ships in each to-be-clustered group according to the confidence and activation weight corresponding to each rule of each to-be-clustered group;

[0018] When the correlation degree is not less than a preset threshold, the two ships in each to-be-clustered group are divided into the same ship group;

[0019] When the correlation degree is less than a preset threshold, the two ships in each to-be-clustered group are not classified into the same ship group.

[0020] Optionally, the determining of the speed obstacle area between the target ship and each ship group using a speed obstacle method according to the relative speed in S4 specifically includes:

[0021] Determining a first radius and a preset radius margin corresponding to each ship group;

[0022] The current position of the target ship is determined, and a speed obstacle area between the target ship and each of the ship groups is determined using a speed obstacle method based on the relative speed, the first radius, the radius margin, and the current position.

[0023] Optionally, the S5 specifically includes:

[0024] Determine the performance indicators of the target ship;

[0025] determining a candidate speed vector of the target ship according to the performance index and each speed obstacle area;

[0026] A collision avoidance speed vector of the target ship is determined based on the candidate speed vectors.

[0027] Optionally, the method further includes:

[0028] S7: Real-time monitoring of the navigation environment within the set range of the target ship;

[0029] S8: When the navigation environment changes, re-collect navigation data of each ship within the set range of the target ship, and re-cluster the ships based on the re-collected navigation data and the confidence rule base to obtain new ship groups;

[0030] S9: regenerating each speed obstacle area using the speed obstacle method based on the new average speed vectors corresponding to each ship group;

[0031] S10: regenerating a collision avoidance path for the target ship according to the regenerated speed obstacle areas, and controlling the target ship to execute the regenerated collision avoidance path.

[0032] Optionally, the method further includes:

[0033] S11: During the execution of the collision avoidance path, the collision avoidance path is monitored in real time, and it is determined whether the collision avoidance path is effectively executed; if not, new speed obstacle areas are generated based on the average speed vectors corresponding to the respective ship groups using the speed obstacle method, and a new collision avoidance path for the target ship is regenerated based on the new speed obstacle areas, and the target ship is controlled to execute the new collision avoidance path.

[0034] This specification provides a multi-ship collision avoidance system based on confidence rule base clustering and speed barrier method, the system includes a data acquisition module, a feature extraction module, a confidence rule base clustering module and a path planning module; wherein:

[0035] The data acquisition module is used to collect the navigation data of all ships within the set range of the target ship in real time;

[0036] The feature extraction module is used to group the ships to determine the groups to be clustered, and determine the core features corresponding to the groups to be clustered based on the navigation data corresponding to the groups to be clustered; wherein the group to be clustered includes two ships, and the core features include the distance, speed difference, heading difference, and average speed between the two ships in the group to be clustered;

[0037] The confidence rule base clustering module is configured to cluster the ships using a pre-built confidence rule base according to the core features corresponding to the groups to be clustered, thereby obtaining ship groups; wherein the correlation between any two ships in each ship group is not less than a preset threshold, and the correlation is obtained using the confidence rule base based on the core features between the two ships;

[0038] The path planning module is configured to determine the average speed vector corresponding to each of the ship groups, determine the relative speed between the target ship and each ship group based on the average speed vector of each ship group, and determine the speed obstacle area between the target ship and each ship group using a speed obstacle method based on the relative speed; determine the collision avoidance speed vector of the target ship based on each speed obstacle area; generate a collision avoidance path for the target ship based on the collision avoidance speed vector, and control the target ship to execute the collision avoidance path.

[0039] Optionally, the confidence rule base clustering module is specifically used to determine the activation weight of each rule of each group to be clustered according to the core features of each group to be clustered and the set values corresponding to each precondition in each rule in the pre-built confidence rule base; determine the correlation between the two ships in each group to be clustered according to the confidence and activation weight corresponding to each rule of each group to be clustered; when the correlation is not less than a preset threshold, the two ships in each group to be clustered are divided into the same ship group; when the correlation is less than the preset threshold, the two ships in each group to be clustered are not divided into the same ship group.

[0040] Optionally, the path planning module is specifically used to determine the first radius and the preset radius margin corresponding to each ship group; determine the current position of the target ship, and use the speed obstacle method to determine the speed obstacle area between the target ship and each ship group based on the relative speed, the first radius, the radius margin and the current position.

[0041] Optionally, the path planning module is specifically used to determine a performance index of the target ship; determine a candidate speed vector of the target ship based on the performance index and each speed obstacle area; and determine a collision avoidance speed vector of the target ship based on the candidate speed vector.

[0042] Optionally, the system further comprises a dynamic update and feedback module;

[0043] The dynamic update and feedback module is used to monitor the navigation environment within the set range of the target ship in real time; when the navigation environment changes, the navigation data of each ship within the set range of the target ship is re-collected, and the ships are re-clustered based on the confidence rule base according to the re-collected navigation data to obtain new ship groups; based on the average speed vectors corresponding to the new ship groups, the speed obstacle method is used to regenerate each speed obstacle area; based on the regenerated speed obstacle areas, the collision avoidance path of the target ship is regenerated, and the target ship is controlled to execute the regenerated collision avoidance path.

[0044] Optionally, the dynamic update and feedback module is also used to monitor the collision avoidance path in real time during the execution of the collision avoidance path, and to determine whether the collision avoidance path is effectively executed; if not, the speed obstacle method is used to generate new speed obstacle areas based on the average speed vectors corresponding to each ship group, and a new collision avoidance path for the target ship is regenerated based on the new speed obstacle areas, and the target ship is controlled to execute the new collision avoidance path.

[0045] At least one of the above technical solutions adopted in this specification can achieve the following beneficial effects:

[0046] This specification provides a multi-ship collision avoidance method based on confidence rule base clustering and speed barrier method. The method can first collect the navigation data of all ships within a set range of the target ship in real time. The ships are grouped to determine each group to be clustered. Based on the navigation data corresponding to each group to be clustered, the core features corresponding to each group to be clustered are determined. The group to be clustered includes two ships. The core features include the distance, speed difference, heading difference, and average speed between the two ships in the group to be clustered. Based on the core features corresponding to each group to be clustered, a pre-built confidence rule base is used to cluster the ships to obtain each ship group. The average speed vector corresponding to each ship group is determined. Based on the average speed vector of each ship group, the relative speed between the target ship and each ship group is determined. Based on the relative speed, the speed barrier method is used to determine the speed barrier area between the target ship and each ship group. Based on each speed barrier area, the collision avoidance speed vector of the target ship is determined. Based on the collision avoidance speed vector, a collision avoidance path for the target ship is generated, and the target ship is controlled to execute the collision avoidance path. The confidence rule base is used to cluster multiple ships near the target ship, grouping ships with similar distances and trends into the same group. This reduces the number of ships that need to be avoided one by one, thereby reducing computational complexity. The confidence rule base, based on dynamic characteristics and confidence judgment capabilities, makes it more flexible in complex navigation environments. The speed barrier method is then used to determine the speed barrier area between the target ship and each clustered group of ships. Based on the speed barrier area, a dynamic collision avoidance path is generated for the target ship, allowing for real-time collision avoidance. This highly adaptable system can flexibly respond to complex environments and improve the accuracy and efficiency of collision avoidance.

[0047] In the present invention, based on the core features such as the distance, speed difference, navigation difference and average speed between two ships in the group to be clustered, activation weights are set between different rules through a trust rule base, and the correlation between the two ships in the group to be clustered is calculated. Based on the correlation, it is determined whether the two ships are clustered into the same ship group, so as to cluster the ships that meet the correlation requirements, thereby classifying ships with similar distances and trends into the same ship group, thereby reducing the number of ships that need to avoid independently, effectively reducing the computational burden, and ensuring real-time avoidance in a dense multi-ship environment.

[0048] The present invention uses a speed barrier method to generate speed barrier zones between each clustered group of ships and the target ship. Based on each speed barrier zone, a collision avoidance path for the target ship is generated. The target ship is dynamically adjusted according to the collision avoidance path, ensuring precise collision avoidance in a changing environment and minimizing collision risks. This ensures that each ship avoids the target while maintaining a safe distance, significantly improving the accuracy and efficiency of collision avoidance.

[0049] The present invention monitors the navigation environment within the set range of the target ship in real time, and automatically triggers multi-ship clustering, reconstruction of the speed obstacle area, and updating of the collision avoidance path when the navigation environment changes, so as to adapt to the collision avoidance needs in the new environment, ensure that accurate collision avoidance paths can be continuously provided in complex and changeable sea conditions, effectively improve adaptability and safety, and avoid the problems of insufficient real-time performance and response delay in complex multi-ship environments.

[0050] In the present invention, during the execution of the collision avoidance path, the collision avoidance path is monitored in real time and it is determined whether the collision avoidance path is effectively executed. If it is not effectively executed, the reconstruction of the speed obstacle area and the update of the collision avoidance path are triggered to ensure the effectiveness and safety of the collision avoidance decision. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The drawings described herein are used to provide a further understanding of this specification and constitute a part of this specification. The exemplary embodiments and descriptions of this specification are used to explain this specification and do not constitute an improper limitation of this specification. In the drawings:

[0052] Figure 1 This is a flow chart of a multi-ship collision avoidance method based on confidence rule base clustering and speed barrier method provided in this specification;

[0053] Figure 2 This is a schematic diagram of a multi-ship collision avoidance method based on confidence rule base clustering and speed barrier method provided in this specification;

[0054] Figure 3 This is a structural diagram of a multi-ship collision avoidance system based on a confidence rule base clustering and speed barrier method provided in this specification. DETAILED DESCRIPTION

[0055] To make the objectives, technical solutions, and advantages of this specification more clear, the following will clearly and completely describe the technical solutions of this specification in conjunction with the specific embodiments of this specification and the corresponding drawings. Obviously, the embodiments described are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this specification.

[0056] With the rapid development of the global shipping industry and the increase in maritime traffic, ship encounters and collision avoidance have become a major factor affecting navigation safety, particularly in densely populated areas such as ports and waterways. In complex multi-ship encounter scenarios, where ships are relatively close to each other and have similar navigational trends, analyzing each ship's collision avoidance path individually would lead to a significant increase in computational complexity, making it difficult to meet the system's real-time requirements. Ensuring real-time and accurate collision avoidance in multi-ship environments is a major challenge in current ship navigation technology.

[0057] In multi-ship encounter scenarios, the collision avoidance path for each ship needs to be calculated and updated one by one. However, this approach imposes a heavy computational burden in densely populated environments, resulting in a slow system response and an inability to meet the needs of real-time collision avoidance, especially in changing sea conditions. Existing technologies for multi-ship collision avoidance typically suffer from high computational load and poor real-time performance, making them difficult to adapt to changing environments and high-density navigation areas. Therefore, there is an urgent need for a method that can quickly and efficiently process collision avoidance decisions in complex multi-ship environments to reduce the computational burden, improve the real-time performance of decisions, and ensure navigation safety.

[0058] Based on this, this specification provides a multi-ship collision avoidance method and system based on confidence rule base clustering and speed barrier method. The following is a detailed description of the technical solutions provided by each embodiment of this specification in conjunction with the accompanying drawings.

[0059] Figure 1 This is a flow chart of a multi-ship collision avoidance method based on confidence rule base clustering and speed barrier method provided in this specification, as shown in FIG. Figure 1 As shown, the method specifically includes the following steps:

[0060] S1: Real-time collection of navigation data of all ships within the set range of the target ship.

[0061] In this specification, the device used for multi-ship collision avoidance can first collect real-time navigation data from all ships within a set range of the target ship. The device used for multi-ship collision avoidance can be a server, a system, or a module within a system. The system can be deployed on a vessel or on an electronic device such as a desktop or laptop computer. For ease of description, the multi-ship collision avoidance method based on confidence rule base clustering and speed barrier method provided in this specification will be described below, using the server as the execution entity.

[0062] The target ship setting range can be a pre-set range, or a range determined based on the current target ship's position according to a preset length. Specifically, the current target ship's position can be used as the starting point to determine an area of a preset length in a preset direction. The determined area is the target ship setting range. The preset direction can be the heading of the target ship, and the preset length can be pre-set data. Of course, the target ship setting range can also be determined in other ways, and this manual does not specifically limit it. The target ship setting range actually represents the area near the target ship, but it should be noted that the target ship is not within the target ship setting range. The target ship is the ship to be avoided, that is, the own ship, which means that the server can control the target ship to avoid collision.

[0063] The aforementioned navigation data can be collected in real time by the server through the Automatic Identification System (AIS). This navigation data may include position, speed, and heading, with the position being the latitude and longitude coordinates of the vessel. The real-time position, speed, and heading information of each vessel is obtained from the AIS (Automatic Identification System), allowing for the subsequent extraction of inter-vessel features such as distance, speed difference, heading difference, and average speed. This serves as the basis for clustering and path planning, ensuring that the BRB clustering and speed barrier method receive accurate input data, providing data support for the generation of real-time collision avoidance decisions.

[0064] S2: Grouping the ships to determine groups to be clustered, and determining core features corresponding to the groups to be clustered based on the navigation data corresponding to the groups to be clustered; wherein the group to be clustered includes two ships, and the core features include the distance, speed difference, heading difference, and average speed between the two ships in the group to be clustered.

[0065] In this specification, the server may first group each ship to determine each group to be clustered, and then determine the core features corresponding to each group to be clustered based on the navigation data corresponding to each group to be clustered. Wherein, the group to be clustered includes two ships. When grouping each ship and determining each group to be clustered, the server divides any two ships from each ship into one group to be clustered to obtain each group to be clustered. In addition, the server may also, for each ship and for each other ship except the ship, group the ship and the other ship as a group to be clustered. Of course, to ensure that each group to be clustered is different, the server may group the ship and the other ship as a first group and remove the same first groups from each first group to obtain each group to be clustered, so that each group to be clustered is different. For example, assuming there are 5 ships within the target ship setting range, namely ships 1 to 3, each first group may be ship 1 and ship 2, ship 1 and ship 3, ship 2 and ship 1, ship 2 and ship 3, ship 3 and ship 1, and ship 3 and ship 2. The same first combination in each first group is removed to obtain each group to be clustered, namely, ship 1 and ship 2, ship 1 and ship 3, and ship 2 and ship 3.

[0066] The core features corresponding to each of the above-mentioned groups to be clustered are obtained based on the navigation data corresponding to the two ships in each group to be clustered, and the core features include the distance, speed difference, heading difference and average speed between the two ships in the group to be clustered. When determining the core features corresponding to each of the above-mentioned groups to be clustered based on the navigation data corresponding to each of the above-mentioned groups to be clustered, the server can determine the core features corresponding to each of the groups to be clustered based on the navigation data corresponding to the two ships in each group to be clustered, that is, the distance, speed difference, heading difference and average speed between the two ships in each group to be clustered. The distance between the two ships in each group to be clustered can be calculated using the Haversine (half-verse) formula, that is, the Haversine formula is used to calculate the spherical distance between the two ships. Specifically, the following formula can be used for calculation:

[0067]

[0068] Among them, D ij Indicates the distance between ship i and ship j in the cluster group to be clustered, R is the radius of the earth, which can be 6378.137 kilometers, lat i is the latitude of ship i, lon i is the longitude of ship i, lat j is the latitude of ship j, lon j is the longitude of ship j. The longitude and latitude are both in radians.

[0069] The speed difference between the two ships in the above clustering group can be calculated using the following formula:

[0070] △V ij =|V i -V j |

[0071] Among them, △V ij represents the speed difference between ship i and ship j in the cluster group, V i represents the speed of ship i, V j represents the speed of ship j.

[0072] The heading difference between the two ships in the above clustering group can be calculated using the following formula:

[0073] △θ ij =|θ i -θ j |

[0074] Among them, △θ ij represents the heading difference between ship i and ship j in the cluster group, θ i represents the heading of ship i, θ j represents the heading of ship j.

[0075] The average speed between two ships in the above clustering group can be calculated using the following formula:

[0076]

[0077] in, Represents the average speed between ship i and ship j in the group to be clustered.

[0078] S3: Clustering the ships according to the core features corresponding to the groups to be clustered using a pre-built confidence rule base to obtain ship groups; wherein the correlation between any two ships in each ship group is not less than a preset threshold, and the correlation is obtained using the confidence rule base based on the core features between the any two ships.

[0079] In this specification, the server can cluster each ship using a pre-built belief rule base based on the core features corresponding to each group to be clustered, thereby obtaining each ship group. The belief rule base (BRB) is pre-built and constructed by the server based on the characteristic values (i.e., distance, speed difference, heading difference, and average speed) between each ship. The belief rule base is used to represent the correlation between different ships. The belief rule base includes rules, each of which includes at least one precondition, which is constructed based on the above-mentioned characteristic values, such as "short distance", "short speed difference", "short heading difference", etc., and each precondition has a pre-set set value. For ease of explanation, the following example is used to illustrate that each rule includes four preconditions. Each rule has a pre-set confidence level, which indicates the degree of trust in the rule when the preconditions included in the rule are met. The confidence level includes two probability values, namely, a clustering probability value and a non-clustering probability value, and the sum of these two probability values is 1. The correlation between any two ships in each ship group is not less than a preset threshold, and the correlation is obtained using a confidence rule base based on the core features between any two ships.

[0080] Specifically, the server can determine the activation weight of each rule in each clustering group based on the core features of each clustering group and the set values corresponding to each precondition in each rule in the pre-built confidence rule library. Based on the confidence and activation weight corresponding to each rule in each clustering group, the server determines the correlation between two ships within each clustering group. If the correlation is not less than a preset threshold, the two ships within each clustering group are classified as the same ship group. If the correlation is less than the preset threshold, the two ships within each clustering group are not classified as the same ship group.

[0081] The activation weight of each rule represents the relevance of each rule when the core features of each clustering group are input, that is, the degree to which each rule is activated when the core features of each clustering group are input. By introducing a confidence rule base, based on features such as distance, speed difference, heading difference, and average speed between ships, ships with similar distances and trends in multi-ship encounter scenarios are grouped together to reduce computational complexity. When clustering multiple ships, a confidence rule base is set to set activation weights between different features and generate dynamic correlations. Ships that meet the correlation requirements (i.e., no less than a preset threshold) are clustered, significantly reducing the number of ships that require independent calculations, effectively reducing computational complexity, and ensuring real-time performance in multi-ship environments. Compared with traditional ship-by-ship calculations, multi-ship clustering achieved through a confidence rule base and calculations based on clustered ship groups can be more flexible in complex navigation environments.

[0082] When determining the activation weight of each rule of each to-be-clustered group based on the core features of each to-be-clustered group and the set values corresponding to the preconditions in each rule in the pre-built confidence rule base, the server may use the following formula for calculation:

[0083]

[0084] Among them, M represents the rule R k The number of preconditions included in , that is, the number of features, x m is the input feature value, that is, the core feature of each group to be clustered, x k,m For rule R k The precondition m includes the setting value, x max is the maximum eigenvalue of the core features of each group to be clustered, x min The minimum eigenvalue of the core features of each group to be clustered. k Representation rule R k The activation weight of .

[0085] When determining the correlation between two ships in each to-be-clustered group based on the confidence and activation weight corresponding to each rule of each to-be-clustered group, the server can use the following formula for calculation:

[0086]

[0087] Where N is the total number of rules included in the confidence rule base, α k is the kth rule (i.e. rule R k ) activation weight, Belief k (x k ) is rule R k Confidence level. R ijis the correlation degree between two ships in each group to be clustered.

[0088] S4: Determine the average speed vector corresponding to each of the ship groups, and determine the relative speed between the target ship and each of the ship groups based on the average speed vector of each ship group; and determine the speed obstacle area between the target ship and each of the ship groups using the speed obstacle method based on the relative speed.

[0089] In this specification, the server can determine the average speed vector corresponding to each ship group, and determine the relative speed between the target ship and each ship group based on the average speed vector of each ship group, and determine the speed obstacle area between the target ship and each ship group based on the relative speed using the speed obstacle method. Among them, the average speed vector is the average value of the speed vectors of all ships in the ship group. The speed obstacle method is used to generate a dynamic collision avoidance path for each group of ships after clustering. By constructing a speed obstacle area, the safe speed range of each ship is determined, and a collision avoidance path is generated accordingly. The speed obstacle method takes into account the speed and direction of the ship when calculating the safe speed vector, ensuring that the ship can safely avoid collisions in a dense environment, significantly improving the accuracy and efficiency of collision avoidance. Compared with the traditional dynamic rule method and the fixed rule base method, the speed obstacle method has a lower real-time calculation amount and is more adaptable, and can respond flexibly in complex environments.

[0090] When determining the relative speed between the target ship and each ship group based on the average speed vector of each ship group, the server may use the following formula for calculation:

[0091] V rel =V a -V b

[0092] Among them, V rel represents the relative speed between the target ship a and the ship group b, V a represents the velocity vector of the target ship a, V b represents the average velocity vector of ship group b.

[0093] When determining the speed obstacle area between the target vessel and each ship group using the speed obstacle method based on relative speed, the server may determine a first radius corresponding to each ship group, determine the current position of the target vessel, and then, based on the relative speed, first radius, and current position, determine the speed obstacle area between the target vessel and each ship group using the speed obstacle method. The first radius represents the effective collision radius of each ship group. This first radius may be determined based on the current positions of all ships in each ship group. Specifically, the server may first determine the current positions of all ships in each ship group and use them as the first positions. Based on each first position, a first circular area is determined, and the radius of this first circular area is determined as the first radius. This first circular area includes the area of all first positions. This first circular area may be a circumscribed circle of a figure formed by all first positions. The specific method for determining the first circular area is not specifically limited in this specification; it is sufficient that the determined first circular area includes all first positions.

[0094] When the speed obstacle method is used to determine the speed obstacle area between the target ship and each ship based on the relative speed, the first radius and the current position, the server may first determine the position corresponding to each ship group, and determine the relative position based on the determined position and the current position of the target ship. Based on the relative position and the first radius, the speed obstacle method is used to determine the angle of the speed obstacle area. Based on the angle, a rotation matrix is generated, and based on the rotation matrix and the relative speed, the boundary vector of the speed obstacle area is determined. Based on the boundary vector, a speed obstacle area is generated. Among them, the position corresponding to each ship group may be the center position or the average position of the current positions of all ships in each ship group. When the speed obstacle method is used to determine the angle of the speed obstacle area based on the relative position and the first radius, the server may first determine the second radius of the target ship, which is the effective collision radius of the target ship, and the second radius may be pre-set. Then, based on the first radius, the second radius and the relative position, the speed obstacle method is used to determine the angle of the speed obstacle area, which may be calculated specifically using the following formula:

[0095]

[0096] Where θ represents the angle of the speed obstacle area, r A Indicates the first radius, r B represents the second radius, ΔP represents the relative position, and ||ΔP|| represents the modulus of the relative position.

[0097] When generating the rotation matrix based on the angle, the following formula can be used for calculation:

[0098]

[0099] Where R(θ) represents the rotation matrix.

[0100] When determining the boundary vector of the speed obstacle area based on the rotation matrix and the relative speed, the server may use the following formula for calculation:

[0101] V boundary1 =R(θ)·V rel

[0102] V boundary2 =R(-θ)·V rel

[0103] Among them, V boundary1 and V boundary2 The boundary vector representing the speed barrier area.

[0104] S5: Determine the collision avoidance speed vector of the target ship according to each speed obstacle area.

[0105] S6: Generate a collision avoidance path for the target ship according to the collision avoidance velocity vector, and control the target ship to execute the collision avoidance path.

[0106] In this specification, the server may first determine the collision avoidance speed vector of the target ship based on each speed obstacle area. Then, based on the collision avoidance speed vector, the collision avoidance path of the target ship is generated, and the target ship is controlled to execute the collision avoidance path. In the above, when determining the collision avoidance speed vector of the target ship based on each speed obstacle area, the server may first determine the performance index of the target ship, and determine the candidate speed vector of the target ship based on the performance index and each speed obstacle area. Based on the candidate speed vector, the collision avoidance speed vector of the target ship is determined. The performance index may include the maximum navigation speed of the target ship. In the above, when determining the candidate speed vector of the target ship based on the performance index and each speed obstacle area, the server may first determine each first speed vector outside each speed obstacle area, and then determine the first speed vector that meets the performance index from each first speed vector, and use it as the candidate speed vector.

[0107] When determining the collision avoidance speed vector of the target ship based on the candidate speed vectors in step S5, the server may use the following formula for calculation:

[0108]

[0109] Among them, V safe represents the collision avoidance velocity vector, VO represents each speed obstacle area, V represents the candidate velocity vector, V rel总 Represents the relative speed between the target ship and each ship group. It should be noted that V rel总 Multiple relative speeds can be expressed.

[0110] When generating the collision avoidance path of the target ship according to the collision avoidance speed vector in the above step S6, the server may generate the collision avoidance path of the target ship according to the current position of the target ship and the collision avoidance speed vector.

[0111] In some embodiments of this specification, to ensure consistent formatting of core features across clustering groups, thereby improving computational stability, in step S3, when clustering ships to obtain ship groups based on the core features corresponding to each clustering group, using a pre-established confidence rule base, the server may first perform feature mapping on the core features of each clustering group to obtain mapped features for each clustering group. Then, based on the mapped features corresponding to each clustering group, the server may use the pre-established confidence rule base to cluster ships to obtain ship groups. The mapped features are in the range of 0 to 1.

[0112] In some embodiments of the present specification, in step S3, when determining the association between two ships within each to-be-clustered group based on the confidence and activation weights corresponding to each rule of each to-be-clustered group, the server may first normalize the activation weights corresponding to each rule of each to-be-clustered group to obtain normalized activation weights. Then, based on the confidence and normalized activation weights corresponding to each rule of each to-be-clustered group, the association between two ships within each to-be-clustered group is determined. The sum of the normalized activation weights is 1.

[0113] In some embodiments of the present specification, to ensure safety between vessels, a radius margin may be added to the calculation when determining the speed obstacle area. Specifically, in step S4, when determining the speed obstacle area between the target vessel and each vessel group using the speed obstacle method based on the relative speed, the server may determine the first radius and a preset radius margin corresponding to each vessel group, determine the current position of the target vessel, and determine the speed obstacle area between the target vessel and each vessel group using the speed obstacle method based on the first radius, the radius margin, and the current position. The specific process of determining the speed obstacle area between the target vessel and each vessel group using the speed obstacle method based on the first radius, the radius margin, and the current position is similar to the process of determining the speed obstacle area between the target vessel and each vessel using the speed obstacle method based on the relative speed, the first radius, and the current position, and will not be further described here.

[0114] In some embodiments of this specification, in order to adapt to the dynamic changes of the navigation environment, the server may further perform the following steps:

[0115] S7: Real-time monitoring of the navigation environment within the set range of the target ship.

[0116] S8: When the navigation environment changes, the navigation data of each ship within the set range of the target ship is re-collected, and the ships are re-clustered based on the re-collected navigation data and the confidence rule base to obtain new ship groups.

[0117] S9: Based on the new average speed vectors corresponding to each ship group, the speed obstacle method is used to regenerate each speed obstacle area.

[0118] S10: Regenerate the collision avoidance path of the target ship according to the regenerated speed obstacle areas, and control the target ship to execute the regenerated collision avoidance path.

[0119] The above-mentioned navigation environment includes the navigation status of the speed, heading, etc. of each ship within the target ship setting range, and may also include whether new ships other than the above-mentioned ships appear within the target ship setting range. When the navigation status of any ship within the target ship setting range changes, or when a new ship other than the above-mentioned ships appears within the target ship setting range, it means that the navigation environment has changed, and the above-mentioned steps S1 to S6 are re-executed. The specific process is as described in the above-mentioned steps S1 to S6, which will not be repeated here. By real-time monitoring of the navigation environment (i.e., position, speed, and heading) near the target ship in the multi-ship encounter scenario (i.e., within the target ship setting range), if a new ship enters or the navigation status of an existing ship changes, the BRB clustering and speed barrier method are automatically triggered to update the path planning to adapt to the collision avoidance requirements in the new environment. This dynamic update mechanism ensures that the collision avoidance strategy adapts to environmental changes and avoids the real-time shortage and response delay problems of traditional collision avoidance in complex multi-ship environments. This mechanism ensures that the system can continuously provide accurate collision avoidance paths in complex and changeable sea conditions, effectively improving adaptability and safety. Traditional solutions usually lack this real-time adjustment capability and are difficult to adapt to rapidly changing navigation environments.

[0120] In some embodiments of this specification, in order to ensure that the target ship avoids collision according to the collision avoidance path, the server may further perform the following steps:

[0121] S11: During the execution of the collision avoidance path, the collision avoidance path is monitored in real time to determine whether it is being effectively executed. If not, a new speed obstacle area is generated using the speed obstacle method based on the average speed vector corresponding to each ship group. A new collision avoidance path is then generated for the target ship based on the new speed obstacle area, and the target ship is controlled to execute the new collision avoidance path.

[0122] The above judgment on whether the collision avoidance path is effectively executed is actually to judge whether the actual navigation path of the target ship deviates from the collision avoidance path, that is, to judge whether the actual navigation path of the target ship is consistent with the collision avoidance path. In addition, it can also be judged whether the distance between the target ship and any ship within the target ship setting range is less than the safety threshold. When the actual navigation path deviates from the collision avoidance path or the distance between the target ship and any ship within the target ship setting range is less than the safety threshold, it means that the collision avoidance path is not effectively executed, so the server can re-execute the above steps S4 to S6. The specific process is as described in the above steps S4 to S6, which will not be repeated here. By real-time monitoring of the execution of the collision avoidance path, if it is detected that the path deviates or the distance between ships is less than the safety threshold, it will trigger the re-planning of the path to ensure the effectiveness and safety of the collision avoidance decision.

[0123] In this manual, if Figure 2 As shown, Figure 2 This is a schematic diagram of a multi-ship collision avoidance method based on confidence rule base clustering and speed barrier method provided in this specification. First, Figure 2 From the "Start" in the target ship, the server can first collect the navigation data of each ship within the target ship setting range in real time through ASI, that is, Figure 2 In the "Collect ASI data (position, speed, heading)". After that, the server can group the ships, determine the groups to be clustered, and determine the core features (i.e., distance, speed difference, heading difference, average speed) corresponding to each group to be clustered according to the navigation data corresponding to each group to be clustered. Figure 2 The server can then determine the activation weight of each rule for each to-be-clustered group based on the core features of each to-be-clustered group and the corresponding set values of each prerequisite in each rule in the pre-built confidence rule base, i.e. Figure 2 After that, the server can determine the correlation between the two ships in each to-be-clustered group according to the confidence and activation weight corresponding to each rule of each to-be-clustered group, that is, Figure 2 When the correlation is not less than the preset threshold, the two ships in each to-be-clustered group are divided into the same ship group. When the correlation is less than the preset threshold, the two ships in each to-be-clustered group are not divided into the same ship group. Figure 2The server can then determine the average speed vector corresponding to each ship group, and based on the average speed vector of each ship group, determine the relative speed between the target ship and each ship group, and based on the relative speed, use the speed obstacle method to determine the speed obstacle area between the target ship and each ship group. Based on each speed obstacle area, the collision avoidance speed vector of the target ship is determined. Based on the collision avoidance speed vector, a collision avoidance path of the target ship is generated, and the target ship is controlled to execute the collision avoidance path, that is, Figure 2 "Path Planning (Based on Speed Obstacle Method)" in.

[0124] In addition, the server can also monitor the navigation environment within the target ship's set range in real time. When the navigation environment changes, the above steps S1 to S6 are re-executed. In addition, during the execution of the collision avoidance path, the collision avoidance path is monitored in real time to determine whether the collision avoidance path is effectively executed. If not, the above steps S4 to S6 are re-executed. Figure 2 Finally, to Figure 2 Until the "End" in the

[0125] based on Figure 1 The present specification also provides a multi-ship collision avoidance system based on a confidence rule base clustering and speed barrier method, specifically as follows Figure 3 As shown, Figure 3 This is a structural diagram of a multi-ship collision avoidance system using a confidence rule base clustering and speed barrier method provided in this specification. The system includes a data acquisition module 100, a feature extraction module 101, a confidence rule base clustering module 102, and a path planning module 103.

[0126] First, the data acquisition module 100 can collect the navigation data of all ships within the set range of the target ship in real time. Afterwards, the feature extraction module 101 can group the ships, determine the groups to be clustered, and determine the core features corresponding to the groups to be clustered based on the navigation data corresponding to the groups to be clustered. The navigation data may include position, speed, and heading. The group to be clustered includes two ships, and the core features include the distance, speed difference, heading difference, and average speed between the two ships in the group to be clustered. The specific execution process of the data acquisition module 100 is the same as that of the above Figure 1 The specific execution process of the feature extraction module 101 is the same as that of step S1 in the above. Figure 1 The process performed in step S2 is the same as that in FIG, and will not be repeated here.

[0127] Then, the above-mentioned belief rule base clustering module 102 can cluster each ship according to the core features corresponding to each group to be clustered, using a pre-constructed belief rule base to obtain each ship group. Among them, the belief rule base (BRB) is pre-constructed and includes various rules. Each rule includes at least one prerequisite, and each prerequisite has a preset set value. Each rule has a preset confidence level, which indicates the degree of trust in the rule when the prerequisite included in the rule is met. The confidence level includes two probability values, namely, a clustering probability value and a non-clustering probability value, and the sum of these two probability values is 1. The correlation between any two ships in each ship group is not less than a preset threshold. The correlation level is obtained using the belief rule base based on the core features between any two ships.

[0128] Specifically, the above-mentioned confidence rule base clustering module 102 can determine the activation weight of each rule of each group to be clustered according to the core features of each group to be clustered and the set values corresponding to each precondition in each rule in the pre-built confidence rule base. According to the confidence and activation weight corresponding to each rule of each group to be clustered, the correlation between the two ships in each group to be clustered is determined. When the correlation is not less than the preset threshold, the two ships in each group to be clustered are divided into the same ship group. When the correlation is less than the preset threshold, the two ships in each cluster group are not divided into the same ship group. The specific execution process of the above-mentioned confidence rule base clustering module 102 is the same as the above-mentioned Figure 1 The process is the same as that of step 3 in , so it will not be repeated here.

[0129] Then, the path planning module 103 can determine the average speed vector corresponding to each ship group, and determine the relative speed between the target ship and each ship group based on the average speed vector of each ship group, and use the speed obstacle method to determine the speed obstacle area between the target ship and each ship group based on the relative speed. According to each speed obstacle area, the collision avoidance speed vector of the target ship is determined. According to the collision avoidance speed vector, the collision avoidance path of the target ship is generated, and the target ship is controlled to execute the collision avoidance path. The calculation method of the above average speed vector and relative speed is the same as that of Figure 1 The calculation method in step S4 is the same as that in step S4, which will not be repeated here. The specific execution process of the above-mentioned path planning module 103 is the same as that in the above-mentioned step S4. Figure 1 The processes performed in steps S4 to S6 are the same and will not be repeated here.

[0130] When the speed obstacle method is used to determine the speed obstacle area between the target ship and each ship group based on the relative speed, the path planning module 103 can determine the first radius corresponding to each ship group and the current position of the target ship, and use the speed obstacle method to determine the speed obstacle area between the target ship and each ship group based on the relative speed, the first radius and the current position. The specific process is the same as Figure 1 The implementation process is consistent with step S4 in , and will not be repeated here.

[0131] When the speed obstacle method is used to determine the speed obstacle area between the target ship and each ship based on the relative speed, the first radius and the current position, the path planning module 103 may first determine the position corresponding to each ship group, and determine the relative position based on the determined position and the current position of the target ship. Based on the relative position and the first radius, the speed obstacle method is used to determine the angle of the speed obstacle area. Based on the angle, a rotation matrix is generated, and based on the rotation matrix and the relative speed, the boundary vector of the speed obstacle area is determined. The speed obstacle area is generated based on the boundary vector. The specific process is the same as Figure 1 The implementation process is consistent with step S4 in , and will not be repeated here.

[0132] When determining the collision avoidance speed vector of the target ship based on each speed obstacle zone, the path planning module 103 may first determine a performance index of the target ship and then determine a candidate speed vector for the target ship based on the performance index and each speed obstacle zone. The collision avoidance speed vector of the target ship is determined based on the candidate speed vectors. The performance index may include the maximum navigation speed of the target ship.

[0133] When determining the candidate speed vectors of the target ship based on the performance index and the speed obstacle areas, the path planning module 103 may first determine the first speed vectors outside the speed obstacle areas, and then determine the first speed vector that meets the performance index from the first speed vectors as the candidate speed vector.

[0134] In some embodiments of this specification, to ensure consistent formatting of core features across clustering groups, thereby providing computational stability, the confidence rule base clustering module 102 may first perform feature mapping on the core features of each clustering group to obtain mapped features for each clustering group, using a pre-established confidence rule base, based on the core features corresponding to each clustering group to obtain ship groups. The pre-established confidence rule base is then used to cluster the ships based on the mapped features corresponding to each clustering group to obtain ship groups. The mapped features are in the range of 0 to 1.

[0135] In some embodiments of the present specification, when determining the association between two ships within each to-be-clustered group based on the confidences and activation weights corresponding to the rules of each to-be-clustered group, the confidence rule base clustering module 102 may first normalize the activation weights corresponding to the rules of each to-be-clustered group to obtain normalized activation weights. Then, based on the confidences and normalized activation weights corresponding to the rules of each to-be-clustered group, the association between the two ships within each to-be-clustered group is determined. The sum of the normalized activation weights is 1.

[0136] In some embodiments of the present specification, to ensure safety between vessels, a radius margin may be added to the calculation when determining the speed obstacle area. Specifically, when the path planning module 103 uses the speed obstacle method to determine the speed obstacle area between the target vessel and each vessel group based on the relative speed, it may determine a first radius and a preset radius margin corresponding to each vessel group, determine the current position of the target vessel, and then use the speed obstacle method to determine the speed obstacle area between the target vessel and each vessel group based on the first radius, the radius margin, and the current position.

[0137] In some embodiments of this specification, to adapt to dynamic changes in the navigation environment, the system may further include a dynamic update and feedback module 104 that monitors the navigation environment within a set range of the target vessel in real time. When the navigation environment changes, navigation data for each vessel within the set range of the target vessel is recollected. Based on the recollected navigation data and a confidence rule base, the vessels are re-clustered to generate new ship groups. Based on the average speed vectors corresponding to the new ship groups, the speed obstacle method is employed to regenerate speed obstacle areas. Based on the regenerated speed obstacle areas, the target vessel's collision avoidance path is regenerated, and the target vessel is controlled to execute the regenerated collision avoidance path. Among them, the above-mentioned navigation environment includes the navigation status such as speed and heading of each ship within the target ship setting range, and may also include whether new ships other than the above-mentioned ships appear within the target ship setting range. When the navigation status of any ship within the target ship setting range changes, or when new ships other than the above-mentioned ships appear within the target ship setting range, it means that the navigation environment has changed, and the data acquisition module 100, feature extraction module 101, confidence rule base clustering module 102, and path planning module 103 are re-executed in this order. The specific process is as described in the above-mentioned data acquisition module 100, feature extraction module 101, confidence rule base clustering module 102, and path planning module 103, and will not be repeated here.

[0138] In some embodiments of the present specification, to ensure that the target vessel avoids collisions according to the collision avoidance path, the dynamic update and feedback module 104 may also monitor the collision avoidance path in real time during the execution of the collision avoidance path and determine whether the collision avoidance path is effectively executed. If not, the dynamic update and feedback module 104 may re-generate new speed obstacle areas based on the average speed vectors corresponding to each ship group using the speed obstacle method. Based on the new speed obstacle areas, a new collision avoidance path for the target vessel is re-generated, and the target vessel is controlled to execute the new collision avoidance path. The above-mentioned determination of whether the collision avoidance path is effectively executed is actually to determine whether the actual navigation path of the target vessel deviates from the collision avoidance path, that is, to determine whether the actual navigation path of the target vessel is consistent with the collision avoidance path. Alternatively, the dynamic update and feedback module 104 may determine whether the distance between the target vessel and any other vessel within the target vessel's set range is less than a safety threshold. If the actual navigation path deviates from the collision avoidance path or the distance between the target vessel and any other vessel within the target vessel's set range is less than the safety threshold, it indicates that the collision avoidance path is not effectively executed. Therefore, the dynamic update and feedback module 104 may re-execute the path planning module 103. The specific process is as described above for the path planning module 103 and will not be repeated here.

[0139] In this specification, the various modules of the system (i.e., data acquisition module 100, feature extraction module 101, confidence rule base clustering module 102, path planning module 103, and dynamic update and feedback module 104) work together to ensure efficient collision avoidance. The system integrates multiple key algorithms and functional modules to achieve real-time, efficient multi-vessel collision avoidance, providing safety assurance for navigation in high-density waters.

[0140] The foregoing is merely an example of the present invention and is not intended to limit the present invention. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

Claims

1. A multi-ship collision avoidance method based on confidence rule base clustering and speed barrier method, characterized in that: include: S1: Real-time collection of navigation data of all ships within the target ship’s set range; S2: Grouping the ships to determine groups to be clustered, and determining core features corresponding to the groups to be clustered based on the navigation data corresponding to the groups to be clustered; wherein the group to be clustered includes two ships, and the core features include the distance, speed difference, heading difference, and average speed between the two ships in the group to be clustered; S3: Clustering the ships according to the core features corresponding to the groups to be clustered using a pre-built confidence rule base to obtain ship groups; wherein the correlation between any two ships in each ship group is not less than a preset threshold, and the correlation is obtained based on the core features between the any two ships using the confidence rule base; S4: determining an average speed vector corresponding to each of the ship groups, and determining a relative speed between the target ship and each of the ship groups based on the average speed vector of each of the ship groups, and determining a speed obstacle area between the target ship and each of the ship groups using a speed obstacle method based on the relative speeds; S5: Determine the collision avoidance speed vector of the target ship according to each speed obstacle area; S6: Generate a collision avoidance path for the target ship according to the collision avoidance velocity vector, and control the target ship to execute the collision avoidance path.

2. The method according to claim 1, wherein The S3 specifically includes: Determining the activation weight of each rule of each to-be-clustered group according to the core features of each to-be-clustered group and the set values corresponding to the respective preconditions in each rule in the pre-built confidence rule library; Determining the degree of association between two ships in each to-be-clustered group according to the confidence and activation weight corresponding to each rule of each to-be-clustered group; When the correlation degree is not less than a preset threshold, the two ships in each to-be-clustered group are divided into the same ship group; When the correlation degree is less than a preset threshold, the two ships in each to-be-clustered group are not classified into the same ship group.

3. The method according to claim 1, wherein The step S4 of determining the speed obstacle area between the target ship and each ship group using the speed obstacle method according to the relative speed specifically includes: Determining a first radius and a preset radius margin corresponding to each ship group; The current position of the target ship is determined, and a speed obstacle area between the target ship and each of the ship groups is determined using a speed obstacle method based on the relative speed, the first radius, the radius margin, and the current position.

4. The method according to claim 1, wherein The S5 specifically includes: Determine the performance indicators of the target ship; determining a candidate speed vector of the target ship according to the performance index and each speed obstacle area; A collision avoidance speed vector of the target ship is determined based on the candidate speed vectors.

5. The method according to claim 1, wherein The method further comprises: S7: Real-time monitoring of the navigation environment within the set range of the target ship; S8: When the navigation environment changes, re-collect navigation data of each ship within the set range of the target ship, and re-cluster the ships based on the re-collected navigation data and the confidence rule base to obtain new ship groups; S9: regenerating each speed obstacle area using the speed obstacle method based on the new average speed vectors corresponding to each ship group; S10: regenerating a collision avoidance path for the target ship according to the regenerated speed obstacle areas, and controlling the target ship to execute the regenerated collision avoidance path.

6. The method according to claim 1, wherein The method further comprises: S11: During the execution of the collision avoidance path, the collision avoidance path is monitored in real time, and it is determined whether the collision avoidance path is effectively executed; if not, new speed obstacle areas are generated based on the average speed vectors corresponding to the respective ship groups using the speed obstacle method, and a new collision avoidance path for the target ship is regenerated based on the new speed obstacle areas, and the target ship is controlled to execute the new collision avoidance path.

7. A multi-ship collision avoidance system based on confidence rule base clustering and speed barrier method, characterized in that: The system includes a data acquisition module, a feature extraction module, a confidence rule base clustering module and a path planning module; wherein: The data acquisition module is used to collect the navigation data of all ships within the set range of the target ship in real time; The feature extraction module is used to group the ships to determine the groups to be clustered, and determine the core features corresponding to the groups to be clustered based on the navigation data corresponding to the groups to be clustered; wherein the group to be clustered includes two ships, and the core features include the distance, speed difference, heading difference, and average speed between the two ships in the group to be clustered; The confidence rule base clustering module is configured to cluster the ships using a pre-built confidence rule base according to the core features corresponding to the groups to be clustered, thereby obtaining ship groups; wherein the correlation between any two ships in each ship group is not less than a preset threshold, and the correlation is obtained using the confidence rule base based on the core features between the two ships; The path planning module is configured to determine the average speed vector corresponding to each of the ship groups, determine the relative speed between the target ship and each ship group based on the average speed vector of each ship group, and determine the speed obstacle area between the target ship and each ship group using a speed obstacle method based on the relative speed; determine the collision avoidance speed vector of the target ship based on each speed obstacle area; generate a collision avoidance path for the target ship based on the collision avoidance speed vector, and control the target ship to execute the collision avoidance path.

8. The system according to claim 7, wherein: The confidence rule base clustering module is specifically used to determine the activation weight of each rule of each to-be-clustered group according to the core features of each to-be-clustered group and the set values corresponding to the respective preconditions in each rule in the pre-built confidence rule base; Determining the degree of association between two ships in each to-be-clustered group according to the confidence and activation weight corresponding to each rule of each to-be-clustered group; When the correlation degree is not less than a preset threshold, the two ships in each to-be-clustered group are divided into the same ship group; When the correlation degree is less than a preset threshold, the two ships in each to-be-clustered group are not classified into the same ship group.

9. The system according to claim 7, wherein: The path planning module is specifically configured to determine a first radius and a preset radius margin corresponding to each ship group; determine a current position of the target ship; and determine a speed obstacle area between the target ship and each ship group using a speed obstacle method based on the relative speed, the first radius, the radius margin, and the current position.

10. The system according to claim 7, wherein: The path planning module is specifically used to determine the performance index of the target ship; determine the candidate speed vector of the target ship according to the performance index and each speed obstacle area; and determine the collision avoidance speed vector of the target ship according to the candidate speed vector.

11. The system according to claim 7, wherein: The system also includes a dynamic update and feedback module; The dynamic update and feedback module is used to monitor the navigation environment within the set range of the target ship in real time; when the navigation environment changes, re-collect navigation data of each ship within the set range of the target ship, and re-cluster the ships based on the re-collected navigation data and the confidence rule base to obtain new ship groups; Based on the new average speed vectors corresponding to each ship group, the speed obstacle method is used to regenerate each speed obstacle area; according to the regenerated speed obstacle area, the collision avoidance path of the target ship is regenerated, and the target ship is controlled to execute the regenerated collision avoidance path.

12. The system according to claim 11, wherein The dynamic update and feedback module is further used to monitor the collision avoidance path in real time during the execution of the collision avoidance path, and to determine whether the collision avoidance path is effectively executed; if not, to regenerate new speed obstacle areas based on the average speed vectors corresponding to each of the ship groups using the speed obstacle method, and to regenerate a new collision avoidance path for the target ship based on the new speed obstacle areas, and to control the target ship to execute the new collision avoidance path.

Images