An intelligent decision-making method for ship collision avoidance based on improved brainstorming optimization algorithm

By improving the brainstorming optimization algorithm, combining K-means clustering and adaptive update to generate new individuals, the problem of local optimal solutions and time-consuming in the ship collision avoidance method is solved, and the rapid finding of the global optimal solutions is achieved. The effect improves the efficiency and safety of ship collision avoidance decisions.

CN115620558BInactive Publication Date: 2024-11-15JIMEI UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210922740.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-08-02
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing ship collision avoidance methods are prone to falling into local optimal solutions, which takes a long time and is difficult to quickly find the global optimal solutions, resulting in low decision-making efficiency.

Method used

Using an improved brainstorming optimization algorithm, the ship's steering angle and navigation navigation status data are obtained by obtaining electronic chart data and ship navigation status data, and the trigger thresholds for new individuals are updated using K-means clustering and adaptive updates to search for ship steering angles and navigation with the best collision avoidance effect. Time, combined with electronic chart display to improve decision-making efficiency.

Benefits of technology

The speed of searching for global optimal solutions is accelerated, the effectiveness and speed of decision-making is improved, the possibility of local optimal solutions is reduced, the global optimization ability of the algorithm is enhanced, and the navigation safety of ship collision avoidance is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115620558B_ABST
    Figure CN115620558B_ABST
Patent Text Reader

Abstract

The present invention relates to an intelligent decision-making method for ship collision avoidance based on an improved brainstorming optimization algorithm, which first obtains electronic chart data and navigation status data of other ships in the ship's navigation waters, as well as the spatial relative position relationship between the ship and the offshore wind farm; then determines the encounter relationship between the ships; and adaptively updates the trigger threshold of the new individual to search for the ship steering angle with the best collision avoidance effect and the sailing time after the ship turns; the calculated best collision avoidance steering angle and steering sailing time are used as one of the optimization sub-goals to prohibit the ship from steering to avoid the wind farm side, which can meet the navigation safety in the waters near the offshore wind farm. The present invention accelerates the search for the global optimal solution and improves the effectiveness and speed of decision-making by improving the brainstorming optimization algorithm.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of ship control, and in particular to an intelligent decision-making method for ship collision avoidance based on an improved brainstorming optimization algorithm. Background Art

[0002] In recent years, in order to reduce the occurrence of ship collision accidents, a variety of ship navigation aids have been installed on ships, including the Global Position System (GPS), Electronic Chart Display and Information System (ECDIS), Automatic Identification System (AIS), and ship radar. In order to provide ship drivers with more quantitative ship collision avoidance decision support, researchers proposed a ship collision avoidance path planning method.

[0003] The existing collision avoidance method is intelligent search, which has the problems of being easily trapped in local optimal solutions and being time-consuming. Summary of the invention

[0004] The object of the present invention is to provide a ship collision avoidance intelligent decision-making method based on an improved brainstorming optimization algorithm, which accelerates the search for a global optimal solution and improves the effectiveness and speed of decision-making by improving the brainstorming optimization algorithm.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] An intelligent decision-making method for ship collision avoidance based on an improved brainstorming optimization algorithm comprises the following steps:

[0007] Step 1: Obtain electronic chart data and navigation status data of other ships near the ship's navigation waters; and obtain the spatial relative position relationship between the ship and the offshore wind farm;

[0008] Step 2: Based on the judgment of the movement trend between the ships, it is judged whether the ships encounter each other and the encounter relationship between the ships is determined;

[0009] Step 3: Search for the best collision avoidance effect of the ship turning angle and the sailing time after the ship turns, and generate an initial solution. Each individual corresponds to the initial solution, including the turning angle x1 and the sailing time x2 after the ship turns.

[0010] The step 3 specifically includes:

[0011] 1) Generate the initial population;

[0012] Each individual uses s k=(x1,x2) means that the first parameter is the turning angle x1 of the ship, and the second parameter is the sailing time x2 after the ship turns;

[0013] 2) Use K-means clustering method to divide individuals into C Clusters, record the cluster center of each cluster;

[0014] 3) Generate a random number r1 in the interval (0, 1). If r1 <p a , then randomly select a cluster center and replace it with a randomly generated individual; where p a is a given constant value.

[0015] 4) Generate a random number r2 in the interval (0, 1); if r2 <p b , then the probability of randomly selecting a i class; generate a random number r3 in the interval (0, 1), if r3 <p c , then select the center of this class and add a random value to generate a new individual; otherwise, when r3>p c , randomly select individuals in this class and update them with random values; where p b and p c is a given constant value.

[0016] 5) If r2>p b , then randomly select two classes to generate new individuals; generate a random number r4 in the interval (0, 1), if r4 <p d , then merge the two cluster centers and add random values ​​to generate new individuals; otherwise, when r4>p d , select individuals from two clusters and merge them, then add random values ​​to generate new individuals; where p d is a given constant value;

[0017] (7)

[0018] in and represents the two individuals to be merged, Represents the combined individual;

[0019] 6) The newly generated individuals are compared with the current individuals, and the individuals with smaller fitness f value are used as the new individuals for the next iteration; the population fitness is composed of the distance to the nearest approach point, the turning amplitude and the distance to the original route, and the prohibition of ships steering to avoid the wind farm side, and is calculated as follows:

[0020] (8)

[0021] Among them, DCPA is the distance from the nearest approach point to the population fitness, w1, w2、 w 3、 w4 is the weight of each sub-goal; sub-goal f4 is to prohibit ships from steering to avoid the wind farm, and its calculation is as follows:

[0022] (9)

[0023] in, It represents the angle between the original route of the ship and the wind farm before the avoidance. It represents the angle between the new route of the ship and the wind farm after the avoidance;

[0024] 7) If N individuals have been updated, go to step 8), otherwise go to step 4);

[0025] 8) If the maximum number of iterations is reached, stop the iteration, otherwise update p as follows a , go to step 3);

[0026] (10)

[0027] in, represents the p of the tth iteration a Value, p a0 Indicates the initial p a value, represents the i-th individual in the t-th iteration, represents the optimal individual of the tth iteration.

[0028] The step 2 is specifically as follows:

[0029] On the basis of judging the trend of the relative distance between the two ships, the relative speed and direction of the two ships are compared. Based on the data collected in step 1, the ship encounter data is dynamically extracted from the data. By judging the ship classification, ship time division, and ship relative motion trend, the ship encounter recognition is performed:

[0030] Assume the ship's speed is , the target ship speed is , the relative speed of the two ships is: (1)

[0031] The relative heading is: (2)

[0032] Said The calculation method is as follows:

[0033] (3)

[0034] R T is the distance between the two ships. If the longitude and latitude of the ship are . The latitude and longitude of the target ship are The relative positions of the two ships are for: (4)

[0035] remember: (5)

[0036] The closest approach distance DCPA between the two ships is: (6)

[0037] The DCPA for the closest approach of the two vessels reflects the highest acceptable navigation risk.

[0038] In open waters, if DCPA < 2 nautical miles, it is judged that the two ships will meet; in narrow waters, if DCPA < 0.5 nautical miles, it is judged that the two ships will meet.

[0039] The method further comprises:

[0040] Step 4: The optimal collision avoidance steering angle and steering navigation time calculated in step 3 are integrated and displayed on the electronic nautical chart.

[0041] In step 1, the navigation status data of the ship includes latitude, longitude, heading, and speed.

[0042] After adopting the above scheme, the present invention first obtains the electronic chart data and the navigation status data of other ships in the ship's navigation waters, as well as the spatial relative position relationship between the ship and the offshore wind farm; then determines the encounter relationship between the ships; and adaptively updates the trigger threshold of the new individual to search for the best collision avoidance effect of the ship steering angle and the sailing time after the ship turns; the calculated best collision avoidance steering angle and steering sailing time are used as one of the optimization sub-goals to prohibit the ship from steering to avoid the wind farm side, which can meet the navigation safety in the waters near the offshore wind farm. The present invention improves the brainstorming optimization algorithm, speeds up the search for the global optimal solution, improves the effectiveness and speed of decision-making, and solves the problems that the existing method is prone to fall into the local optimal solution and is time-consuming. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a flow chart of the method of the present invention;

[0044] Figure 2 To judge the encounter pattern of ships;

[0045] Figure 3 This is the convergence curve of the improved brainstorming optimization algorithm. DETAILED DESCRIPTION

[0046] like Figure 1 As shown, the present invention discloses an intelligent decision-making method for ship collision avoidance based on an improved brainstorming optimization algorithm, which comprises the following steps:

[0047] Step 1: Obtain electronic chart data and navigation status data of other ships near the ship's navigation area; and obtain the spatial relative position relationship between the ship and the offshore wind farm. The navigation status data of the ship includes latitude, longitude, heading, speed and other information.

[0048] Step 2: Based on the judgment of the movement trends between the ships, it is judged whether the ships encounter each other and the encounter relationship between the ships is determined.

[0049] Specifically, on the basis of judging the trend of the relative distance between the two ships, the relative speed and direction of the two ships are compared, and based on the data collected in step 1, the ship encounter data is dynamically extracted from the data, and the ship encounter recognition is performed through ship classification, ship time division, and ship relative motion trend judgment:

[0050] Assume the ship's speed is , the target ship speed is , the relative speed of the two ships is: (1)

[0051] The relative heading is: (2)

[0052] Said The calculation method is as follows:

[0053] (3)

[0054] R T is the distance between the two ships. If the longitude and latitude of the ship are . The latitude and longitude of the target ship are The relative positions of the two ships are for: (4)

[0055] remember: (5)

[0056] The closest approach distance DCPA between the two ships is: (6)

[0057] The DCPA of the two ships reflects the highest acceptable navigation risk. According to the practice of seafarers, it is generally set at 2 nautical miles in open waters and 0.5 nautical miles in narrow waters. If DCPA < 0.5 nautical miles, it is judged that the two ships will meet.

[0058] Step 3: Search for the best collision avoidance effect of the ship turning angle and the sailing time after the ship turns, and generate an initial solution. Each individual corresponds to the initial solution, including the turning angle x1 and the sailing time x2 after the ship turns.

[0059] The details are as follows:

[0060] 3) Generate the initial population.

[0061] Each individual uses s k =(x1,x2) indicates that the first parameter is the ship's turning angle x1. If the turning range is too large, the ship will deviate from the original route to a large extent, increasing unnecessary energy consumption. When the ship completes the turning collision avoidance operation, it will resume its original course. The second parameter is the sailing time x2 after the ship turns. If the sailing time of the ship on the new route is too short, other target ships will not be able to correctly judge the next action, which will also increase the difficulty of ship manipulation. If the time is too long, the ship will seriously deviate from the original route, increasing the difficulty of restoring the original route. After sailing on the new route, the ship will resume operation and return to the original route.

[0062] 2) Use K-means clustering method to divide individuals into C clusters and record the cluster center of each cluster.

[0063] 3) Generate a random number r1 in the interval (0, 1). If r1 <p a , then randomly select a cluster center and replace it with a randomly generated individual. a is a given constant value.

[0064] 4) Generate a random number r2 in the interval (0, 1). If r2 <p b , then the probability of randomly selecting a i Generate a random number r3 in the interval (0, 1), if r3 <p c , then select the center of this class and add a random value to generate a new individual; otherwise, when r3>p c , randomly select individuals in this class and update them with random values. b and p c is a given constant value.

[0065] 5) If r2>p b , then randomly select two classes to generate new individuals. Generate a random number r4 in the interval (0, 1). If r4 <p d , then merge the two cluster centers and add random values ​​to generate new individuals; otherwise, when r4>p d , select individuals from two clusters and merge them, then add random values ​​to generate new individuals. d is a given constant value.

[0066] (7)

[0067] in and represents the two individuals to be merged, Represents the combined individual.

[0068] 6) The individuals with smaller fitness f value are used as the new individuals for the next iteration compared with the current individuals. The fitness of the population is composed of the distance to the nearest approach point (DCPA), the turning amplitude and the distance to the original route, and the prohibition of the ship from steering to avoid the wind farm. It can effectively evaluate the risk status encountered by the ship and provide navigation auxiliary decision support for the driver.

[0069] The calculation is as follows:

[0070] (8)

[0071] Among them, w1, w 2、 w 3、 w4 is the weight of each sub-goal.

[0072] Sub-goal f4 is to prohibit the ship from steering to avoid the wind farm, and its calculation is as follows:

[0073] (9)

[0074] in, It represents the angle between the original route of the ship and the wind farm before the avoidance. It indicates the angle between the new route of the ship after avoidance and the wind farm.

[0075] 7) If the update of N individuals is completed, go to step 8), otherwise go to step 4).

[0076] 8) If the maximum number of iterations is reached, stop the iteration, otherwise update p as follows a , go to step 3).

[0077] (10)

[0078] in represents the p of the tth iteration a Value, p a0 Indicates the initial p a value, represents the i-th individual in the t-th iteration, Represents the optimal individual of the tth iteration. The trigger threshold p of the new individual is generated by adaptively updating a , which is conducive to reducing the possibility of the algorithm falling into the local optimal solution and enhancing the global optimization ability of the algorithm.

[0079] Step 4: The optimal collision avoidance steering angle and steering navigation time calculated in step 3 are integrated and displayed on the electronic nautical chart to enhance the visualization effect.

[0080] In summary, the present invention first obtains electronic chart data and navigation status data of other ships in the ship's navigation waters, as well as the spatial relative position relationship between the ship and the offshore wind farm; then determines the encounter relationship between the ships; and adaptively updates the trigger threshold of the new individual to search for the best ship steering angle and the sailing time after the ship turns; the calculated best collision avoidance steering angle and steering sailing time are used as one of the optimization sub-goals to prohibit the ship from steering to avoid the wind farm side, which can meet the navigation safety in the waters near the offshore wind farm.

[0081] The present invention adaptively updates the trigger threshold for generating new individuals, which is conducive to reducing the possibility of the brainstorming optimization algorithm falling into the local optimal solution and enhancing the global optimization ability of the algorithm. Intelligent collision avoidance mechanism for ship collision avoidance based on improved swarm intelligence. In addition, in order to speed up the search speed of the optimal decision-making system for automatic collision avoidance of ships, the present invention adds a random normal distribution disturbance term on the basis of the original population movement, which is conducive to reducing the possibility of falling into the local optimal solution and enhancing the global optimization ability.

[0082] The present invention can provide intelligent routes and collision avoidance warning services, and send warnings to the ship in advance when there is a risk of collision. It meets the demand for dynamically changing maritime safety information in modern navigation and can help improve crew efficiency and ship navigation safety.

[0083] In ship navigation, the information obtained from electronic charts, AIS and other navigation equipment is used to superimpose and integrate the results of prediction analysis and auxiliary decision-making, so as to improve the safety of ship navigation and the accuracy of decision-making. For example, ship collision avoidance auxiliary decision-making information is displayed on electronic charts to minimize maritime accidents caused by subjective misjudgment and reduce the workload and intensity of drivers. The present invention improves the brainstorming optimization algorithm, speeds up the search for the global optimal solution, improves the effectiveness and speed of decision-making, and solves the problems that the existing methods are prone to fall into local optimal solutions and are time-consuming.

[0084] The above description is merely an embodiment of the present invention and does not limit the technical scope of the present invention. Therefore, any minor modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. An intelligent decision-making method for ship collision avoidance based on an improved brainstorming optimization algorithm. Features: The method comprises the following steps: Step 1: Obtain electronic chart data and navigation status data of other ships near the ship's navigation waters; and obtain the spatial relative position relationship between the ship and the offshore wind farm; Step 2: Based on the judgment of the movement trend between the ships, it is judged whether the ships encounter each other and the encounter relationship between the ships is determined; Step 3: Search for the best collision avoidance effect of the ship steering angle and the sailing time after the ship turns, and generate an initial solution. Each individual corresponds to the initial solution, including the steering angle x 1 and the sailing time x after the ship turns 2 ; The step 3 specifically includes: 1) Generate the initial population; Each individual uses s k =(x 1 ,x 2 ) indicates that the first parameter is the ship's turning angle x 1 , the second parameter is the sailing time x after the ship turns 2 ; 2) Use K-means clustering method to divide individuals into C Clusters, record the cluster center of each cluster; 3) Generate a random number r in the interval (0, 1) 1 , if r 1 <p a , then randomly select a cluster center and replace it with a randomly generated individual; where p a is a given constant value; 4) Generate a random number r in the interval (0, 1) 2 ; if r 2 <p b , then the probability of randomly selecting a i Class; generates a random number r in the interval (0, 1) 3 , if r 3 <p c , then select the center of this class and add a random value to generate a new individual; otherwise, when r 3 >p c , randomly select individuals in this class and update them with random values; where p b and p c is a given constant value; 5) If r 2 >p b , then two classes are randomly selected to generate new individuals; the random number r generated is within the interval (0, 1) 4 , if r 4 <p d , then merge the two cluster centers and add random values ​​to generate new individuals; otherwise, when r 4 >p d , select individuals from two clusters and merge them, then add random values ​​to generate new individuals; where p d is a given constant value; (7) in and represents the two individuals to be merged, Represents the combined individual; 6) The newly generated individuals are compared with the current individuals, and the individuals with smaller fitness f value are used as the new individuals for the next iteration; the population fitness is composed of the distance to the nearest approach point, the turning amplitude and the distance to the original route, and the prohibition of ships steering to avoid the wind farm side, and is calculated as follows: (8) Among them, DCPA is the distance from the population fitness to the nearest approach point, w 1 、w 2、 w 3、 w 4 is the weight of each sub-goal; sub-goal f4 is to prohibit ships from steering to avoid the wind farm, and its calculation is as follows: (9) in, It represents the angle between the original route of the ship and the wind farm before the avoidance. It represents the angle between the new route of the ship and the wind farm after the avoidance; 7) If N individuals have been updated, go to step 8), otherwise go to step 4); 8) If the maximum number of iterations is reached, stop the iteration, otherwise update p as follows a , go to step 3); (10) in, represents the p of the tth iteration a Value, p a0 Indicates the initial p a value, represents the i-th individual in the t-th iteration, represents the optimal individual of the tth iteration.

2. According to claim 1, a ship collision avoidance intelligent decision-making method based on an improved brainstorming optimization algorithm, Features: The step 2 is specifically as follows: On the basis of judging the trend of the relative distance between the two ships, the relative speed and direction of the two ships are compared. Based on the data collected in step 1, the ship encounter data is dynamically extracted from the data. By judging the ship classification, ship time division, and ship relative motion trend, the ship encounter recognition is performed: Assume the ship's speed is , the target ship speed is , the relative speed of the two ships is: (1) The relative heading is: (2) Said The calculation method is as follows: (3) R T is the distance between the two ships. If the longitude and latitude of the ship are . The latitude and longitude of the target ship are The relative positions of the two ships are for: (4) remember: (5) The closest approach distance DCPA between the two ships is: (6) The DCPA for the closest approach of the two vessels reflects the highest acceptable navigation risk.

3. According to claim 2, a ship collision avoidance intelligent decision-making method based on an improved brainstorming optimization algorithm, Features: In open waters, if DCPA < 2 nautical miles, it is judged that the two ships will meet; in narrow waters, if DCPA < 0.5 nautical miles, it is judged that the two ships will meet.

4. A ship collision avoidance intelligent decision-making method based on an improved brainstorming optimization algorithm according to any one of claims 1 to 3, Features: The method further comprises: Step 4: The optimal collision avoidance steering angle and steering navigation time calculated in step 3 are integrated and displayed on the electronic nautical chart.

5. A ship collision avoidance intelligent decision-making method based on an improved brainstorming optimization algorithm according to any one of claims 1 to 3, Features: In step 1, the navigation status data of the ship includes latitude, longitude, heading, and speed.