Ship meeting collision avoidance method based on improved grey wolf algorithm
By improving the initialization and convergence factor optimization of the gray wolf algorithm, combined with the gold dig strategy, the fitness function is constructed, and the limitations of traditional ship collision avoidance methods under complex situations are solved, and a safe and economic collision avoidance path is generated, which reduces the collision risk.
Patent Information
- Application Number
- CN202510507409.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-22
- Publication Date
- 2025-07-22
AI Technical Summary
Traditional ship collision avoidance methods are difficult to take into account safety and economy when facing complex navigation situations. Especially when static obstacles and single ships encounter situations, there are problems such as uneven distribution of initial populations, slow convergence speed, and easy to fall into local optimality.
The population is initialized by Tent chaotic mapping, the convergence factor is modified to be a nonlinear form, and the exploration mechanism of the gold rush optimization algorithm is combined with the exploration mechanism of the gold rush optimization algorithm to optimize the position update mechanism of the α wolf, and the fitness function is constructed to optimize the collision avoidance path, and the minimum safe distance between the ship and the obstacle and collision avoidance steering angle are generated by improving the gray wolf algorithm.
It improves the algorithm's global search capability and convergence speed, generates a safe and economical collision avoidance path, reduces the collision risk of ships in complex seas, and improves the safety and economical navigation.
Smart Images
Figure CN120351935A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a ship encounter collision avoidance method based on an improved grey wolf algorithm, belonging to the fields of intelligent navigation and path planning. Background Art
[0002] With the continuous development and utilization of marine resources, the traffic flow of ships at sea is increasing day by day, and the collision risk faced by ships during navigation is also correspondingly increasing. Ship collision avoidance technology, as one of the key technologies to ensure the safety of maritime navigation, is of great importance. Traditional ship collision avoidance methods mainly rely on the experience and intuition of crew members, as well as some rule-based expert systems. However, these methods often have limitations when dealing with complex navigation situations, especially when facing static obstacles and single-ship encounter situations.
[0003] Static obstacles, such as reefs, fishing areas, and shallow waters, pose a serious threat to the navigation safety of ships. When planning a detour path, traditional collision avoidance methods often have difficulty taking into account both safety and economy, resulting in the ship needing to make a large detour, increasing the navigation time and fuel consumption. In addition, there are also many challenges in collision avoidance decisions in single-ship encounter situations. Crew members need to make accurate judgments and decisions in a short time to avoid collision accidents. However, human factors account for a large proportion in ship collision accidents. According to statistics, more than 80% of ship collision accidents are caused by human factors. This shows that the traditional collision avoidance method relying on the experience and intuition of crew members has great uncertainty and is difficult to meet the needs of modern maritime navigation safety.
[0004] Traditional grey wolf algorithms have problems such as uneven initial population distribution, slow convergence speed, and easy entrapment in local optima in ship static obstacle collision avoidance and single-ship encounter scenarios.
[0005] Therefore, it is of great practical significance to study a ship collision avoidance method that can effectively handle static obstacles and single-ship encounter situations. Summary of the Invention
[0006] Aiming at the problem that the existing collision avoidance methods cannot take into account both safety and economy when planning paths in the face of complex navigation situations, the present invention provides a ship encounter collision avoidance method based on an improved grey wolf algorithm.
[0007] A ship encounter collision avoidance method based on an improved grey wolf algorithm of the present invention includes:
[0008] Using Tent chaotic mapping for the traditional grey wolf algorithm to generate uniformly distributed initial path turning points, modifying the convergence factor into a non-linear form, and optimizing the position update mechanism of the alpha wolf by combining the exploration mechanism of the gold panning optimization algorithm to obtain an improved grey wolf algorithm;
[0009] Construct a fitness function for the static obstacle scenario or a fitness function for the single-ship encounter scenario based on the improved grey wolf algorithm; obtain a collision avoidance path for the ship to maintain the minimum safe distance from the obstacle based on the fitness function for the static obstacle scenario; obtain the collision avoidance steering angle and steering point based on the fitness function for the single-ship encounter scenario.
[0010] According to the ship encounter collision avoidance method based on the improved grey wolf algorithm of the present invention, the iterative formula of the Tent chaotic mapping is:
[0011]
[0012] where k is the iteration number of the Tent chaotic mapping, and x k is the chaotic variable value of the k-th iteration, and the value range of x k is [0, 1]; initialize the position of the grey wolf population based on the iterative formula.
[0013] According to the ship encounter collision avoidance method based on the improved grey wolf algorithm of the present invention, the non-linear convergence factor is expressed as a:
[0014]
[0015] where t is the position update iteration number of the alpha wolf, and Max_it is the maximum iteration number.
[0016] According to the ship encounter collision avoidance method based on the improved grey wolf algorithm of the present invention, the position update mechanism of the alpha wolf is:
[0017]
[0018] where is the current target position of the ship, is the position of the alpha wolf, which is the position of the alpha wolf determined after the initialization of the grey wolf population position; is the weight coefficient matrix for controlling the search direction and intensity, is the distance vector between the ship and the alpha wolf, l1 is the position convergence factor, is the random vector one, is the perturbation vector, is the ship position, is the random vector two.
[0019] According to the ship encounter collision avoidance method based on the improved grey wolf algorithm of the present invention, the fitness function F1 in the static obstacle scenario is:
[0020]
[0021] where F distance is the fitness function of economy; F collisionIt is the safety fitness function. If a collision occurs on the planned route, it is 0; otherwise, it is 1.
[0022]
[0023] In the formula, n is the number of turning points, and (x i, y i ) is the horizontal coordinate of the i-th turning point.
[0024] Determine the collision avoidance path based on all the obtained turning points.
[0025] According to the ship encounter collision avoidance method based on the improved grey wolf algorithm of the present invention, the fitness function F2 in the single-ship encounter scenario is:
[0026] F2 = ω1f1 + ω2f2 + ω3f3, where ω1 + ω2 + ω3 = 1.
[0027] In the formula, f1 is the safety fitness function, f2 is the route economy fitness function, f3 is the fitness function based on collision avoidance rules, ω1 is the weight coefficient of f1, ω2 is the weight coefficient of f2, and ω3 is the weight coefficient of f3.
[0028] f1 = maxCRI(v0, c0, v t , c t , T, R),
[0029]
[0030] In the formula, CRI represents the collision risk degree, v0 is the ship's speed, c0 is the ship's course, v t is the target ship's speed, c t is the target ship's course, T is the target ship's true bearing, and R is the relative speed between the ship and the target ship.
[0031] N is the number of turning points, and d I is the vertical distance from the I-th turning point to the original planned trajectory.
[0032] Obtain the collision avoidance turning angle based on the determined ship's course c0, and obtain the turning point by combining the turning point coordinates determined by d I .
[0033] The improved grey wolf algorithm combines the fitness function F2 to iteratively update the collision avoidance turning angle and the turning point. By comparing the values of the fitness function F2 in each iteration, gradually determine the optimal collision avoidance turning angle and turning point, and determine the best collision avoidance path.
[0034] According to the ship encounter collision avoidance method based on the improved grey wolf algorithm of the present invention, ω1, ω2, and ω3 are sequentially taken as 0.5, 0.25, and 0.25.
[0035] Advantages of the present invention: By designing the real-time reconstruction principle of chaotic mapping initialization, factor optimization, and gold panning strategy fusion, a multi-objective fitness function is constructed, and the implementation path planning of the ship in the scenarios of static obstacles and single-ship encounters is obtained based on the real-time reconstruction principle.
[0036] The method of the present invention improves the global search ability and convergence speed of the algorithm by improving the initialization method, convergence factor, and individual update method of the gray wolf optimization algorithm. For static obstacles and single-ship encounter situations, fitness functions are respectively established to evaluate the safety and economy of collision avoidance paths. The effectiveness of the method of the present invention is verified by simulation experiments, indicating that the method of the present invention can successfully avoid the target ship and output the optimal decision that complies with collision avoidance rules. The present invention provides an efficient and intelligent solution for ship collision avoidance in complex sea areas and has important practical application value.
[0037] The method of the present invention solves the limitations of the traditional gray wolf algorithm in complex collision avoidance scenarios. By initializing the population with chaotic mapping, it ensures the uniform distribution of the initial population in the search space and avoids local optima; by factor optimization, it dynamically adjusts the algorithm parameters and improves the convergence speed; by fusing the gold panning strategy, it enhances the local development ability of the algorithm and ensures the accuracy and reliability of the collision avoidance path. The method of the present invention can plan a safe and economical path in static obstacle collision avoidance and can quickly generate the optimal decision that complies with collision avoidance rules in single-ship encounter situations, significantly reducing the collision risk of ships in complex sea areas and improving the safety and economy of navigation. Description of the Drawings
[0038] Figure 1 is the flowchart of the ship encounter collision avoidance method based on the improved gray wolf algorithm of the present invention;
[0039] Figure 2 is the schematic diagram of the path planned by the method of the present invention in Scenario 1 of static obstacles; 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;
[0040] Figure 3 is the schematic diagram of the path planned by the method of the present invention in Scenario 2 of static obstacles;
[0041] Figure 4 is the schematic diagram of the path planned by the method of the present invention in Scenario 3 of static obstacles;
[0042] Figure 5 is the schematic diagram of the path planned by the method of the present invention in Scenario 4 of static obstacles;
[0043] Figure 6 is the comparison diagram of the path planning between the method of the present invention and the traditional gray wolf algorithm in the static obstacle scenario;
[0044] Figure 7 Yes Figure 6 Comparison chart of fitness change curves for path planning by two methods;
[0045] Figure 8 Collision avoidance path diagram for head-on situation obtained by using the method of the present invention in a single-ship encounter scenario; In the figure, the original path of the own ship coincides with that of the target ship;
[0046] Figure 9 Yes Figure 8 Fitness curve of the collision avoidance path shown;
[0047] Figure 10 Yes Figure 8 Schematic diagram of the distance between two ships under the collision avoidance path shown;
[0048] Figure 11 Collision avoidance schematic diagram for small-angle crossing encounter situation obtained by using the method of the present invention in a single-ship encounter scenario;
[0049] Figure 12 Yes Figure 11 Fitness curve of the collision avoidance path shown;
[0050] Figure 13 Yes Figure 11 Schematic diagram of the distance between two ships under the collision avoidance path shown;
[0051] Figure 14 Collision avoidance schematic diagram for large-angle crossing encounter situation obtained by using the method of the present invention in a single-ship encounter scenario;
[0052] Figure 15 Yes Figure 14 Fitness curve of the collision avoidance path shown;
[0053] Figure 16 Yes Figure 14 Schematic diagram of the distance between two ships under the collision avoidance path shown;
[0054] Figure 17 Collision avoidance path schematic diagram for overtaking situation obtained by using the method of the present invention in a single-ship encounter scenario in the overtaking situation of ships;
[0055] Figure 18 Yes Figure 17 Fitness curve of the collision avoidance path shown;
[0056] Figure 19 Yes Figure 17 Schematic diagram of the distance between two ships under the collision avoidance path shown. Detailed implementation method
[0057] Next, in combination with the accompanying drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.
[0058] 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.
[0059] The present invention will be further described below in conjunction with the accompanying drawings, but it is not a limitation of the present invention.
[0060] Combined Figure 1 As shown, the present invention provides a ship encounter collision avoidance method based on an improved grey wolf algorithm, including
[0061] Using Tent chaotic mapping to initialize the traditional grey wolf algorithm to generate uniformly distributed initial path turning points, modifying the convergence factor into a non-linear form, and optimizing the position update mechanism of the alpha wolf by combining the exploration mechanism of the gold panning optimization algorithm to obtain the improved grey wolf algorithm;
[0062] Based on the improved grey wolf algorithm, construct a fitness function in a static obstacle scenario or a fitness function in a single ship encounter scenario; obtain a collision avoidance path for the ship to maintain the minimum safe distance from the obstacle based on the fitness function in the static obstacle scenario; obtain a collision avoidance steering angle and turning point that comply with the international collision avoidance rules based on the fitness function in the single ship encounter scenario.
[0063] Furthermore, the iterative formula of Tent chaotic mapping is:
[0064]
[0065] where k is the iteration number of Tent chaotic mapping, x k is the chaotic variable value of the k-th iteration, and the value range of x k is [0, 1]; initialize the position of the grey wolf population based on the iterative formula.
[0066] The non-linear convergence factor is expressed as a:
[0067]
[0068] where t is the iteration number of the position update of the alpha wolf, Max_it is the maximum iteration number, and the default value is 100.
[0069] a is the non-linear convergence factor, which controls the search range of grey wolf individuals and dynamically decays with the iteration number.
[0070] The position update mechanism of the alpha wolf is as follows:
[0071]
[0072]
[0073] In the formula is the current target position of the ship, is the position of the alpha wolf, which is the position of the alpha wolf determined after the initialization of the gray wolf population position; is the weight coefficient matrix for controlling the search direction and intensity, is the distance vector between the 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 ship, is the random vector two. The modulus of takes a random number between [0,1];
[0074] Furthermore, the fitness function F1 in the static obstacle scenario is:
[0075]
[0076] In the formula F distance is the fitness function of economy; F collision is the safety fitness function, which is 0 if a collision occurs on the planned route and 1 otherwise;
[0077]
[0078] In the formula, n is the number of turning points, (x i, y i ) is the horizontal plane coordinate of the i-th turning point;
[0079] Determine the collision avoidance path based on all the obtained turning points.
[0080] Furthermore, the fitness function F2 in the single-ship encounter scenario is:
[0081] F2 = ω1f1 + ω2f2 + ω3f3, ω1 + ω2 + ω3 = 1,
[0082] In the formula, f1 is the safety fitness function, f2 is the route economy fitness function, f3 is the fitness function based on collision avoidance rules, ω1 is the weight coefficient of f1, ω2 is the weight coefficient of f2, and ω3 is the weight coefficient of f3;
[0083] f1 = maxCRI(v0, c0, v t , c t , T, R),
[0084]
[0085]
[0086] where CRI represents the collision risk degree, v0 is the ship's speed, c0 is the ship's course, v t is the target ship's speed, c t is the target ship's course, T is the target ship's true bearing, and R is the relative speed between the ship and the target ship;
[0087] N is the number of turning points, and d I is the perpendicular distance of the I-th turning point from the original planned trajectory;
[0088] Based on the determined ship's course c0, the collision avoidance steering angle is obtained, and the turning points are obtained by combining the turning point coordinates determined by d I ;
[0089] The improved grey wolf algorithm combines with the fitness function F2 to iteratively update the collision avoidance steering angle and turning points. By comparing the values of the fitness function F2 in each iteration, the optimal collision avoidance steering angle and turning points are gradually determined, and the best collision avoidance path is determined.
[0090] As an example, ω1, ω2, and ω3 are successively taken as 0.5, 0.25, and 0.25.
[0091] Verification experiment: Figures 2 to 5 Shows the simulation experimental results of ship static collision avoidance path planning based on the improved grey wolf optimization algorithm. In the experiment, this ship starts from the starting point and successfully plans a safe path to avoid obstacles under different obstacle layouts through the improved grey wolf algorithm and reaches the end point. The results show that the improved algorithm of the present invention can effectively avoid static obstacles and generate a shorter path at the same time. The path lengths are 13.2675, 13.0775, 13.0613, and 13.3017 nautical miles (nmile) respectively, which reflects the high efficiency and adaptability of the algorithm in complex environments. Through the improvement of initializing the population by Tent mapping and the nonlinear convergence factor, the global search ability and convergence speed of the algorithm are significantly improved, verifying its superiority in static collision avoidance path planning.
[0092] Figure 6 Compares the path planning effects of the original grey wolf algorithm and the improved grey wolf algorithm. The results show that the path length planned by the improved algorithm is 13.1879 n mile, which is shorter than 13.6018 n mile of the original algorithm, reflecting the high efficiency of the improved algorithm in path optimization. Figure 7The fitness change curves of the two algorithms were further compared. The improved algorithm converged stably in less than 50 generations, while the original algorithm took about 120 generations to reach a stable state, indicating that the improved algorithm has a significant advantage in the convergence speed. The improved grey wolf algorithm has obvious improvements in both the convergence speed and the optimization effect of the algorithm.
[0093] Figures 8 to 10 The simulation experiment results of ship encounter situation collision avoidance based on the improved grey wolf optimization algorithm are shown. In the experiment, the speed of the target ship is 16 Kn, the course is 225°, the relative bearing is 0°, and the distance from the own ship is 10.3 n mile. Figure 8 It shows that the collision avoidance path planned by the improved algorithm conforms to the "give-way" principle required by the collision avoidance rules, and it is recommended that the own ship take a large right turn measure to avoid the target ship. Figure 9 The fitness curve of... indicates that the algorithm tends to be stable at about 75 generations and successfully finds a better solution. Figure 10 It is further verified that the minimum distance between the two ships is always greater than 2 n mile, meeting the safety requirements of the collision avoidance rules.
[0094] Figures 11 to 16 The simulation experiment results of ship collision avoidance in small-angle and large-angle crossing situations are shown. In the crossing situation, there are two cases. When the target ship is in area B, it is a small-angle crossing encounter. According to the collision avoidance rules, the own ship should take a right turn strategy. When the target ship is in area C, it is a large-angle crossing encounter. According to the collision avoidance rules, the own ship should take a left turn strategy until the own ship has passed the target ship, and then take the measure of resuming the original course. In the small-angle crossing situation collision avoidance experiment, the speed of the target ship is 16 Kn, the course is 270°, the relative bearing is 21°, and the distance from the own ship is 9.8 n mile. Figure 11 It shows that the collision avoidance path planned by the improved algorithm conforms to the collision avoidance rules, and it is recommended that the own ship take a right turn measure to avoid the target ship. Figure 12 The fitness curve of... indicates that the algorithm tends to be stable at about 75 generations and successfully finds a better solution. Figure 13 It is further verified that the minimum distance between the two ships is always greater than the safe passing distance of 2 n mile. In the large-angle crossing situation collision avoidance experiment, the speed of the target ship is 10 Kn, the course is 0°, the relative bearing is 90°, and the distance from the own ship is 2.8 n mile. Figure 14 It shows that the path planned by the improved algorithm conforms to the collision avoidance rules, and it is recommended that the own ship turn left to avoid the target ship. Figure 15 The fitness curve of... indicates that the algorithm tends to be stable at about 70 generations. Figure 16 It is verified that the minimum distance between the two ships is always greater than 2 n mile, ensuring the safety of the collision avoidance process. The improved grey wolf algorithm optimizes the steering angle and path length, effectively reducing the collision risk in small / large-angle crossing situations and reducing the voyage loss at the same time.
[0095] Combined with Figures 17 to 19 In the simulation experiment of the ship overtaking situation, the target ship approaches the own ship at a speed of 5.0 knots and a course of 45°, with an initial relative bearing of 0° and a distance of 4.2 nautical miles. The collision avoidance path planned by the improved grey wolf optimization algorithm shows that the own ship takes a right turn measure in line with the rules and successfully avoids the target ship. The fitness curve indicates that the algorithm stabilizes after about 75 iterations and quickly finds the optimal solution. The curve of the distance change between the two ships further verifies that the minimum safe distance during the collision avoidance process always remains above 2 nautical miles, ensuring the navigation safety.
[0096] So far, using the ship encounter collision avoidance method based on the improved grey wolf algorithm, considering multiple static obstacles and the situation of large and small angle crossings, etc., by optimizing the steering angle and path length, the voyage loss has been significantly reduced, ensuring the safety of the collision avoidance operation.
[0097] Although the present invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed, as long as they do not depart from the spirit and scope of the present invention as defined by the appended claims. It should be understood that different dependent claims and the features described herein can be combined in a manner different from that described in the original claims. It should also be understood that the features described in connection with a single embodiment can be used in other described embodiments.
Claims
1. A ship encounter collision avoidance method based on an improved grey wolf algorithm, characterized in that Including, using Tent chaotic mapping for the traditional grey wolf algorithm to generate uniformly distributed initial path turning points, modifying the convergence factor into a non-linear form, and combining the exploration mechanism of the gold panning optimization algorithm to optimize the position update mechanism of the alpha wolf to obtain an improved grey wolf algorithm; Constructing a fitness function in a static obstacle scenario or a fitness function in a single-ship encounter scenario based on the improved grey wolf algorithm; obtaining a collision avoidance path for the ship to maintain the minimum safe distance from the obstacle based on the fitness function in the static obstacle scenario; obtaining the collision avoidance steering angle and turning points based on the fitness function in the single-ship encounter scenario.
2. The ship encounter collision avoidance method based on the improved grey wolf algorithm according to claim 1, wherein The iterative formula of Tent chaotic mapping is: where k is the number of iterations of the Tent chaotic map, and x k is the value of the chaotic variable at the k-th iteration, and the value range of x k is [0, 1]; Initializing the position of the grey wolf population based on the iterative formula.
3. The ship encounter collision avoidance method based on the improved grey wolf algorithm according to claim 2, wherein The non-linear convergence factor is expressed as a: where t is the position update iteration number of the alpha wolf, and Max_it is the maximum iteration number.
4. The ship encounter collision avoidance method based on the improved grey wolf algorithm according to claim 3, wherein The position update mechanism of the alpha wolf is: In the formula is the current target position of the ship is the position of the alpha wolf, which is determined after the initial position of the gray wolf population is initialized 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 first random vector is the perturbation vector is the position of the ship is the second random vector 5. The ship encounter collision avoidance method based on the improved grey wolf algorithm according to claim 4, wherein The fitness function F1 in the static obstacle scenario is: where F distance is the fitness function for economy; F collision is the safety fitness function, which is 0 if a collision occurs on the planned route and 1 otherwise; where n is the number of turning points, and (x i, y i ) is the horizontal coordinate of the i-th turning point; Determining the collision avoidance path based on all the obtained turning points.
6. The ship encounter collision avoidance method based on the improved grey wolf algorithm according to claim 4, wherein The fitness function F2 in the single-ship encounter scenario is: F2 = ω1f1 + ω2f2 + ω3f3, ω1 + ω2 + ω3 = 1, where f1 is the safety fitness function, f2 is the route economy fitness function, f3 is the fitness function based on collision avoidance rules, ω1 is the weight coefficient of f1, ω2 is the weight coefficient of f2, and ω3 is the weight coefficient of f3; f1 = maxCRI(v0, c0, v t , c t , T, R), where CRI represents the collision risk degree, v0 is the ship speed, c0 is the ship course, v t is the target ship speed, c t is the target ship course, T is the true bearing of the target ship, and R is the relative speed between the ship and the target ship; N is the number of turning points, and d I is the vertical distance from the i-th turning point to the original planned trajectory; Obtain a collision avoidance steering angle based on the determined ship course c0, and obtain a turning point in combination with the turning point coordinates determined by d I Determine the turning point; The improved grey wolf algorithm combines with the fitness function F2 to iteratively update the collision avoidance steering angle and turning points, and gradually determines the optimal collision avoidance steering angle and turning points by comparing the values of the fitness function F2 in each iteration, and determines the best collision avoidance path.
7. The ship encounter collision avoidance method based on the improved grey wolf algorithm according to claim 6, wherein ω1, ω2 and ω3 take values of 0.5, 0.25 and 0.25 in sequence.