Collision avoidance auxiliary decision-making method for multi-ship encounter situation
By improving the gray wolf algorithm and gray correlation analysis, the intelligent problem of ship collision avoidance decisions is solved, and the safe and efficient collision avoidance of multiple ships is achieved under the situation, which improves the safety and stability of ship navigation.
Patent Information
- Application Number
- CN202510507407.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-22
- Publication Date
- 2025-07-25
AI Technical Summary
The existing ship collision avoidance decision-making methods have not been intelligent and rely on the crew’s experience and subjective judgment, resulting in safety hazards in complex multi-target maritime environments.
The improved gray wolf algorithm is used to initialize the population through Tent mapping, modify the convergence factor to be a nonlinear cosine function, and combine the gold rush optimization idea to calculate the ship's collision avoidance objective function, and use gray correlation analysis to determine the collision risk order, and output the optimal steering angle and navigation time.
It has achieved intelligent collision avoidance decisions under the situation of multiple ships, improved navigation safety and stability, and reduced uncertainty caused by human factors.
Smart Images
Figure CN120370950A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a collision avoidance assistance decision-making method for multi-ship encounter situations, belonging to the field of ship navigation and collision avoidance decision-making. Background Art
[0002] Improving the collision avoidance decision-making ability is one of the key links in promoting the overall development of intelligent ships and has become an important bottleneck restricting the technological progress of intelligent ships. Therefore, exploring scientific and reliable collision avoidance decision-making methods has important practical significance. An efficient collision avoidance assistance technology can not only assist the crew in making accurate navigation decisions but also effectively reduce their work burden and improve the navigation safety and automation level.
[0003] The core of ship collision avoidance decision-making is to plan an optimal or relatively reasonable avoidance path for the ships in navigation. This path should not only ensure that the ships can safely avoid static and dynamic obstacles but also minimize the additional voyage distance and time caused by collision avoidance to balance the economy and efficiency of navigation. In short, the collision avoidance path should, on the basis of ensuring safety, minimize the interference with the established route and time plan. In recent years, high-precision navigation and perception devices such as AIS, ARPA, ECDIS, and GIS have played an important role in the ship collision avoidance process, capable of providing key situation awareness information and providing auxiliary judgment support for the crew when dealing with complex multi-ship encounters. However, these systems essentially still belong to the information provision level and have not achieved true intelligent collision avoidance decision-making. In current actual operations, once a potential collision risk appears, the avoidance decision still relies on the experience and subjective judgment of the crew. Although this method has the ability to respond quickly in some scenarios, it is also prone to safety hazards due to human negligence or misjudgment.
[0004] Therefore, the existing collision avoidance assistance mechanisms still have obvious deficiencies when facing complex, dynamic, and multi-target marine environments, and there is an urgent need to build a more intelligent and reliable automatic collision avoidance decision-making system to reduce the uncertainty brought by human factors and improve the overall safety of ship navigation. Summary of the Invention
[0005] Aiming at the problem that the existing ship collision avoidance methods cannot achieve intelligent collision avoidance decision-making, the present invention provides a collision avoidance assistance decision-making method for multi-ship encounter situations.
[0006] A collision avoidance assistance decision-making method for multi-ship encounter situations according to the present invention includes:
[0007] Initializing the population of the traditional grey wolf algorithm by using the Tent mapping method, and at the same time modifying the convergence factor a of the grey wolf algorithm into a non-linear decreasing cosine function form, and improving the method of calculating the current target position of the ship by the α wolf by using the gold panning optimization idea to obtain the improved grey wolf algorithm;
[0008] Obtain the position information and motion state information of the own ship and all collision avoidance target ships, and use the collision avoidance geometric method to calculate the important parameters affecting the collision risk; establish a ship collision avoidance objective function, and obtain the collision risk order of all collision avoidance target ships through grey relational analysis; based on the collision risk order, adopt the improved grey wolf algorithm to obtain the optimal steering angle, navigation time and resuming navigation angle of the own ship relative to the current collision avoidance target ship.
[0009] According to the collision avoidance auxiliary decision-making method for multi-ship encounter situations of the present invention, the method for calculating the convergence factor a of the grey wolf algorithm is:
[0010]
[0011] where t represents the current iteration number, and Max_it represents the maximum iteration number.
[0012] According to the collision avoidance auxiliary decision-making method for multi-ship encounter situations of the present invention, the method for improving the calculation of the current target position of the own ship by the alpha wolf using the gold panning optimization idea is specifically:
[0013] Adopt the migration formula of gold miners to gold mines in the gold panning optimization algorithm to improve the method for calculating the current target position of the own ship by the alpha wolf.
[0014] According to the collision avoidance auxiliary decision-making method for multi-ship encounter situations of the present invention, the current target position of the own ship is expressed as
[0015]
[0016] where is the position of the alpha wolf, is the weight coefficient matrix for controlling the search direction and intensity, is the distance vector between the own ship and the alpha wolf, l1 is the position convergence factor, is the random vector one, is the perturbation vector, is the position of the own ship, is the random vector two.
[0017] According to the collision avoidance auxiliary decision-making method for multi-ship encounter situations of the present invention, the important parameters affecting the collision risk include the closest distance of approach, the shortest time of encounter and the safe distance of encounter.
[0018] According to the collision avoidance auxiliary decision-making method for multi-ship encounter situations of the present invention, the ship collision avoidance objective function is expressed as F:
[0019]
[0020] where f1 is the economic fitness function and f2 is the safety fitness function;
[0021]
[0022] where r is the radius of the inscribed circle of the turning angle of the planned collision avoidance path, c1 is the corresponding central angle of the inscribed circle, and θ is the optimal turning angle. is the resumption angle, a θ is the turning angle weight coefficient, n is the number of collision avoidance target ships, and R i is the grey correlation degree between the ship's own turning to the optimal turning angle θ and the i-th collision avoidance target ship, and CRI θi is the collision risk degree between the ship's own turning to the optimal turning angle θ and the i-th collision avoidance target ship. is the resumption angle weight coefficient, k i is the ship's own turning resumption angle and the grey correlation degree between the ship's own turning to the resumption angle and the i-th collision avoidance target ship. is the collision risk degree between the ship's own turning to the resumption angle
[0023] According to the collision avoidance assistance decision-making method for multi-ship encounter situations of the present invention, the criterion for grey correlation analysis is:
[0024] Set the ideal target situation of the ship's own as Y0:
[0025] Y0 = (y0(1), y0(2),... y0(N)),
[0026] where y0(N) is the N-th target situation point in the ideal target situation;
[0027] Express the current situations of n collision avoidance target ships as:
[0028]
[0029] where y n (N) is the N-th current situation point of the n-th collision avoidance target ship;
[0030] Calculate the correlation degree γ(Y0, Y i ): i ) of Y0 and Y
[0031]
[0032] where γ(y0(m), y i (m)) is the correlation coefficient between Y0 and Y i at the m-th point;
[0033] If the correlation degree γ(Y0, Y i)If it satisfies the four axioms of grey correlation, then γ(Y0,Y i ) is called the grey correlation degree;
[0034] Sort all the grey correlation degrees from largest to smallest to obtain the collision risk order of n collision avoidance target ships.
[0035] According to the collision avoidance auxiliary decision-making method for multi-ship encounter situations of the present invention, the four axioms of grey correlation include:
[0036] Normativity:
[0037] Integrality:
[0038] For Y I , Y J ∈{Y s |s = 0, 1, 2, …, n}, there is γ(Y I , Y J ) ≠ γ(Y J , Y I ), I ≠ J;
[0039] Even-pair symmetry:
[0040] For Y I , Y J ∈Y, Y is the set of Y I and Y J :
[0041]
[0042] Proximity: The smaller |y0(m) - y i (m)| is, the larger γ(y0(m), y i (m)) is.
[0043] The beneficial effects of the present invention: The method of the present invention initializes the population in the way of Tent mapping, which can ensure that the initial population is evenly and diversely distributed in the search space, thus helping the grey wolf optimization algorithm to effectively jump out of the local optimal solution during the search process and approaching the global optimum faster. The introduction of Tent mapping can also improve the stability of the algorithm, enabling it to show good performance when facing different problems. Replacing the linear decrease of algebra with the non-linear decrease of the convergence factor in the form of a cosine function helps the algorithm maintain a high exploration ability in the initial stage of the search, effectively balance the global search and the local search, and also improve the convergence speed and accuracy of the algorithm.
[0044] The method of the present invention considers the encounter and collision avoidance problems of multiple ships during navigation, and realizes intelligent collision avoidance decision-making by using the improved grey wolf algorithm, ensuring the stability and safety of ships during the task execution process. Brief Description of the Drawings
[0045] Figure 1 is the flow chart of the collision avoidance assistance decision-making method for multi-ship encounter situations described in the present invention;
[0046] Figure 2 is the flow chart of the improved grey wolf algorithm;
[0047] Figure 3 is the flow chart of multi-ship collision avoidance decision-making;
[0048] Figure 4 is the schematic diagram of the three-ship encounter situation;
[0049] Figure 5 is the collision avoidance decision-making diagram for the three-ship encounter situation; in the figure, x / n mile represents the abscissa of the ship's position, with the unit of nautical mile, and y is the ordinate of the ship's position; yellow represents the own ship, and green represents other ships in the encounter environment;
[0050] Figure 6 is the curve graph of the change in the risk degree of the target ship in the three-ship encounter situation; in the figure, t / s is time / second;
[0051] Figure 7 is the curve graph of the change in the relative distance of the target ship in the three-ship encounter situation;
[0052] Figure 8 is the schematic diagram of the four-ship encounter situation;
[0053] Figure 9 is the collision avoidance decision-making diagram for the four-ship encounter situation;
[0054] Figure 10 is the curve graph of the change in the risk degree of the target ship in the four-ship encounter situation;
[0055] Figure 11 is the curve graph of the change in the relative distance of the target ship in the four-ship encounter situation;
[0056] Figure 12 is the schematic diagram of the five-ship encounter situation;
[0057] Figure 13 is the collision avoidance decision-making diagram for the five-ship encounter situation;
[0058] Figure 14 is the curve graph of the change in the risk degree of the target ship in the five-ship encounter situation;
[0059] Figure 15 is the curve graph of the change in the relative distance of the target ship in the five-ship encounter situation. Specific implementation manner
[0060] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.
[0061] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.
[0062] The present invention will be further described below in conjunction with the accompanying drawings, but it is not a limitation of the present invention.
[0063] Combined Figures 1 to 3 As shown, the present invention provides a collision avoidance assistance decision-making method for multi-ship encounter situations, including
[0064] Initializing the population of the traditional grey wolf algorithm by using Tent mapping, at the same time modifying the convergence factor a of the grey wolf algorithm into a non-linear decreasing cosine function form, and improving the method of calculating the current target position of the ship by the alpha wolf by using the gold panning optimization idea to obtain the improved grey wolf algorithm;
[0065] Obtaining the position information and motion state information of the ship itself and all collision avoidance target ships through the Automatic Identification System (AIS), calculating the important parameters affecting the collision risk by using collision avoidance geometric methods and combining the basic information of the two ships; outputting a practical collision avoidance strategy through the improved and optimized grey wolf algorithm: including establishing a ship collision avoidance objective function, obtaining the collision risk order of all collision avoidance target ships through grey relational analysis; based on the collision risk order, using the improved grey wolf algorithm to obtain the best steering angle, navigation time and resuming navigation angle of the ship relative to the current collision avoidance target ship.
[0066] Tent mapping utilizes its good randomness and ergodicity to ensure that the initial population is evenly and diversely distributed within the search space, thereby helping the grey wolf optimization algorithm to effectively jump out of the local optimal solution during the search process and approaching the global optimum faster. The introduction of Tent mapping can also improve the stability of the algorithm, enabling it to exhibit good performance when facing different problems.
[0067] Replacing the linear decrease of the algebraic form with the non-linear decrease of the convergence factor a in the traditional grey wolf algorithm in the form of a cosine function helps the algorithm maintain a high exploration ability in the initial stage of the search, effectively balance global search and local search, and also improve the convergence speed and accuracy of the algorithm.
[0068] Furthermore, the method for calculating the convergence factor a of the grey wolf algorithm is:
[0069]
[0070] Where \(t\) represents the current iteration number and \(Max\_it\) represents the maximum number of iterations.
[0071] In the traditional Grey Wolf Algorithm It linearly decreases to 0 as the current iteration number increases.
[0072] Compared with the linear decrease, the improved function of the convergence factor \(a\) makes the convergence factor \(a\) decrease slowly in the initial stage, which helps the algorithm maintain a high exploration ability in the initial search to find potential optimal solutions globally and avoid falling into local optimal solutions prematurely. As the algorithm enters the later stage with the progress of iterations, the faster decrease speed of the convergence factor \(a\) helps to quickly narrow the search range, enabling individuals to conduct fine searches within the local range and improving the solution accuracy of the algorithm.
[0073] The method of improving the calculation of the current target position of the own ship by the \(\alpha\) wolf using the gold panning optimization idea is specifically as follows:
[0074] Adopt the migration formula of gold prospectors towards gold mines in the gold panning optimization algorithm to improve the method of calculating the current target position of the own ship by the \(\alpha\) wolf.
[0075] Express the current target position of the own ship as
[0076]
[0077] Where is the position of the \(\alpha\) wolf, is the weight coefficient matrix for controlling the search direction and intensity, is the distance vector between the own ship and the \(\alpha\) wolf, \(l1\) is the position convergence factor, is the random vector one, is the perturbation vector, is the position of the own ship, is the random vector two.
[0078] In the traditional Grey Wolf pyramid, the \(\alpha\) wolf is at the top of the group. It is not only the most decision-making individual but also responsible for coordinating various important affairs in the wolf pack. Followed by the \(\beta\) wolves, as the capable assistants of the \(\alpha\) wolf, they assist the leader in making decisions and take over the leadership position when necessary. The \(\delta\) wolves are in the middle layer of the hierarchical system. They execute the orders of the upper layer and are responsible for tasks such as reconnaissance and sentry duty. And the \(\omega\) wolves at the bottom of the pyramid, although with a lower status, play an indispensable role in maintaining the balance of the internal relationship of the population and performing basic work. In the mathematical description of the improved Grey Wolf Algorithm, The modulus of takes a random number between \([0,1]\).
[0079] The important parameters affecting the collision risk include the closest distance of approach, the shortest time of approach, and the safe distance of approach.
[0080] Going further, the ship collision avoidance objective function is expressed as F:
[0081]
[0082] Where f1 is the economic fitness function, and f2 is the safety fitness function;
[0083]
[0084] Where r is the radius of the inscribed circle of the planned collision avoidance path, c1 is the corresponding centripetal angle of the inscribed circle, θ is the optimal steering angle, is the return angle, a θ is the steering angle weight coefficient, n is the number of target ships to avoid collision, R i CRI is the grey correlation between the ship and the i-th collision avoidance target ship after the ship turns to the optimal turning angle θ. θi is the collision risk with the i-th collision avoidance target ship after the ship turns to the optimal steering angle θ, is the weight coefficient of the re-travel angle, k i The ship's turning angle The grey correlation degree between the latter and the i-th collision avoidance target ship is The ship turns back to the original angle The collision risk with the i-th collision avoidance target ship is calculated.
[0085] When the collision risk reaches 0.5, it means that there is an obvious collision risk. At this time, appropriate collision avoidance measures must be taken. Then, the optimized gray wolf algorithm can output a feasible collision avoidance strategy. θ and is the weight coefficient, n represents the number of target ships; in order to meet the requirements of the International Regulations for Preventing Collisions at Sea, the avoiding vessel should take "substantial action" to ensure that other ships can clearly observe and take corresponding actions. Therefore, the value of the avoidance angle θ is set between [30°, 80°]; in order to ensure that the ship has passed the most dangerous moment without taking collision avoidance measures at the beginning of the re-entry, the avoidance navigation time is set between [max (TCPA), 400], that is, between the maximum value of the shortest encounter time TCPA of all target ships and 400min. For the re-entry angle It can be set between [30°,60°].
[0086] In this implementation, the criteria for grey relational analysis are:
[0087] Grey relational analysis is a method for comprehensively evaluating the overall system. It depends on the similarity degree between the geometric curves formed by data sequences for analysis and comparison. When the shapes of these geometric curves are more similar, it indicates that the development trends they represent are closer, and further implies that the correlation between them is stronger. Even when the amount of data is small or the data quality is not high, grey relational analysis can still work effectively, helping to solve the uncertainty problem of influencing factors among multiple indicators and being an effective supplement to traditional statistical models.
[0088] Set the ideal target situation of the own ship as Y0:
[0089] Y0 = (y0(1), y0(2),... y0(N)),
[0090] where y0(N) is the Nth target situation point in the ideal target situation;
[0091] Express the current situations of n collision avoidance target ships as:
[0092]
[0093] where y n (N) is the Nth current situation point of the nth collision avoidance target ship;
[0094] Calculate the correlation degree γ(Y0, Y i ) of: i ):
[0095]
[0096] where γ(y0(m), y i (m)) is the correlation coefficient of Y0 and Y i at point m; γ(y0(m), y i (m)) is a real number;
[0097] If the correlation degree γ(Y0, Y i ) satisfies the four axioms of grey correlation, then γ(Y0, Y i ) is called the grey correlation degree;
[0098] Sort all the grey correlation degrees from largest to smallest to obtain the collision risk order of n collision avoidance target ships.
[0099] The four axioms of grey correlation include:
[0100] Normativity:
[0101] Integrality:
[0102] For Y I , Y J ∈{Ys |s = 0, 1, 2, …, n}, there is γ(Y I , Y J ), ≠ γ(Y J , Y I ), I ≠ J;
[0103] Pair symmetry:
[0104] For Y I , Y J ∈ Y, Y is the set of Y I and Y J :
[0105]
[0106] Proximity: |y0(m) - y i (m)| The smaller, γ(y0(m), y i (m)) The larger.
[0107] If the reference sequence and the sequence to be compared are highly similar, that is, the ideal target situation of the ship itself and the current situation of a certain collision avoidance target ship are highly similar, the value of the grey correlation degree will be very large, then the best sequence can be screened out, and then the advantages and disadvantages of each evaluation object can be compared and sorted.
[0108] Example:
[0109] Combined with Figure 2 shown, the process of the improved grey wolf algorithm includes:
[0110] Step 1, Initialization stage: Initialize the positions of the grey wolf population, the convergence factor a, and the coefficient matrices A, C;
[0111] Step 2, Fitness evaluation: Calculate the fitness value of each grey wolf, and record the top three grey wolves with the current best fitness value, denoted as α, β, δ respectively.
[0112] Step 3, Iterative update: Update the parameters a, A, C. According to the positions of α, β, δ, guide the position update of other grey wolves. Calculate the position vectors of each grey wolf and update its current position. Recalculate the fitness values of all grey wolves. Update the positions and fitness of α, β, δ according to the new fitness values.
[0113] Step 4, Termination judgment whether the preset maximum number of iterations is reached: If not, continue to iterate; if so, the algorithm terminates and outputs the current optimal solution.
[0114] Combined with Figure 3 shown, the multi - ship collision avoidance decision - making process is specifically as follows:
[0115] Step 1, Obtain information: Obtain relevant information of the own ship and the target ship, including parameters such as speed, course, and position.
[0116] Step 2, Calculate key parameters: Calculate indicators related to collision avoidance, including the distance of closest point of approach (DCPA), the time to closest point of approach (TCPA), relative position, relative speed, ship speed ratio, and the collision risk index (CRI), etc.
[0117] Step 3, Preliminary judgment of collision risk: Judge whether the CRI value is greater than 0.5: If so, it indicates a relatively high collision risk and further collision avoidance assessment is required; if not, return to the information update stage and continue monitoring.
[0118] Step 4, Judgment of the intention of rule - based avoidance: Judge whether the target ship has the responsibility to avoid according to the International Regulations for Preventing Collisions at Sea and identify its avoidance intention.
[0119] Step 5, Judgment of whether there is an avoidance intention: If the target ship does not show an avoidance behavior, enter the next risk analysis; if it has shown, continue to observe its actions.
[0120] Step 6, Risk level assessment: Evaluate the collision risk level with the target ship through the grey relational analysis method.
[0121] Step 7, Decision optimization and generation of avoidance actions: Activate the decision optimization module to calculate operation parameters such as the optimal course angle, navigation time, and speed change amplitude of the own ship in the avoidance situation.
[0122] Step 8, Verification of collision avoidance effect: Judge whether the DCPA after collision avoidance is greater than the set safe distance threshold (SDA): If the condition is met, the collision avoidance is successful; if not, re - perform the avoidance decision optimization.
[0123] Step 9, End: When the collision avoidance safety requirements are met, complete the auxiliary decision - making process.
[0124] Combined with Figures 4 to 15 , consider the problem of collision avoidance auxiliary decision - making in the situation of multi - ship encounter.
[0125] In the figure, OS (Own Ship) represents the own ship, and TS (Target Ship) represents other ships in the encounter environment. The decision - making process adopted by the own ship can keep the risk level between the own ship and the target ship below 0.5 all the time; from the relative distance change curve between the own ship and the target ship, it can be obtained that the relative distance between the own ship and the target ship is always greater than the safe encounter distance during this process, ensuring the safety of the collision avoidance process.
[0126] So far, using the collision avoidance auxiliary decision - making method for the situation of multi - ship encounter, considering the collision avoidance problems of multiple ships during navigation, ensures the stability and safety of ships during the task execution process.
[0127] Although the present invention has been described herein with reference to particular embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. Accordingly, it should be understood that numerous modifications may be made to the exemplary embodiments, and other arrangements may be devised, without departing from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the different dependent claims and the features described herein may be combined in ways different from those described in the original claims. It should also be understood that the features described in connection with separate embodiments may be used in other described embodiments.
Claims
1. A collision avoidance assistance decision-making method for multi-ship encounter situations, characterized in that including, Initializing the population of the traditional grey wolf algorithm by using Tent mapping, modifying the convergence factor a of the grey wolf algorithm into the form of a non-linear decreasing cosine function, and improving the method of calculating the current target position of the ship by the alpha wolf by using the gold panning optimization idea to obtain the improved grey wolf algorithm; Obtaining the position information and motion state information of the ship itself and all collision avoidance target ships, calculating the important parameters affecting the collision risk by using the collision avoidance geometry method; establishing a ship collision avoidance objective function, and obtaining the collision risk order of all collision avoidance target ships through grey relational analysis; based on the collision risk order, using the improved grey wolf algorithm to obtain the optimal steering angle, navigation time and resumption angle of the ship relative to the current collision avoidance target ship.
2. The collision avoidance auxiliary decision-making method for multi-ship encounter situations according to claim 1, characterized in that The method for calculating the convergence factor a of the grey wolf algorithm is: where t represents the current iteration number and Max_it represents the maximum iteration number.
3. The collision avoidance auxiliary decision-making method for multi-ship encounter situations according to claim 2, characterized in that The method of improving the method of calculating the current target position of the ship by the alpha wolf by using the gold panning optimization idea is specifically: Using the migration formula of gold miners to gold mines in the gold panning optimization algorithm to improve the method of calculating the current target position of the ship by the alpha wolf.
4. The collision avoidance auxiliary decision-making method for multi-ship encounter situations according to claim 3, characterized in that Represent the current target position of this vessel as In the formula is the position of the alpha wolf, is the weight coefficient matrix for controlling the search direction and intensity, is the distance vector between the ship and the alpha wolf, and l1 is the position convergence factor, is the random vector one, is the perturbation vector, is the ship's position, is the random vector two.
5. The collision avoidance auxiliary decision-making method for multi-ship encounter situations according to claim 4, characterized in that The important parameters affecting the collision risk include the closest distance of approach, the shortest time of approach and the safe distance of approach.
6. The collision avoidance auxiliary decision-making method for multi-ship encounter situations according to claim 5, characterized in that The ship collision avoidance objective function is expressed as F: where f1 is the economic fitness function and f2 is the safety fitness function; where r is the radius of the inscribed circle of the turning angle of the planned collision avoidance path, c1 is the corresponding central angle of the inscribed circle, θ is the optimal turning angle, is the resumption angle, a θ is the turning angle weight coefficient, n is the number of collision avoidance target ships, R i is the grey correlation degree between the ship after turning to the optimal turning angle θ and the i-th collision avoidance target ship, CRI θi is the collision risk degree between the ship after turning to the optimal turning angle θ and the i-th collision avoidance target ship, is the resumption angle weight coefficient, k i is the resumption angle of the ship's turning and the grey correlation degree between the ship and the i-th collision avoidance target ship, is the resumption angle of the ship's turning and the collision risk degree between the ship and the i-th collision avoidance target ship.
7. The collision avoidance auxiliary decision-making method for multi-ship encounter situations according to claim 6, characterized in that The criteria for grey relational analysis are: Setting the ideal target situation of the ship itself as Y0: Y0 = (y0(1), y0(2),... y0(N)), where y0(N) is the Nth target situation point in the ideal target situation; Expressing the current situations of n collision avoidance target ships as: where y n (N) is the Nth current situation point of the nth collision avoidance target ship; Calculate the correlation degree γ(Y0, Y i ) between Y0 and Y i ): where γ(y0(m), y i (m)) is the correlation coefficient between Y0 and Y i at point m; If the correlation degree γ(Y0, Y i ) satisfies the four axioms of grey correlation, then γ(Y0, Y i ) is called the grey correlation degree; Sorting all grey relational degrees from largest to smallest to obtain the collision risk order of n collision avoidance target ships.
8. The collision avoidance assistance decision-making method for multi-ship encounter situations according to claim 7, wherein The four axioms of grey relational analysis include: Normativeness: Integrity: For Y I , Y J ∈ {Y s | s = 0, 1, 2, …, n}, there is γ(Y I , Y J ) ≠ γ(Y J , Y I ), I ≠ J; Even pair symmetry: For Y I , Y J ∈Y, where Y is the set of Y I and Y J : Proximity: The smaller |y0(m) - y i (m)| is, the larger γ(y0(m), y i (m)) is.