Path planning and collision avoidance decision-making method of unmanned surface ships during meeting of multiple ships

Abstract

The invention provides a path planning and collision avoidance decision-making method of unmanned surface ships during multi-ship encounter, which comprises the following steps: optimizing a collision risk degree model, and improving the calculation precision of the collision risk degree in combination with a quaternary ship field; according to the ship collision risk degree, the voyage, the steering angle, the international seaborne collision avoidance rule and the constraint of the optimal collision avoidance distance, constructing a fitness function to evaluate the quality of the path points; determining the weight of each evaluation factor in the fitness function through an analytic hierarchy process; an elite archiving strategy and a parameter adaptive adjustment mechanism are introduced, an adaptive differential evolution algorithm model combined with analytic hierarchy process is generated, a reasonable path is planned, and collision with obstacles and dynamic ships is avoided. According to the method, the calculation precision of the collision risk degree is remarkably improved, and the autonomous collision avoidance capability of the ship in a complex navigation environment is enhanced; and optimization of path point selection is ensured to be based on comprehensive consideration of multiple important factors, so that path planning is more scientific and reasonable.

Description

Technical Field

[0001] The present invention relates to the technical field of ship collision avoidance and path planning, and in particular to a path planning and collision avoidance decision method for an unmanned surface ship when multiple ships meet. Background Art

[0002] Globally, as the use of unmanned surface vessels (USVs) in commercial and military fields increases, their automated path planning and collision avoidance technologies are becoming increasingly important. USVs are able to operate without direct human control when performing ocean monitoring, marine resource exploration, underwater archaeology, border patrols, and search and rescue missions. However, these operating environments are often complex and changeable, containing a variety of static and dynamic obstacles, such as other ships, buoys, and weather changes, all of which place extremely high demands on the navigation system of unmanned vessels.

[0003] The core challenge of collision avoidance and path planning is how to accurately handle dynamic and static obstacles in complex sea areas in real time. Traditional methods usually rely on predetermined rules and simple sensor inputs, which may be sufficient in open and relatively simple sea areas, but in heavy traffic or complex sea conditions, these methods may not provide sufficient reaction time and decision support, especially when encountering unexpected situations.

[0004] Traditional collision avoidance technologies are mainly based on preset rules and algorithms, such as following the International Regulations for Preventing Collisions at Sea for simple obstacle avoidance operations, or implementing responses through basic sensor data. Although these methods provide collision avoidance functions to a certain extent, they often seem to be powerless when dealing with decisions in multi-target, high-speed, and complex environments. For example, in crowded waters, when multiple ships approach at the same time, traditional systems find it difficult to quickly and effectively handle complex dynamic collision avoidance decisions. In addition, these technologies usually rely on repetitive path planning algorithms and lack the ability to respond quickly to changes in the dynamic ocean environment, which may lead to inefficiency or untimely response in practical applications.

[0005] Although existing path planning technologies such as A* and Dijkstra algorithm perform well in computer navigation and map path planning, the paths planned in a dynamically changing ocean environment often do not meet the actual needs of navigation, and it is difficult to ensure the safety and economy of the path at the same time. Summary of the invention

[0006] Based on the above technical problems such as unmanned ships planning paths in complex marine environments being too close to obstacles or other ships, and long paths, a path planning and collision avoidance decision-making method for unmanned surface ships when multiple ships meet is provided, which has better performance and provides an efficient and accurate technical solution for autonomous collision avoidance of unmanned ships in complex sea areas.

[0007] The technical means adopted by the present invention are as follows:

[0008] A path planning and collision avoidance decision method for an unmanned surface vessel when multiple ships meet, comprising:

[0009] S1. Optimize the collision risk model and improve the calculation accuracy of collision risk by combining the four-dimensional ship field;

[0010] S2. Construct a fitness function to evaluate the quality of the path points based on the constraints of ship collision risk, range, turning angle, international maritime collision avoidance regulations, and optimal collision avoidance distance;

[0011] S3. Determine the weight of each evaluation factor in the fitness function through the hierarchical analysis method to enhance the availability and reliability of the fitness function;

[0012] S4. Introduce elite archiving strategy and parameter adaptive adjustment mechanism to generate an adaptive differential evolution algorithm model combined with hierarchical analysis to plan a reasonable path and avoid collisions with obstacles and dynamic ships.

[0013] Furthermore, step S1 specifically includes:

[0014] S11, select the closest encounter distance DCPA, the minimum encounter time TCPA, the distance between the two ships D, the relative position of the two ships A, and the ship speed ratio K to construct a collision risk model, and establish a collision risk factor set U = {DCPA, TCPA, D, A, K};

[0015] S12, defining the membership function of the closest approach distance DCPA, the minimum approach time TCPA, the distance between the two ships D, the relative position of the two ships A, and the ship speed ratio K, and at the same time introducing a restriction factor into the DCPA membership function for optimization;

[0016] S13, according to the importance of each factor in obtaining CRI, a weight set W = {W DCPA , W TCPA , W D , W A , W K}, where W DCPA >W TCPA >W D >W A >W K And W DCPA +W TCPA +W D +W A +W K =1; construct the optimized collision risk model as follows:

[0017] CRI=WDCPA k DCPA +W TCPA k TCPA +W D k D +W A k A +W K k K

[0018] Among them, W DCPA , W TCPA , W D , W A , W K are the weights of the closest encounter distance, the minimum encounter time, the distance between the two ships, the relative position of the two ships, and the ship speed ratio, respectively. DCPA , k TCPA , k D , k A , k K They are the membership functions after optimization of the five factors.

[0019] Furthermore, step S12 specifically includes:

[0020] S121, establish a coordinate system with the ship as the origin, due east as the positive direction of the x-axis, due north as the positive direction of the y-axis, and the ship's coordinates (x O ,y O ), speed v O ,course Target ship coordinates (x T ,y T ), speed v T ,course The distance between the two ships is D, and the true bearing of the own ship to the target ship and the true bearing of the target ship to the own ship are a OT 、a TO , v R is the relative speed of the two ships;

[0021] S122. A coordinate system is established with the target ship as the origin of coordinates, the direction of the bow as the positive direction of the y-axis, and the direction perpendicular to the bow to the right as the positive direction of the x-axis. The coordinate transformation of the own ship's position is performed:

[0022] x O1 =Dsinβ0

[0023] y O1 =Dcosβ0

[0024]

[0025] in,

[0026] S123, according to the converted coordinates (x O1 ,y O1 ), and the relative motion line equation of the own ship relative to the target ship is obtained:

[0027]

[0028]

[0029]

[0030] S124. Transform the relative motion line equation:

[0031]

[0032] Determine whether the ship's territory is invaded by analyzing whether the relative motion line intersects with the boundary of the ship's territory;

[0033] S125. Introduce a limiting factor to optimize the DCPA membership function as follows:

[0034]

[0035] Among them, r1 is the minimum safe encounter distance, r2 is the safe passing distance of the ship, p1 is the number of intersections between the relative motion line of the own ship relative to the target ship and the boundary of the target ship's ship area, and p2 is the number of intersections between the relative motion line equation of the target ship relative to the own ship and the boundary of the own ship's ship area;

[0036] S126. Define the membership function of TCPA as follows:

[0037]

[0038] in, d3 is the latest avoidance distance for the give-way vessel, and d4 is the distance at which the vessel can take measures to avoid the vessel;

[0039] S127. Define the membership function of the distance D between the two ships as follows:

[0040]

[0041] S128. Define the membership function of the relative position A as follows:

[0042]

[0043] S129, define the membership function of the speed ratio K as follows:

[0044]

[0045] Wherein, K0 is the threshold value of K, and K0=1.

[0046] Furthermore, step S2 specifically includes:

[0047] S21. Construct the objective function F1 based on CRI as follows:

[0048] F1=CRI

[0049] S22, construct the objective function F2 based on the voyage, as follows:

[0050]

[0051] Among them, (x i ,y i ) is the current path point, (x i-1 ,y i-1 ) is the previous path point, (x n ,y n ) is the destination point;

[0052] S23, construct an objective function F3 based on the steering amplitude, as follows:

[0053]

[0054] S24. According to whether the ship is acting as a give-way vessel, a design objective function F4 based on the International Regulations for Preventing Collisions at Sea is constructed as follows:

[0055]

[0056] Among them, θ r It represents the true position of the target ship relative to the own ship. This objective function is constructed according to whether the own ship is a give-way ship.

[0057] S25, construct an objective function F5 based on the optimal collision avoidance distance, as follows:

[0058]

[0059] Where D represents the distance between the two ships, and L0 represents the optimal collision avoidance distance;

[0060] S26. Construct the fitness function Fitness as follows:

[0061] Fitness=W1F1+W2F2+W3F3+W4F4+W5F5

[0062] Among them, W1, W2, W3, W4, and W5 are the weights of safety, total range, deflection angle, International Regulations for Preventing Collisions at Sea, and optimal collision avoidance distance, respectively.

[0063] Furthermore, in step S25, a fuzzy evaluation method is used to calculate the optimal collision avoidance distance L0, so that the decision-making process is more accurate and efficient. The specific method is as follows:

[0064] Construct a set P = {P1, P2, P3, P4, P5, P6, P7} containing multiple evaluation factors and an evaluation result set L = {L1, L2, L3, L4, L5}, where P1, P2, P3, P4, P5, P6, P7 are the relative motion speed judgment vector, the azimuth judgment vector, the speed ratio judgment vector, the ship density judgment vector, the navigation waters condition judgment vector, the meteorological condition judgment vector, and the ship size judgment vector respectively;

[0065] The evaluation result set L is used to define the possible values ​​of the optimal collision avoidance distance, specifically:

[0066] L1=2, L2=3, L3=4, L4=5, L5=6

[0067] This method allows the system to calculate the most appropriate collision avoidance distance by comprehensively considering the above factors, thereby improving navigation safety;

[0068] Construct the optimal collision avoidance distance judgment matrix E:

[0069]

[0070] The weight of each evaluation factor is X = (0.4, 0.2, 0.15, 0.1, 0.05, 0.05, 0.05), the evaluation result F = X·E, and L0 is determined by the maximum membership principle.

[0071] Further, step S3 specifically includes:

[0072] By constructing a judgment matrix, normalizing the judgment matrix, adding the processed matrix row by row and then normalizing it again, we can obtain the eigenvector and the weight of each factor W1=0.3913, W2=0.1353, W3=0.1813, W4=0.1156, W5=0.1765. The calculation results show that the calculated weights are consistent through consistency test, which ensures the accuracy and reliability of the evaluation process.

[0073] Further, step S4 specifically includes:

[0074] S41. Introducing the elite archiving strategy makes the mutation process have both the accuracy of local search and the breadth of global search;

[0075] S42. Introduce a parameter adaptive adjustment mechanism to dynamically optimize the algorithm's search capabilities.

[0076] Furthermore, step S41 specifically includes:

[0077] Sort the M individuals in the population according to the value of the fitness function, put the OR individuals with better fitness into the set P to form an elite population, and put the (M-OR) individuals with poorer fitness into the set Q to form a non-elite population; P and Q should satisfy The individual variation formula is as follows:

[0078]

[0079] in, and are individuals randomly selected from the elite population P, and and are individuals randomly selected from the non-elite population Q.

[0080] Further, step S42 specifically includes:

[0081] S421, calculating the adaptive scaling factor F G , the formula is as follows:

[0082]

[0083] Among them, F0 and F1 are the scaling factors of the first and second iterations respectively, and F G is the scaling factor at the Gth iteration, and are the fitness values ​​of the individuals at the Gth iteration and the 1st iteration respectively;

[0084] S422, calculate the cross factor CR n , the formula is as follows:

[0085]

[0086] in, and are the fitness value of the nth individual and the average fitness value of all individuals, is the minimum value of fitness.

[0087] Compared with the prior art, the present invention has the following advantages:

[0088] 1. The present invention provides a path planning and collision avoidance decision-making method for an unmanned surface vessel when encountering multiple ships. The collision risk model is optimized. By combining it with the four-dimensional ship field and introducing a restriction factor to optimize the membership function of DCPA, the calculation accuracy of the collision risk is significantly improved, the autonomous collision avoidance capability of the ship in a complex navigation environment is enhanced, and the potential collision risk is effectively reduced.

[0089] 2. The present invention provides a path planning and collision avoidance decision-making method for an unmanned surface vessel when encountering multiple ships. The improved collision risk model, the International Maritime Collision Prevention Regulations and the optimal collision avoidance distance are used as evaluation factors of the fitness function. The model determines the weight of each evaluation factor through the hierarchical analysis method, thereby improving the rationality and accuracy of the fitness function, ensuring that the optimization of path point selection is based on comprehensive consideration of multiple important factors, making path planning more scientific and reasonable.

[0090] 3. The present invention provides a path planning and collision avoidance decision-making method for an unmanned surface vessel when encountering multiple ships. By introducing an elite archiving strategy in a differential evolution algorithm, the accuracy of local search and the breadth of global search are effectively combined. The global search capability in the early stage of the algorithm is enhanced by adaptively adjusting the scaling factor F and the crossover factor CR to prevent premature convergence, and the local search capability is enhanced in the later stage of the algorithm to improve the accuracy and optimization level of path planning.

[0091] Based on the above reasons, the present invention can be widely promoted in the fields of ship collision avoidance and path planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0093] Figure 1 The figure is a flow chart of the method of the present invention.

[0094] Figure 2 A quaternion ship area map provided by an embodiment of the present invention.

[0095] Figure 3 This is a diagram showing the intersection of the ship area boundary and the relative motion line provided by an embodiment of the present invention.

[0096] Figure 4 A flow chart of an improved differential evolution algorithm provided by an embodiment of the present invention.

[0097] Figure 5 A real-time distance diagram between the ship and other targets in a two-ship encounter provided by an embodiment of the present invention.

[0098] Figure 6 A diagram of path planning and collision avoidance effects in a two-ship encounter scenario provided by an embodiment of the present invention.

[0099] Figure 7 A real-time distance diagram between the own ship and other targets in a four-ship encounter provided by an embodiment of the present invention.

[0100] Figure 8 This is a diagram of path planning and collision avoidance effects in a four-ship encounter scenario provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0101] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0102] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0103] like Figure 1 As shown, the present invention provides a path planning and collision avoidance decision method for an unmanned surface ship when multiple ships meet, comprising:

[0104] S1. Optimize the collision risk model and improve the calculation accuracy of the collision risk by combining it with the four-dimensional ship field. In this embodiment, CRI is used to evaluate the safety factor in the fitness function of a single path point, and the collision risk model is optimized and combined with the four-dimensional ship field theory to ensure that the estimation of potential collision risks is more accurate and reliable.

[0105] S2. According to the constraints of ship collision risk, range, steering angle, international regulations for preventing collisions at sea, and optimal collision avoidance distance, a fitness function is constructed to evaluate the quality of the path points; in this embodiment, the performance of the ship automatic collision avoidance system is enhanced by comprehensively considering the collision risk index (CRI), the international regulations for preventing collisions at sea (COLREGs), and the optimal collision avoidance distance. Specifically, these factors are integrated into the evaluation of the fitness function to form a comprehensive evaluation set F, with safety, economy, compliance with international regulations for preventing collisions at sea, and optimal collision avoidance distance as the core evaluation factors.

[0106] S3. Determine the weight of each evaluation factor in the fitness function through the analytic hierarchy process to enhance the availability and reliability of the fitness function; in this embodiment, the analytic hierarchy process is a mathematical method widely used in the decision-making process. It decomposes complex decision-making problems into smaller and more manageable parts, and systematically evaluates and compares these parts through quantitative methods. It not only improves the operating efficiency of the unmanned ship automatic collision avoidance system, but also enhances the system's ability to cope with complex navigation environments. It enables us to accurately measure the relative importance of factors such as collision risk, compliance with international maritime collision avoidance regulations, economy and optimal collision avoidance distance, and assign appropriate weights to them accordingly.

[0107] S4. Introduce elite archiving strategy and parameter adaptive adjustment mechanism to generate an adaptive differential evolution algorithm model combined with hierarchical analysis to plan a reasonable path and avoid collisions with obstacles and dynamic ships.

[0108] In specific implementation, as a preferred embodiment of the present invention, step S1 specifically includes:

[0109] S11, select the closest encounter distance DCPA, the minimum encounter time TCPA, the distance between the two ships D, the relative position of the two ships A, and the ship speed ratio K to construct a collision risk model, and establish a collision risk factor set U = {DCPA, TCPA, D, A, K};

[0110] S12, defining the membership function of the closest approach distance DCPA, the minimum approach time TCPA, the distance between the two ships D, the relative position of the two ships A, and the ship speed ratio K, and at the same time introducing a restriction factor into the DCPA membership function for optimization;

[0111] S13, according to the importance of each factor in obtaining CRI, a weight set W = {W DCPA , W TCPA , W D , W A , W K}, where W DCPA >W TCPA >W D >W A >W K And W DCPA +W TCPA +W D +W A +W K =1; construct the optimized collision risk model as follows:

[0112] CRI=W DCPA k DCPA +W TCPA k TCPA +W D kD +W A k A +W K k K

[0113] Among them, W DCPA , W TCPA , W D , W A , W K are the weights of the closest encounter distance, the minimum encounter time, the distance between the two ships, the relative position of the two ships, and the ship speed ratio, respectively. DCPA , k TCPA , k D , k A , k K They are the membership functions after optimization of the five factors.

[0114] In specific implementation, as a preferred embodiment of the present invention, step S12 specifically includes:

[0115] S121, establish a coordinate system with the ship as the origin, due east as the positive direction of the x-axis, due north as the positive direction of the y-axis, and the ship's coordinates (x O ,y O ), speed v O ,course Target ship coordinates (x T ,y T ), speed v T ,course The distance between the two ships is D, and the true bearing of the own ship to the target ship and the true bearing of the target ship to the own ship are a OT 、a TO , v R is the relative speed of the two ships;

[0116] S122. A coordinate system is established with the target ship as the origin of coordinates, the direction of the bow as the positive direction of the y-axis, and the direction perpendicular to the bow to the right as the positive direction of the x-axis. The coordinate transformation of the own ship's position is performed:

[0117] x O1 =Dsinβ0

[0118] y O1 =Dcosβ0

[0119]

[0120] in,

[0121] S123, according to the converted coordinates (x O1 ,y O1 ), and the relative motion line equation of the own ship relative to the target ship is obtained:

[0122]

[0123]

[0124]

[0125] S124. Transform the relative motion line equation:

[0126]

[0127] Determine whether the ship's territory is invaded by analyzing whether the relative motion line intersects with the boundary of the ship's territory; Figure 3 Schematic diagram of the intersection of the ship area boundary and the relative motion line.

[0128] S125. Introduce a limiting factor to optimize the DCPA membership function as follows:

[0129]

[0130] Among them, r1 is the minimum safe encounter distance, r2 is the safe passing distance of the ship, p1 is the number of intersections between the relative motion line of the own ship relative to the target ship and the boundary of the target ship's ship area, and p2 is the number of intersections between the relative motion line equation of the target ship relative to the own ship and the boundary of the own ship's ship area. In the optimized membership function, p1 and p2 are used to judge whether the ship area of ​​the own ship and the ship area of ​​the target ship are invaded, thereby improving the reliability of the selection of the membership function;

[0131] S126. Define the membership function of TCPA as follows:

[0132]

[0133] in, d3 is the latest avoidance distance for the give-way vessel, and d4 is the distance at which the vessel can take measures to avoid the vessel;

[0134] S127. Define the membership function of the distance D between the two ships as follows:

[0135]

[0136] S128. Define the membership function of the relative position A as follows:

[0137]

[0138] S129, define the membership function of the speed ratio K as follows:

[0139]

[0140] Wherein, K0 is the threshold value of K, and K0=1.

[0141] In specific implementation, as a preferred embodiment of the present invention, step S2 specifically includes:

[0142] S21. Safety is the primary factor in the assessment, which is mainly determined by the collision risk index (CRI). CRI is a quantitative indicator used to assess the collision risk when ships meet under specific navigation conditions. The objective function F1 based on CRI is constructed as follows:

[0143] F1=CRI

[0144] S22, economic efficiency is evaluated by considering the voyage length and turning range. When designing a ship route, it is necessary not only to ensure the safety of the route, but also to consider its economic benefits, that is, to choose the navigation path with the lowest cost. The optimization of the voyage length and turning range can significantly reduce fuel consumption and increase navigation efficiency, thereby achieving economic optimization while ensuring safety. The objective function F2 based on the voyage is constructed as follows:

[0145]

[0146] Among them, (x i ,y i ) is the current path point, (x i-1 ,y i-1 ) is the previous path point, (x n ,y n ) is the destination point;

[0147] S23, construct an objective function F3 based on the steering amplitude, as follows:

[0148]

[0149] S24. Compliance with COLREGs is also an important part of this fitness function. COLREGs provide detailed guidelines to regulate the behavior of ships in encounter situations and ensure that all operations comply with internationally recognized safety standards. By incorporating these rules into the fitness function, the decision-making of the automatic collision avoidance system is ensured to be consistent with international regulations, further enhancing regulatory compliance and operational safety. According to whether the ship is a give-way vessel, the design objective function F4 based on the international maritime collision avoidance regulations is constructed as follows:

[0150]

[0151] Among them, θ r It represents the true position of the target ship relative to the own ship. This objective function is constructed according to whether the own ship is a give-way ship.

[0152] S25. The optimal collision avoidance distance is a crucial parameter in the unmanned ship collision avoidance system. It determines when the ship should start taking collision avoidance measures to ensure safety. This distance not only reflects the current speed and position of the ship, but also takes into account the dynamic changes and possible interaction paths of the surrounding ships, thus providing sufficient time and space for effective collision avoidance operations. By scientifically setting the optimal collision avoidance distance, the potential collision risk caused by untimely response or improper operation can be minimized. The objective function F5 based on the optimal collision avoidance distance is constructed as follows:

[0153]

[0154] Where D represents the distance between the two ships, and L0 represents the optimal collision avoidance distance;

[0155] S26. Construct the fitness function Fitness as follows:

[0156] Fitness=W1F1+W2F2+W3F3+W4F4+W5F5

[0157] Among them, W1, W2, W3, W4, and W5 are the weights of safety, total range, deflection angle, International Regulations for Preventing Collisions at Sea, and optimal collision avoidance distance, respectively.

[0158] In specific implementation, as a preferred embodiment of the present invention, in step S25, a fuzzy evaluation method is used to calculate the optimal collision avoidance distance L0, so that the decision-making process is more accurate and efficient. The specific method is as follows:

[0159] Construct a set P = {P1, P2, P3, P4, P5, P6, P7} containing multiple evaluation factors and an evaluation result set L = {L1, L2, L3, L4, L5}, where P1, P2, P3, P4, P5, P6, P7 are the relative motion speed judgment vector, the azimuth judgment vector, the speed ratio judgment vector, the ship density judgment vector, the navigation waters condition judgment vector, the meteorological condition judgment vector, and the ship size judgment vector respectively;

[0160] The evaluation result set L is used to define the possible values ​​of the optimal collision avoidance distance, specifically:

[0161] L1=2, L2=3, L3=4, L4=5, L5=6

[0162] This method allows the system to calculate the most appropriate collision avoidance distance by comprehensively considering the above factors, thereby improving navigation safety;

[0163] Construct the optimal collision avoidance distance judgment matrix E:

[0164]

[0165] The weight of each evaluation factor is X = (0.4, 0.2, 0.15, 0.1, 0.05, 0.05, 0.05), the evaluation result F = X·E, and L0 is determined by the maximum membership principle.

[0166] In specific implementation, as a preferred embodiment of the present invention, step S3 specifically includes:

[0167] By constructing a judgment matrix, normalizing the judgment matrix, adding the processed matrix row by row and then normalizing it again, we can obtain the eigenvector and the weight of each factor W1=0.3913, W2=0.1353, W3=0.1813, W4=0.1156, W5=0.1765. The calculation results show that the calculated weights are consistent through consistency test, which ensures the accuracy and reliability of the evaluation process.

[0168] In specific implementation, as a preferred embodiment of the present invention, step S4 specifically includes:

[0169] S41. Introducing the elite archiving strategy makes the mutation process have both the accuracy of local search and the breadth of global search;

[0170] S42. Introduce a parameter adaptive adjustment mechanism to dynamically optimize the algorithm's search capabilities.

[0171] In specific implementation, as a preferred embodiment of the present invention, step S41 specifically includes:

[0172] Sort the M individuals in the population according to the value of the fitness function, put the OR individuals with better fitness into the set P to form an elite population, and put the (M-OR) individuals with poorer fitness into the set Q to form a non-elite population; P and Q should satisfy The individual variation formula is as follows:

[0173]

[0174] in, and are individuals randomly selected from the elite population P, and and are individuals randomly selected from the non-elite population Q.

[0175] In specific implementation, as a preferred embodiment of the present invention, step S42 specifically includes:

[0176] S421, calculating the adaptive scaling factor F G , the formula is as follows:

[0177]

[0178] Among them, F0 and F1 are the scaling factors of the first and second iterations respectively, and F G is the scaling factor at the Gth iteration, and are the fitness values ​​of the individuals at the Gth iteration and the 1st iteration respectively; in this embodiment, the scaling factor F is an important parameter for adjusting the disturbance intensity of the individuals in the search space. A higher F value will increase the disturbance of the individuals and improve the global search capability of the algorithm, but may lead to a slower convergence speed. On the contrary, a lower F value reduces the disturbance intensity, speeds up the convergence of the algorithm, and enhances the local search efficiency, but may reduce the diversity of the population and increase the risk of falling into a local optimum.

[0179] S422, calculate the cross factor CR n , the formula is as follows:

[0180]

[0181] in, and are the fitness value of the nth individual and the average fitness value of all individuals, is the minimum value of fitness. Figure 4 The flowchart of the improved differential evolution algorithm is shown in FIG. 1 . In this embodiment, the crossover factor CR determines the possibility that the vectors of each dimension of the initial individual are replaced by the vectors of each dimension of the mutant individual. A high CR value means that the characteristics of the mutant individual can be transmitted to the initial individual more frequently, thereby enhancing the fusion of information. A low CR value reduces this information exchange, allowing individuals to maintain higher independence, but at the same time may reduce the efficiency of information integration.

[0182] Example

[0183] In order to verify the effectiveness of the present invention, a detailed simulation study of the method was carried out on the MATLAB platform, and the experimental contents are as follows:

[0184] The simulation experiment in the present invention is designed to evaluate the effect of the proposed multi-ship collision avoidance path planning method in an open water environment with good visibility. The influence of environmental factors such as wind, waves, and currents is excluded in the experiment to more accurately measure the performance of the algorithm itself. The improved differential evolution algorithm (AHP-ADE) is compared with the improved particle swarm algorithm (I-PSO), and the effectiveness of each algorithm is verified by using factors such as safety, economy, and algorithm operation efficiency as evaluation indicators.

[0185] The specific parameters of the experimental objects are shown in the following table:

[0186] Table 1 Specific parameters of experimental objects

[0187]

[0188] The simulation experiment includes two typical scenarios: two-ship encounter (S1) and four-ship encounter (S2). The initial states of the experimental objects are shown in Table 2. In the scenario of two-ship encounter, static obstacles are added to the experimental design to further test the obstacle avoidance ability of the algorithm. This setting is intended to simulate obstacles that may be encountered in actual navigation and to test the algorithm's response strategy in complex environments. In the experiment, the target ship is set to keep sailing straight, while the ship plays the role of a give-way ship. When considering avoidance measures, the ship avoids by adjusting its course instead of slowing down or stopping. The key parameters of the algorithm are set as follows: number of iterations K = 100, population size M = 50, and individual dimension n = 10.

[0189] Table 2 Initial state of experimental subjects

[0190]

[0191] Two-ship encounter simulation experiment:

[0192] A scenario is set in which two ships move relative to each other in open waters, forming a crossing situation. The goal is to test whether the path planning algorithm can effectively identify potential collision risks and automatically adjust the course to avoid collision. Specifically, the initial position, speed, and course of each ship are set as shown in Table 2, and the collision avoidance actions and path adjustment process of the ships are simulated according to the instructions generated by the algorithm. Through this scenario, the reaction speed and decision-making accuracy of the algorithm when dealing with crossing encounter conflicts are evaluated. Figure 5 is the real-time distance between the ship and the target ship and the static obstacle. Figure 6 This is the simulation result of the encounter between two ships.

[0193] As shown in Table 3, the experimental results show that the I-PSO algorithm has deficiencies in path safety, especially the frequent close contact with the target ship, which increases the risk of collision. In contrast, the AHP-ADE algorithm can effectively maintain a larger safety distance from obstacles and target ships, with the closest distance to the target ship being about 1.98 nautical miles, ensuring navigation safety.

[0194] Table 3 Comparison results between I-PSO algorithm and AHP-ADE algorithm

[0195]

[0196] In addition, the I-PSO algorithm deviates from the scheduled route at the beginning of the voyage, resulting in a large deviation distance, which reduces the economic efficiency of the route. The AHP-ADE algorithm allows the ship to travel along the original path in the initial stage and only adjusts the course when necessary to avoid obstacles, thereby optimizing the deviation distance and improving economic efficiency. Therefore, the AHP-ADE algorithm performs better in terms of safety and economy.

[0197] Four-ship encounter simulation experiment:

[0198] This scenario is more complex and involves collision avoidance behaviors of four ships in the same waters. This scenario is designed to test the performance of the algorithm in a complex multi-ship interactive environment. The simulated situations include crossing routes and multi-point encounters. The specific starting point, speed, and heading of each ship are set as shown in Table 2. This scenario specifically examines the algorithm's ability to handle sudden collision avoidance situations in a dynamic multi-ship environment, and to minimize path deviation while maintaining navigation efficiency. Figure 7 is the real-time distance between the own ship and the target ship, Figure 8 This is the simulation result of four-ship encounter.

[0199] As shown in Table 4, the experimental results show that the AHP-ADE algorithm and the I-PSO algorithm perform equally well in terms of path economy, with total deviation distances of 16.15 nautical miles and 15.8 nautical miles, respectively, both showing good economy. However, in terms of safety, the I-PSO algorithm performs poorly. It violates the International Regulations for Preventing Collisions at Sea when meeting target ship 2, and fails to maintain a safe distance when meeting target ship 3, with the closest distance being only 0.25 nautical miles, increasing the risk of collision. In contrast, the AHP-ADE algorithm significantly improves navigation safety by adjusting the heading, following the collision avoidance regulations, and maintaining a larger safe distance during the meeting.

[0200] Table 4 Comparison results between I-PSO algorithm and AHP-ADE algorithm

[0201]

[0202] Based on the simulation experiments of two-ship and four-ship encounters, the AHP-ADE algorithm performs well in multiple dimensions, including deviation distance, the closest distance to the target ship and obstacles, and algorithm running time. This effectively verifies the applicability and advantages of the algorithm and proves its effectiveness and reliability in complex navigation environments.

[0203] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A path planning and collision avoidance decision method for an unmanned surface vessel when multiple ships meet, characterized in that: include: S1. Optimize the collision risk model and improve the calculation accuracy of collision risk by combining the four-dimensional ship field; S2. Construct a fitness function to evaluate the quality of the path points based on the constraints of ship collision risk, range, turning angle, international maritime collision avoidance regulations, and optimal collision avoidance distance; S3. Determine the weight of each evaluation factor in the fitness function through the hierarchical analysis method to enhance the availability and reliability of the fitness function; S4. Introduce elite archiving strategy and parameter adaptive adjustment mechanism to generate an adaptive differential evolution algorithm model combined with hierarchical analysis to plan a reasonable path and avoid collisions with obstacles and dynamic ships.

2. The path planning and collision avoidance decision method for an unmanned surface vessel when multiple ships meet according to claim 1, characterized in that: Step S1 specifically includes: S11, select the closest encounter distance DCPA, the minimum encounter time TCPA, the distance between the two ships D, the relative position of the two ships A, and the ship speed ratio K to construct a collision risk model, and establish a collision risk factor set U = {DCPA, TCPA, D, A, K}; S12, defining the membership function of the closest approach distance DCPA, the minimum approach time TCPA, the distance between the two ships D, the relative position of the two ships A, and the ship speed ratio K, and at the same time introducing a restriction factor into the DCPA membership function for optimization; S13, according to the importance of each factor in obtaining CRI, a weight set W = {W DCPA , W TCPA , W D , W A , W K }, where W DCPA >W TCPA >W D >W A >W K And W DCPA +W TCPA +W D +W A +W K =1; construct the optimized collision risk model as follows: CRI=W DCPA k DCPA +W TCPA k TCPA +W D k D +W A k A +W K k K Among them, W DCPA , W TCPA , W D , W A , W K are the weights of the closest encounter distance, the minimum encounter time, the distance between the two ships, the relative position of the two ships, and the ship speed ratio, respectively. DCPA , k TCPA , k D , k A , k K They are the membership functions after optimization of the five factors.

3. The path planning and collision avoidance decision method for an unmanned surface vessel when multiple ships meet according to claim 2, characterized in that: Step S12 specifically includes: S121, establish a coordinate system with the ship as the origin, due east as the positive direction of the x-axis, due north as the positive direction of the y-axis, and the ship's coordinates (x O ,y O ), speed v O ,course Target ship coordinates (x T ,y T ), speed v T ,course The distance between the two ships is D, and the true bearing of the own ship to the target ship and the true bearing of the target ship to the own ship are a OT 、a TO , v R is the relative speed of the two ships; S122. A coordinate system is established with the target ship as the origin of coordinates, the direction of the bow as the positive direction of the y-axis, and the direction perpendicular to the bow to the right as the positive direction of the x-axis. The coordinate transformation of the own ship's position is performed: x O1 =Dsinβ0 and O1 =Dcosβ0 in, S123, according to the converted coordinates (x O1 ,y O1 ), and the relative motion line equation of the own ship relative to the target ship is obtained: S124. Transform the relative motion line equation: Determine whether the ship's territory is invaded by analyzing whether the relative motion line intersects with the boundary of the ship's territory; S125. Introduce a limiting factor to optimize the DCPA membership function as follows: Among them, r1 is the minimum safe encounter distance, r2 is the safe passing distance of the ship, p1 is the number of intersections between the relative motion line of the own ship relative to the target ship and the boundary of the target ship's ship area, and p2 is the number of intersections between the relative motion line equation of the target ship relative to the own ship and the boundary of the own ship's ship area; S126. Define the membership function of TCPA as follows: in, d3 is the latest avoidance distance for the give-way vessel, and d4 is the distance at which the vessel can take measures to avoid the vessel; S127. Define the membership function of the distance D between the two ships as follows: S128. Define the membership function of the relative position A as follows: S129, define the membership function of the speed ratio K as follows: Wherein, K0 is the threshold value of K, and K0=1.

4. The path planning and collision avoidance decision method for an unmanned surface vessel when multiple ships meet according to claim 1, characterized in that: Step S2 specifically includes: S21. Construct the objective function F1 based on CRI as follows: F1=CRI S22, construct the objective function F2 based on the voyage, as follows: Among them, (x i ,y i ) is the current path point, (x i-1 ,y i-1 ) is the previous path point, (x n ,y n ) is the destination point; S23, construct an objective function F3 based on the steering amplitude, as follows: S24. According to whether the ship is acting as a give-way vessel, a design objective function F4 based on the International Regulations for Preventing Collisions at Sea is constructed as follows: Among them, θ r It represents the true position of the target ship relative to the own ship. This objective function is constructed according to whether the own ship is a give-way ship. S25, construct an objective function F5 based on the optimal collision avoidance distance, as follows: Where D represents the distance between the two ships, and L0 represents the optimal collision avoidance distance; S26. Construct the fitness function Fitness as follows: Fitness=W1F1+W2F2+W3F3+W4F4+W5F5 Among them, W1, W2, W3, W4, and W5 are the weights of safety, total range, deflection angle, International Regulations for Preventing Collisions at Sea, and optimal collision avoidance distance, respectively.

5. The path planning and collision avoidance decision method for an unmanned surface vessel when multiple ships meet according to claim 4, characterized in that: In step S25, a fuzzy evaluation method is used to calculate the optimal collision avoidance distance L0, making the decision process more accurate and efficient. The specific method is as follows: Construct a set P = {P1, P2, P3, P4, P5, P6, P7} containing multiple evaluation factors and an evaluation result set L = {L1, L2, L3, L4, L5}, where P1, P2, P3, P4, P5, P6, P7 are the relative motion speed judgment vector, the azimuth judgment vector, the speed ratio judgment vector, the ship density judgment vector, the navigation waters condition judgment vector, the meteorological condition judgment vector, and the ship size judgment vector respectively; The evaluation result set L is used to define the possible values ​​of the optimal collision avoidance distance, specifically: L1=2, L2=3, L3=4, L4=5, L5=6 This method allows the system to calculate the most appropriate collision avoidance distance by comprehensively considering the above factors, thereby improving navigation safety; Construct the optimal collision avoidance distance judgment matrix E: The weight of each evaluation factor is X = (0.4, 0.2, 0.15, 0.1, 0.05, 0.05, 0.05), the evaluation result F = X·E, and L0 is determined by the maximum membership principle.

6. The path planning and collision avoidance decision method for an unmanned surface vessel when multiple ships meet according to claim 1, characterized in that: Step S3 specifically includes: By constructing a judgment matrix, normalizing the judgment matrix, adding the processed matrix row by row and then normalizing it again, we can obtain the eigenvector and the weight of each factor W1=0.3913, W2=0.1353, W3=0.1813, W4=0.1156, W5=0.1765. The calculation results show that the calculated weights are consistent through consistency test, which ensures the accuracy and reliability of the evaluation process.

7. The path planning and collision avoidance decision method for an unmanned surface vessel when multiple ships meet according to claim 1, characterized in that: Step S4 specifically includes: S41. Introducing the elite archiving strategy makes the mutation process have both the accuracy of local search and the breadth of global search; S42. Introduce a parameter adaptive adjustment mechanism to dynamically optimize the search capability of the algorithm.

8. The path planning and collision avoidance decision method for an unmanned surface vessel when multiple ships meet according to claim 7, characterized in that: Step S41 specifically includes: Sort the M individuals in the population according to the value of the fitness function, put the OR individuals with better fitness into the set P to form an elite population, and put the (M-OR) individuals with poorer fitness into the set Q to form a non-elite population; P and Q should satisfy P∪Q=M, The individual variation formula is as follows: in, and are individuals randomly selected from the elite population P, and and are individuals randomly selected from the non-elite population Q.

9. The path planning and collision avoidance decision method for an unmanned surface vessel when multiple ships meet according to claim 7, characterized in that: Step S42 specifically includes: S421, calculating the adaptive scaling factor F G , the formula is as follows: Among them, F0 and F1 are the scaling factors of the first and second iterations respectively, and F G is the scaling factor at the Gth iteration, and are the fitness values ​​of the individuals at the Gth iteration and the 1st iteration respectively; S422, calculate the cross factor CR n , the formula is as follows: in, and are the fitness value of the nth individual and the average fitness value of all individuals, is the minimum value of fitness.

Images