A method for multi-ship intelligent avoidance optimization in restricted waters considering route planning and a medium

Abstract

The application relates to a kind of limited water area multi-ship intelligent avoidance optimization method and medium considering fixed line system, which comprises the following steps: obtaining the environmental information in limited water area and the state information of each ship;Based on the environmental information and state information, and combined with fixed line system navigation mark, artificial potential field algorithm is used to construct navigation scene image;Under the navigation scene image, the collision risk of the ship and the safety distance of the ship are calculated;The collision risk of the ship and the safety distance of the ship are used as the collision avoidance condition, and the global and local combined path planning algorithm is used to plan the optimal navigation path of each ship, and the search algorithm is used to optimize the navigation state of each ship, and the multi-ship intelligent avoidance optimization process is completed.Compared with the prior art, the application has the advantages of improving multi-ship avoidance effect, ensuring the navigation safety of water area near the port and the like.

Description

Technical Field

[0001] This invention relates to the field of ship collision avoidance technology, and in particular to a method and medium for optimizing intelligent collision avoidance among multiple vessels in restricted waters, taking into account the route system. Background Technology

[0002] Intelligent collision avoidance is a challenging core issue in maritime traffic safety, especially in port access channels. There is an urgent need for technologies adapted to restricted waterways to ensure the safety and reliability of autonomous avoidance between encountering vessels. The traffic environment faced by ships navigating in port access channels is exceptionally complex, including not only natural obstacles such as land and islands, but also crisscrossing vessels. Although numerous researchers have conducted studies on intelligent collision avoidance methods in restricted waterways, they have not fully considered lane marking factors and the principles of "early, large, wide, and clear" collision avoidance. Therefore, developing multi-vehicle intelligent collision avoidance optimization technologies for lane marking in restricted waterways is a key issue for ensuring navigational safety and achieving risk management.

[0003] Intelligent collision avoidance among multiple vessels is a primary challenge for navigation safety in restricted waters. Intelligent collision avoidance manifests as both path planning and obstacle avoidance problems. Classic intelligent collision avoidance algorithms include A*, APF, VO, and DWA, which effectively address single collision avoidance problems such as static obstacle avoidance, dynamic obstacle avoidance, global planning, and local planning. However, in actual navigation, multiple problems are often encountered simultaneously, leading to the growing popularity of hybrid algorithms. For example, A* and DWA can be combined to solve both local and global path planning problems. With the inclusion of more factors such as multi-vehicle navigation, complex waters, and COLREGs, simple algorithm fusion is no longer sufficient to solve these complex problems. While collision avoidance in open waters primarily involves maneuvering, in restricted waters, not only maneuvering but also deceleration is crucial.

[0004] When facing collision risks in exceptionally complex waters, it is often difficult to ensure safety if only one vessel takes evasive action. Based on the principles of "early, large, wide, and clear," all vessels should be considered as capable of taking evasive action to ensure safe navigation in order to avoid imminent danger. Summary of the Invention

[0005] The purpose of this invention is to provide a method and medium for intelligent collision avoidance optimization of multiple vessels in restricted waters, taking into account the route system, to improve the safety of navigation of multiple vessels in restricted waters.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A method for intelligent collision avoidance optimization of multiple vessels in restricted waters considering route-fixed systems includes the following steps:

[0008] To obtain environmental information and the status information of each vessel within the restricted waters;

[0009] Based on the aforementioned environmental and status information, and in conjunction with the fixed-route navigation beacon, an artificial potential field algorithm is used to construct a navigation scenario profile.

[0010] Under the aforementioned navigation scenario profile, the risk of ship collision is calculated and the safe distance between ships is determined;

[0011] Based on the navigation scenario profile, the collision risk of the ships and the safe distance between the ships are used as collision avoidance conditions. A path planning algorithm combining global and local methods is used to plan the optimal navigation path for each ship. At the same time, a search algorithm is used to optimize the navigation status of each ship, thus completing the intelligent collision avoidance optimization process for multiple ships.

[0012] Furthermore, the steps for constructing the general aviation scenario profile include:

[0013] Based on the environmental information within the restricted water area, the contours of natural obstacles are extracted using the convex hull algorithm to obtain the natural obstacle region;

[0014] Based on the natural obstacle areas, navigational beacons, and status information, an artificial potential field algorithm is used to construct a navigation scenario profile.

[0015] Furthermore, the step of obtaining the natural obstacle region includes:

[0016] Let X be any subset of the Euclidean space W. The smallest convex set containing X is called the convex hull of X, denoted as conv(X) = {X1, X2, ..., X}. n}, where each convex hull corresponds to a natural obstacle object;

[0017] Connect the outermost points of each convex hull to form a convex polygon, which serves as the natural obstacle region.

[0018] Furthermore, the steps for constructing the general aviation scenario profile include:

[0019] Based on the repulsive force relationship between natural obstacles and navigational beacons in the natural obstacle area, a repulsive force field is constructed, wherein the expression of the repulsive force field is:

[0020]

[0021] K(t,t0)=||t-t0||2

[0022] In the formula, G k(t) is the repulsive potential field function, g is the proportionality coefficient, K(t,t0) is a vector, its direction is from the obstacle to the ship, and its magnitude is the Euclidean distance |t-t0| between the ship and the obstacle, where t0 is a constant representing the maximum influence range of the obstacle on the ship, and K λ It is a constant representing the maximum distance at which an obstacle exerts its effect on a ship;

[0023] Based on the gravitational relationship between each ship's own position and the target position, a gravitational field is constructed, wherein the expression for the gravitational field is:

[0024]

[0025] k(t,t i )=||tt i ||2

[0026] In the formula, G ω (t) is the gravitational field function, k(t,t) i () is a vector representing the ship's position t and the target point's position t. i Euclidean distance between |tt i |,k λ β is a constant, and β is the proportional gain coefficient;

[0027] The repulsive and gravitational fields are used as a model for general aviation scenarios.

[0028] Furthermore, the calculation process for the ship collision risk includes:

[0029] Based on the status information of each vessel, calculate the DCPA and TCPA parameters, where the expressions for the DCPA and TCPA parameters are as follows:

[0030]

[0031] In the formula, DCPA t For DCPA parameters, α rt Let t be the relative heading of the target ship, TCPA t For TCPA parameters, μ r Let μ1 be the velocity of the target ship relative to the ship at time t, μ2 be the velocity of the ship at time t, and D be the velocity of the target ship at time t. t Let β be the distance between the two ships at time t. Tt Let t be the phase of the target ship and the bearing of this ship.

[0032] Based on the DCPA and TCPA parameters, a negative exponential function is used to fit and calculate the ship collision risk, wherein the expression for the ship collision risk is:

[0033] CRI t=ω×RDCPA t +τ×RTCPA t

[0034]

[0035] In the formula, CRI t For the ship collision risk level, ω and τ are weighting coefficients, a and b are adjustment coefficients, and DCPA′ is the coefficient for DCPA. t The dimensionless DCPA is the result of the DCPA, TCPA′ is the dimensionless TCPA, and RDCPA is the result of the TCPA. t As a risk component of DCPA, RTCPA t It is a risk component of TCPA.

[0036] Furthermore, the safe distance of the vessel is defined using a quaternion-based vessel domain, and the safe distance is expressed as:

[0037]

[0038] in:

[0039]

[0040] In the formula, η k Let (x, y) be the boundary of the ship's domain, (x, y) be the ship's position, S be a quaternion, and D be the boundary of the ship's domain. fore D is the distance from the bow of the ship. aft D is the distance from the stern of the ship. starb D is the distance to the starboard side of the ship. portt ξ is the distance to the port side of the ship. AD ξ is the gain of the approach distance. DT The gain is the gyroscopic diameter, L is the ship length, and A is the gyroscopic diameter. D For the advance distance, D T V is the initial diameter of the cycle. own This is the speed of the ship.

[0041] Furthermore, the A* algorithm is used as the global path planning algorithm to plan the global shortest path, and the artificial potential field algorithm is used as the local path planning algorithm to plan the local path, so as to obtain the optimal navigation path for ships. During the A* algorithm path planning process, a value function is used to evaluate the value of each ship from the starting point S. i To the target point G i The total cost estimate is used to find the shortest path. The shortest path is obtained when f(k) > 0 and 0 < g(k) ≤ g*(k), where the expression for the value function is:

[0042] φ*(k)=f*(k)+g*(k)

[0043] In the formula, φ*(k) is the value function, f*(k) is the distance function from the starting point, g*(k) is the distance function to the target point, f(k) is the actual cost from the initial state to state k, and g(k) is the estimated cost of the optimal path from state k to the target state.

[0044] Furthermore, the steps of the local path planning include:

[0045] The artificial potential field algorithm applies a repulsive field and a gravitational field to the ship to form a resultant potential field, which guides the ship to move along a local path. The expression for the resultant potential field is as follows:

[0046] G(q)=G k (t)+G ω (t)

[0047] In the formula, G(q) is the artificial potential field, that is, the resultant potential field. k (t) represents the gravitational field, G ω (t) represents the repulsive force field.

[0048] Furthermore, the search algorithm is an artificial bee colony algorithm, and the steps for optimizing the navigation status of each vessel using the artificial bee colony algorithm include:

[0049] Initial Phase: Initialize all nectar sources using scout bees, where the location of the nectar sources represents the ship's navigation status, including heading and speed. The initialization expression is:

[0050] f nj =k j +rand(0,1)×(ω j -k j )

[0051] In the formula, f nj Let n be the nectar source, j be the population size, and k be the number of variables. j ω j f nj The upper and lower limits of the value;

[0052] Employment stage:

[0053] Foraging bees search for neighbors based on their memory of nectar source locations and calculate fitness to determine neighboring nectar sources:

[0054]

[0055] The fitness problem is solved by minimizing the following equation:

[0056]

[0057] In the formula, gnj For the neighbor's honey source, f tj For randomly selected nectar sources, These are random numbers used to calculate the fitness value of new nectar sources. Let be the objective function value of the nth nectar source;

[0058] Observation phase:

[0059] The observation bee randomly selects a nectar source based on the nectar source information obtained by the foraging bee, and generates new nectar sources near the selected source. The fitness of the new nectar source is calculated, and the probability of it being selected by the observation bee is:

[0060]

[0061] In the formula, ψ n This represents the probability that a nectar source is selected.

[0062] Reconnaissance phase:

[0063] If a foraging bee decides to abandon a nectar source after a given number of attempts, it becomes a scout bee and restarts the initial phase until it finds the optimal nectar source.

[0064] The present invention also provides a computer-readable storage medium including one or more programs executable by one or more processors of an electronic device, the one or more programs including instructions for performing the intelligent collision avoidance optimization method for multiple vessels in restricted waters considering the line-keeping system as described above.

[0065] Compared with the prior art, the present invention has the following beneficial effects:

[0066] (1) This invention considers environmental information and route information in restricted waters to construct a scenario profile of ship avoidance. Combining collision hazards and safe distances, it uses a path planning algorithm that combines global and local approaches to plan the navigation path of each ship. At the same time, it optimizes the navigation operation status of each ship according to a search algorithm, thereby achieving simultaneous optimization of the navigation path and navigation status of multiple ships and improving the safety of multiple ships navigating in restricted waters.

[0067] (2) This invention uses the convex hull algorithm to extract the contours of natural obstacles and considers navigational markers to construct a static obstacle map. The use of the artificial potential field method to process the static obstacle map into a potential field can obtain a static obstacle profile in restricted waters that includes multiple information such as natural obstacles and navigational markers, providing a scenario for subsequent steps.

[0068] (3) This invention uses the A* algorithm to obtain the global path planning effect and uses the APF for local planning to consider the safe distances of static obstacles and dynamic vessels. It can autonomously adjust the safety factor according to the magnitude of the perceived collision risk of the vessel and the size of the vessel's domain, providing personalized avoidance solutions for vessels.

[0069] (4) This invention introduces an artificial bee colony algorithm to optimize the avoidance path, avoidance speed, and turning radius, enabling turning and deceleration avoidance in restricted waters. It can fully consider the fixed-line system factors and scientifically allocate the turning radius and speed to achieve efficient and intelligent collision avoidance when multiple ships meet in restricted waters, ensuring the navigation safety of ships. Attached Figure Description

[0070] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0071] Figure 2 The image shown is an environment perception map under the convex hull algorithm of the present invention, wherein (a) represents local nautical map islands and reefs, (b) represents grayscale processing, (c) represents convex hull extraction contour, (d) represents local nautical map islands and reefs, (e) represents global nautical map appearance, (f) represents global nautical map overlay, (g) represents contour extraction, and (h) represents channel division.

[0072] Figure 3 The image depicts the waterway scene based on the artificial potential field method of the present invention, wherein (a) represents the repulsive field and (b) represents the gravitational field;

[0073] Figure 4 This invention relates to risk perception under certain circumstances at a given moment.

[0074] Figure 5 The following are the optimization results for avoiding confined waters in this invention: (a) Figure shows the multi-target avoidance result of A*+APF, (b) Figure shows the optimization result of artificial bee colony, (c) Figure shows the optimization result from a three-dimensional perspective, and (d) Figure shows the optimization result under nautical chart overlay.

[0075] Figure 6 For the evaluation of the multi-vehicle avoidance effect in restricted waters according to the present invention, (a) figure represents the vessel heading, (b) figure represents the vessel speed, (c) figure represents the distance between each pair of vessels, and (d) figure represents the risk between each pair of vessels. Detailed Implementation

[0076] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0077] This embodiment provides an intelligent collision avoidance optimization method for multiple vessels in restricted waters considering route-fixed systems, such as... Figure 1 As shown, the method includes the following steps:

[0078] Step 1: Perceiving Natural Obstacles. Based on the convex hull algorithm, natural obstacles on the nautical chart are identified. Utilizing the convex hull concept, the avoidance scenario is reconstructed using navigational aids, providing a scenario foundation for subsequent experiments.

[0079] (1) Natural obstacle contour extraction

[0080] Land and islands are common natural obstacles. To ensure safe navigation, the location and outline of these obstacles must be comprehensively considered. A convex hull is a geometric concept, typically defined mathematically as follows: Let X be any subset of Euclidean space W. The smallest convex set containing X is called the convex hull of X, denoted as conv(X) = {X1, X2, ..., Xn}. Therefore, it can be formed by connecting the outermost points to create a convex polygon that comprehensively contains all points in the point set, i.e., the natural obstacle region.

[0081] (2) Integration of the outlines of navigational beacons and natural obstacles

[0082] The environment of restricted waters is extremely complex, with significant ship traffic. The implementation of route-fixed navigation systems has greatly reduced the occurrence of ship collisions in restricted waters. Therefore, route-fixed navigation marks are another key factor to consider in intelligent collision avoidance scenarios for ships in restricted waters. Furthermore, considering the relatively fixed routes in restricted waters, the original subset of natural obstacles and route-fixed navigation marks are set as new points. Again, using the convex hull concept, each point is connected to obtain the static obstacle environment of the restricted waters.

[0083] Step 2: Construct a navigation scene profile. Given that vessels need to maintain a sufficient safe distance from static obstacles, an artificial potential field is introduced to profile the identified scene, enabling personalized avoidance strategies and providing safe distance limits in subsequent obstacle avoidance operations.

[0084] A method for profiling general aviation scenarios based on artificial potential field algorithms:

[0085] In intelligent collision avoidance, static obstacles constructed by environmental factors can be considered to have a repulsive force field with the ship, which can be expressed by mathematical formula (1):

[0086] Commonly used repulsive potential field functions:

[0087]

[0088] Where g is a proportionality coefficient, K(t,t0) is a vector with a direction from the obstacle to the ship and a magnitude of the Euclidean distance |t-t0| between the ship and the obstacle, and t0 is a constant representing the maximum range of influence of the obstacle on the ship.

[0089] It is assumed that there is a gravitational field between the target position and the current position, which can be described by mathematical formula (3):

[0090]

[0091] Where β is the proportional gain coefficient, and k is a vector representing the ship's position t and the target point's position t. n Euclidean distance between |tt n |, where the vector direction is from the ship's position to the target point's position.

[0092] k(t,t i )=||tt i ||2 (4)

[0093] Step 3: Perceive Ship Risks. Risk perception involves identifying ship collision hazards and the ship's surroundings, and determining the priority of ship avoidance and the safe distance for avoidance.

[0094] This step considers the avoidance obligation and the avoidance mode of dangerous situations, that is, it takes into account the avoidance principle of "early, large, wide, and clear" in the context of the shipping industry. Based on the risk perception results, a multi-vessel avoidance mechanism suitable for restricted waters is established, providing a basis for subsequent intelligent algorithms to implement avoidance. Risk perception, also known as situational awareness, is a prerequisite for intelligent collision avoidance. Ship encounter situational awareness determines whether avoidance is necessary and the priority of avoidance by identifying the danger of the vessels currently forming a dangerous situation and the target vessel.

[0095] (1) Calculation of ship collision risk

[0096] DCPA (Distance to Closest Point of Approach) and TCPA (Time to Closest Point of Approach) have always been key factors in determining whether there is a risk of collision between two ships. Their calculation formulas are as follows:

[0097]

[0098] In the formula, μ1 is the speed of the current ship at time t, μ2 is the speed of the target ship at time t, and μ rt Let α be the velocity of the target ship relative to the ship at time t. rt μ at time t r The direction.

[0099] The Collison Risk Index (CRI) is a standard for measuring the risk of collision with an encountering vessel. This invention uses a negative exponential function to fit and calculate the collision risk, as shown in the following formula:

[0100] CRI t =ω×RDCPAt +τ×RTCPA t (8)

[0101]

[0102] In the formula, a and b are adjustment coefficients, which are generally set according to the characteristics of marine applications; ω and τ are weighting coefficients.

[0103] (2) Scope of the shipping industry

[0104] This invention uses the Quaternion Maritime Distance (QSD) to define the safe distance between ships. Its formula can be expressed as:

[0105]

[0106] The definition of sgn(.) in the formula is as follows:

[0107]

[0108] The radius parameters are as follows:

[0109]

[0110] Where L is the captain; ξ AD and ξ DT The calculation is as follows:

[0111]

[0112] Step 4: Plan navigation routes. Based on the groundwork laid in the first three steps, the A* algorithm and APF algorithm are introduced to achieve intelligent collision avoidance for multiple vessels in confined waters.

[0113] Using the aforementioned ship collision risks and ship domain scope as safety constraints, a multi-ship intelligent obstacle avoidance method based on A* and APF algorithms is employed for path planning:

[0114] (1) Global path planning based on A*

[0115] The A* algorithm is a global programming algorithm. It's a heuristic search algorithm derived by adding certain constraints to the evaluation function φ(k) = f(k) + g(k) of the A* algorithm. The specific process of the A* algorithm is as follows: Assume the starting point S... i and target point G i Let f*(k) be the distance function from the starting point; and g*(k) be the distance function to the target point. Then the value function of the A* algorithm is φ*(k) = f*(k) + g*(k). If the A* algorithm takes the element with the smallest f from the priority queue each time and then updates the adjacent states, then the A* algorithm has an optimal solution when f(k) > 0 and 0 < g(k) ≤ g*(k).

[0116] (2) Local path planning based on APF

[0117] Artificial potential field (APF) is a classic path planning method commonly used for local path planning in mobile robots. Its main idea is to guide the movement of a robot using the combined potential field resulting from the attraction of the target and the repulsion of obstacles. The expression for this combined potential field is:

[0118] G(q)=G k (t)+G ω (t) (15)

[0119] In the formula, G(q) is the artificial potential field, i.e., the resultant potential field. k (t) represents the gravitational field, G ω (t) represents the repulsive field. The expressions for the repulsive potential field and the gravitational potential field are given in formulas (1) and (3).

[0120] The optimal navigation path for ships is formed based on the shortest path in the global plan and the path in the local navigation area.

[0121] Step 5: Optimize vessel traffic flow. Considering the obstacle avoidance conditions in restricted waters, optimize turning and deceleration for obstacle avoidance, obtaining smooth obstacle avoidance trajectories and scientifically safe obstacle avoidance methods. It also includes a function to evaluate obstacle avoidance effectiveness. Finally, coordinate transformation technology is used to present the obstacle avoidance effect on the electronic nautical chart.

[0122] Intelligent obstacle avoidance optimization involving artificial bee swarms of multiple vessels:

[0123] Intelligent collision avoidance for multiple vessels requires not only reasonable avoidance paths but also more scientific course and speed allocation to maintain course safety during collision avoidance. The artificial bee colony algorithm is a metaheuristic algorithm based on bee colonies that can optimize numerical problems, thereby achieving optimization of course and speed during collision avoidance. This method mainly includes four steps:

[0124] Initial stage

[0125] All nectar sources are initialized using scout bees, where the location of the nectar sources represents the navigation status of the ship, including its heading and speed. The initialization expression is:

[0126] f nj =k j +rand(0,1)×(ω j -k j (16)

[0127] • Employment stage

[0128] Foraging bees search for neighbors based on their memory of nectar source locations and calculate fitness to determine neighboring nectar sources:

[0129]

[0130] The fitness problem is solved by minimizing the following equation:

[0131]

[0132] In the formula, g nj For the neighbor's honey source, f tj For randomly selected nectar sources, This is a random number used to calculate the fitness value after a new nectar source is generated.

[0133] • Observation phase

[0134] The observation bee randomly selects a nectar source based on the nectar source information obtained by the foraging bee, and generates new nectar sources near the selected source. The fitness of the new nectar source is calculated, and the probability of it being selected by the observation bee is:

[0135]

[0136] In the formula, ψ n Here is the probability formula for nectar sources. RM represents the number of foraging bees and observation bees.

[0137] • Reconnaissance phase

[0138] If a foraging bee does not improve its performance due to observation by a bee within a certain number of optimization attempts, it will restart the initial phase after reaching that number of attempts, with the expression still being formula (16), until the optimal nectar source is found.

[0139] This invention uses a waterway near a certain location in the Yangtze River Estuary deep-water channel as a case study. This waterway is one of the busiest waterways. The current system guides vessels of all sizes to their respective lanes through a fixed-line system and a "north-main, south-auxiliary" navigation pattern, ensuring they do not interfere with each other and improving navigation safety. However, this waterway features diverse vessel types and a complex traffic environment, including not only merchant ships but also small vessels such as fishing boats that weave through the water. This poses a safety hazard to navigation in this area. The purpose of this invention is to establish a multi-vehicle avoidance mode that fully considers the risk of collision, preventing the formation of tense situations by having both vessels take avoidance actions as early as possible, and scientifically formulating avoidance plans using intelligent optimization algorithms.

[0140] Table 1 Initial Parameters of the Ship

[0141] S1 S2 S3 S4 S5 Starting point (nm) (15,0) (2.5,12) (0,14.2) (18,7.7) (5.3 8.9) End point (nm) (1,15.5) (18,6.5) (13.9,0) (1,15.5) (18,6.5) Heading (°) 314 127 136 237 46 Speed ​​(kn) 10 8 12 15 7

[0142] To address the problem of identifying natural obstacles and navigational aids in this waterway, this embodiment of the invention uses a convex hull algorithm to extract land and island contour information and reconstruct the scene from the captured nautical chart image of this waterway. Figure 2 ). Figure 2 Figures (a)-(d) in the figure illustrate the process of grayscale processing of a nautical chart image of islands and reefs and extracting their contours using the convex hull algorithm. Figure 2 Figures (e)-(h) in the figure illustrate the process of geometrically processing the outline of the entire selected water area and reconstructing the experimental scene by taking into account the navigational beacons. Figure 2 (h) The gray line segment in the upper left corner represents the division of the waterway into left and right sections in the alignment system, guiding ships to follow their designated routes. Simultaneously, this invention also utilizes the artificial potential field method to create a portrait of the environmental scene. Figure 3 The image includes a repulsive potential field. Figure 3 (a) diagram) and gravitational potential field ( Figure 3 (Figure (b)). This prepares for the subsequent intelligent multi-ship collision avoidance. Focusing on the multi-ship collision avoidance problem in complex encounter situations of ships in confined waters, five ships sailing in different directions were first designed in the above scenario (Table 1). Then, the collision risk of the five ships in the encounter situation was calculated to determine the priority of avoidance. Next, the ship domain was considered to ensure that the ships have a sufficiently safe distance for navigation. Finally, a multi-ship intelligent collision avoidance mode considering ship collision risk and ship domain was established. Figure 4 ). Figure 4 The results depict the collision risks perceived by vessel S3 from target vessels S1, S2, S4, and S5 at a certain moment. S2 is the highest priority target considered by S3 for avoidance, followed by S1. Based on this avoidance pattern, a collision risk threshold of 0.2 and a vessel domain boundary of 0.5 nm are set. Multi-vessel intelligent collision avoidance is achieved by combining the A* and APF algorithms. Figure 5 (Figure (a) in the middle). Figure 5 Figure (a) shows that the A* and APF algorithms can achieve intelligent collision avoidance for multiple vessels in confined waters, but the enlarged view in the upper right corner reveals that the trajectory is still not smooth. Furthermore, to further optimize the speed and heading for collision avoidance, this invention introduces an artificial bee colony algorithm. The optimization effect is as follows... Figure 5 As shown in Figure (b), the distribution of the ship's territory demonstrates that the boundary was maintained throughout the avoidance process, without any intrusion into the ship's territory. Furthermore, to more clearly illustrate the avoidance effect, this invention expands on the avoidance effect from a temporal perspective. Figure 5 Figure (c) shows the spatiotemporal trajectories of multiple vessels avoiding each other. Finally, using coordinate transformation technology, the electronic chart display was achieved.

[0143] To highlight the intelligent multi-ship collision avoidance decision-making scheme designed for confined waters, this invention analyzes the speed and heading of ships during the collision avoidance process. Figure 6 (See Figures (a)-(b) in the diagram). Since the implementation scenario already has lane separation and diversion under the fixed route system, turning measures will not be taken unless necessary, especially in the established route. Figure 6The heading change result in Figure (a) conforms to the actual heading operation. Given the constraints of steering, this invention considers deceleration and obstacle avoidance; the optimized speed change curve ( Figure 6 Figure (b) shows the speed changes of the five vessels during the avoidance process. To evaluate the safety and scientific validity of the avoidance scheme, this invention performed real-time vessel-to-vessel distance analysis and vessel-to-vessel collision risk analysis during the avoidance process. Figure 6 (Figures (c)-(d) in the text). Figure 6 Figure (c) shows that the minimum distance between the vessels during the avoidance process was 0.6 nm, which is in line with the limits in the maritime field. Figure 6 Figure (d) shows that the maximum collision risk during the avoidance process is approximately 0.7, indicating that no urgent situation has arisen during the encounter. Based on the above experimental verification, this invention is effective for multi-vehicle avoidance in confined waters and complies with the limitations imposed by maritime regulations and collision risk requirements.

[0144] In summary, this invention proposes an intelligent collision avoidance optimization method for multiple vessels in restricted waters, considering route control. It takes into account natural obstacles and route control information acquisition and avoidance scenarios, and introduces the artificial potential field method and A* algorithm to obtain multi-vehicle avoidance paths by combining collision hazards and the ship domain. It integrates the artificial bee colony algorithm to achieve intelligent collision avoidance for multiple vessels in restricted waters, obtaining the optimal route, avoidance heading, and speed, providing regulatory guidance for shore-based supervisors. This invention applies classical intelligent algorithms and biomimetic algorithms to the fields of maritime traffic risk management and complex waterway traffic risk engineering. It fully considers multiple information such as route control in restricted waters and artificial intelligence algorithms to provide collision avoidance optimization schemes for vessel navigation in restricted waters, realizing intelligent port traffic control demonstration applications and knowledge discovery, which is beneficial to ensuring navigation safety in waters near ports. This invention can serve as an auxiliary tool for shore-based control personnel to regulate vessel navigation.

[0145] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0146] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0147] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0148] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0149] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0150] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention.

[0151] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for intelligent collision avoidance optimization of multiple vessels in confined waters considering fixed-lineage conditions, characterized in that, Includes the following steps: To obtain environmental information and the status information of each vessel within the restricted waters; Based on the aforementioned environmental and status information, and in conjunction with the fixed-route navigation beacon, an artificial potential field algorithm is used to construct a navigation scenario profile. Under the aforementioned navigation scenario profile, the risk of ship collision is calculated and the safe distance between ships is determined; Based on the navigation scenario profile, the collision risk of the ships and the safe distance between the ships are used as collision avoidance conditions. A path planning algorithm combining global and local methods is used to plan the optimal navigation path for each ship. At the same time, a search algorithm is used to optimize the navigation status of each ship, thus completing the multi-ship intelligent collision avoidance optimization process. The steps for constructing a general aviation scenario profile include: Based on the environmental information within the restricted water area, the contours of natural obstacles are extracted using the convex hull algorithm to obtain the natural obstacle region; Based on the natural obstacle areas, navigational beacons, and status information, an artificial potential field algorithm is used to construct a navigation scenario profile. The steps for constructing a general aviation scenario profile include: Based on the repulsive force relationship between natural obstacles and navigational beacons in the natural obstacle area, a repulsive force field is constructed, wherein the expression of the repulsive force field is: K(t,t0)=||t-t0||2 In the formula, G k (t) is the repulsive potential field function, g is the proportionality coefficient, K(t,t0) is a vector, its direction is from the obstacle to the ship, and its magnitude is the Euclidean distance |t-t0| between the ship and the obstacle, where t0 is a constant representing the maximum influence range of the obstacle on the ship, and K λ It is a constant representing the maximum distance at which an obstacle exerts its effect on a ship; Based on the gravitational relationship between each ship's own position and the target position, a gravitational field is constructed, wherein the expression for the gravitational field is: k(t,t i )=||t-t i ||2 In the formula, G ω (t) is the gravitational field function, k(t,t) i () is a vector representing the ship's position t and the target point's position t. i Euclidean distance between |tt i |,k λ β is a constant, and β is the proportional gain coefficient; The repulsive and gravitational fields are used as a model for general aviation scenarios; The calculation process for the ship collision risk includes: Based on the status information of each vessel, calculate the DCPA and TCPA parameters, where the expressions for the DCPA and TCPA parameters are as follows: In the formula, DCPA t For DCPA parameters, α rt Let t be the relative heading of the target ship, TCPA t For TCPA parameters, μ r Let μ1 be the velocity of the target ship relative to the ship at time t, μ2 be the velocity of the ship at time t, and D be the velocity of the target ship at time t. t Let β be the distance between the two ships at time t. Tt Let t be the phase of the target ship and the bearing of this ship. Based on the DCPA and TCPA parameters, a negative exponential function is used to fit and calculate the ship collision risk, wherein the expression for the ship collision risk is: CRI t =ω×RDCPA t +τ×RTCPA t In the formula, CRI t For the ship collision risk level, ω and τ are weighting coefficients, a and b are adjustment coefficients, DCPA t ′ represents the dimensionless DCPA. TCPA ′ represents the dimensionless TCPA, RDCPA t As a risk component of DCPA, RTCPA t It is a risk component of TCPA.

2. The intelligent collision avoidance optimization method for multiple vessels in restricted waters considering route planning, as described in claim 1, is characterized in that... The steps for obtaining the natural obstacle area include: Let X be any subset of the Euclidean space W. The smallest convex set containing X is called the convex hull of X, denoted as conv(X) = {X1, X2, ..., X}. n }, where each convex hull corresponds to a natural obstacle object; Connect the outermost points of each convex hull to form a convex polygon, which serves as the natural obstacle region.

3. The intelligent collision avoidance optimization method for multiple vessels in restricted waters considering route planning, as described in claim 1, is characterized in that... The safe distance of the vessel is defined using quaternions in the vessel domain, and the safe distance is expressed as: in: In the formula, η k Let (x, y) be the boundary of the ship's domain, (x, y) be the ship's position, S be a quaternion, and D be the boundary of the ship's domain. fore D is the distance from the bow of the ship. aft D is the distance from the stern of the ship. starb D is the distance to the starboard side of the ship. portt ξ is the distance to the port side of the ship. AD ξ is the gain of the approach distance. DT The gain is the gyroscopic diameter, L is the ship length, and A is the gyroscopic diameter. D For the advance distance, D T V is the initial diameter of the cyclic curve. own This is the speed of the ship.

4. The intelligent collision avoidance optimization method for multiple vessels in restricted waters considering route planning, as described in claim 1, is characterized in that... The A* algorithm is used as the global path planning algorithm to plan the global shortest path, and the artificial potential field algorithm is used as the local path planning algorithm to plan the local path, so as to obtain the optimal navigation path for ships. In the A* algorithm path planning process, a value function is used to evaluate the value of each ship from the starting point S. i To the target point G i The total cost estimate is used to find the shortest path. The shortest path is obtained when f(k) > 0 and 0 < g(k) ≤ g*(k), where the expression for the value function is: φ*(k)=f*(k)+g*(k) In the formula, φ*(k) is the value function, f*(k) is the distance function from the starting point, g*(k) is the distance function to the target point, f(k) is the actual cost from the initial state to state k, and g(k) is the estimated cost of the optimal path from state k to the target state.

5. The intelligent collision avoidance optimization method for multiple vessels in restricted waters considering route-fixed systems, as described in claim 4, is characterized in that... The steps of the local path planning include: The artificial potential field algorithm applies a repulsive field and a gravitational field to the ship to form a resultant potential field, which guides the ship to move along a local path. The expression for the resultant potential field is as follows: G(q)=G k (t)+G ω (t) In the formula, G(q) is the artificial potential field, that is, the resultant potential field. k (t) represents the gravitational field, G ω (t) represents the repulsive force field.

6. The intelligent collision avoidance optimization method for multiple vessels in restricted waters considering route-fixed systems, as described in claim 1, is characterized in that... The search algorithm is an artificial bee colony algorithm, and the steps for optimizing the navigation status of each vessel using the artificial bee colony algorithm include: Initial Phase: Initialize all nectar sources using scout bees, where the location of the nectar sources represents the ship's navigation status, including heading and speed. The initialization expression is: f nj k j +rand(0,1)×(ω j -k j ) In the formula, f nj Let n be the nectar source, j be the population size, and k be the number of variables. j ω j f nj The upper and lower limits of the value; Employment stage: Foraging bees search for neighbors based on their memory of nectar source locations and calculate fitness to determine neighboring nectar sources: The fitness problem is solved by minimizing the following equation: In the formula, g nj For the neighbor's honey source, f tj For randomly selected nectar sources, These are random numbers used to calculate the fitness value of new nectar sources. Let be the objective function value of the nth nectar source; Observation phase: The observation bee randomly selects a nectar source based on the nectar source information obtained by the foraging bee, and generates new nectar sources near the selected source. The fitness of the new nectar source is calculated, and the probability of it being selected by the observation bee is: In the formula, ψ n This represents the probability that a nectar source is selected. Reconnaissance phase: If a foraging bee decides to abandon a nectar source after a given number of attempts, it becomes a scout bee and restarts the initial phase until it finds the optimal nectar source.

7. A computer-readable storage medium, characterized in that, Includes one or more programs executed by one or more processors of an electronic device, said one or more programs including instructions for executing the intelligent collision avoidance optimization method for multiple vessels in restricted waters considering the line-keeping system as described in any one of claims 1-6.

Images