Unmanned ship global path planning method and device based on ant colony algorithm, and medium

Abstract

The application discloses an unmanned ship global path planning method and device based on an ant colony algorithm, and a medium. The method comprises the following steps: step one, establishing a task environment model; step two, establishing an unmanned ship navigation cost function; and step three, solving the unmanned ship navigation cost by using an improved ant colony algorithm. The application has the following advantages: a relatively perfect unmanned ship global path planning optimization mathematical model is established, so that the model itself has a relatively accurate description for the navigation process of the unmanned ship; the navigation and power characteristics of the unmanned ship are considered, the navigation cost is introduced into the ant colony algorithm to represent pheromone and a heuristic factor, and on this basis, an initial pheromone setting method is proposed to accelerate the calculation speed of the algorithm; a pheromone updating rule is proposed to solve the problem that the traditional ant colony algorithm is prone to falling into a local optimum, and the global search ability of the algorithm is improved; and a pheromone updating mechanism of dead lock points is proposed to guarantee the solution diversity in the later optimization process of the algorithm.

Description

TECHNICAL FIELD

[0001] The present application relates to the water surface unmanned ship path planning technical field, and particularly relates to an unmanned ship global path planning method, device and medium based on an ant colony algorithm. BACKGROUND

[0002] The unmanned ship has a wide application prospect in the marine environment monitoring, marine rights maintenance and military fields, and has become a research hotspot in the field of marine intelligent equipment. The unmanned ship path planning enables it to reach the destination according to the preset, and is a prerequisite and guarantee for its autonomous task execution. According to the environmental information, the path planning of the unmanned ship is divided into global path planning and local path planning. Among them, the global path planning is the global path planning under the known task environmental information (usually satellite remote sensing map information), and the local path planning is the local real-time path re-planning according to the local environmental information obtained by the ship-borne equipment on the basis of the global path. Therefore, the global path planning is the basis and key of the unmanned ship path planning, and determines the performance of the unmanned ship path planning.

[0003] The global path planning problem of the unmanned ship is a typical combinatorial optimization problem, and the methods for solving this problem mainly include genetic algorithm, particle swarm optimization algorithm, ant colony algorithm and artificial potential field method. Compared with other intelligent optimization algorithms, the ant colony algorithm has the characteristics of parallel search, fast convergence speed, strong robustness, simple implementation and the like when applied to path planning, but has the shortcomings of low initial search efficiency, slow convergence speed and easy falling into local optimum. Therefore, it is necessary to improve the ant colony algorithm from aspects of pheromone update and heuristic information construction, so as to improve the adaptability of the algorithm to application environment and application object. SUMMARY

[0004] The purpose of the present application is to provide an unmanned ship global path planning method, device and medium based on an ant colony algorithm, which improves the operation speed, avoids falling into local optimum, and can guarantee the solution diversity in the later optimization process.

[0005] The purpose of the present application is achieved by the following technical solutions:

[0006] An unmanned ship global path planning method based on an ant colony algorithm, comprising the following steps:

[0007] Step 1: Establish a task environment model:

[0008] (1) Original image down-sampling. Since the sea area and land area are both large structure information in the original satellite remote sensing image, and according to the multi-resolution theory, it is beneficial to the segmentation processing of large structure information to use low-resolution images, therefore, the original image is down-sampled first. The down-sampling process of the original image is described as:

[0009] I s = D n×n [I o

[0010] In the formula, I o represents the original input image; I s represents the image after downsampling; D n×n [·] represents the operation of reducing the original image by n times in length and width respectively, where 1 < n < N, and N is the number of pixels in the length and width of the image itself.

[0011] (2) Fuzzy c-means clustering segmentation. By analyzing the downsampled image, it can be seen that the entire sea area can be divided into three categories: the deep sea area, the coastal area, and the land area, and the unmanned boat can perform tasks at sea. Therefore, it is also necessary to use the fuzzy c-means clustering segmentation algorithm to segment the sea area image. This method compares each pixel value in the RGB space with the center point value of each class, and then assigns a fuzzy value between 0 and 1 to each pixel to indicate which class center point the pixel is closer to, thereby achieving segmentation. The fuzzy rule is that the sum of the values of the pixel for all classes is 1. The process of fuzzy c-means clustering segmentation is described as:

[0012] {I d1 , I d2 , I d3} = C[I d

[0013] In the formula, I d1 , I d2 , I d3 respectively represent the three categories of the deep sea area, the coastal area, and the land area; the C[·] operator represents the fuzzy c-means clustering segmentation algorithm.

[0014] (3) Upsampling and reconstruction of the segmented image. The segmented deep sea area I d1 , the coastal area I d2 and the land area I d3 are upsampled and reconstructed to the same size as the original image. The processing process is described as:

[0015] I o1 = U n×n [I d1

[0016] I o2 = U n×n [I d2

[0017] I o3 = U n×n [I d3

[0018] In the formula, I​​​​​o1 , I o2 , I o3 respectively represent the reconstructed deep sea area, near sea area and land area; the U[·] operator uses a common bilinear interpolation method to make the edges of the reconstructed image smoother.

[0019] (4) Mask method to extract target information. First, the original RGB image is converted into a gray image, and then the reconstructed deep sea area I o1 , near sea area I o2 and land area I o3 are used as masks to segment the original image. The mask image is mapped to the original image I o , and the deep sea area and near sea area identified as navigable are mapped to white, and the land area identified as non-navigable is mapped to black. The specific mask method processing process is described as:

[0020]

[0021] where (x, y) represents the specific position of the pixel point, I s (x, y) represents the sea-land segmentation image. After sea-land segmentation, the black land area, white deep sea area and near sea area in the unmanned ship navigation sea area are separated.

[0022] Step two: Establish the unmanned ship navigation cost function, where the optimization objectives include the length of the track, the navigation energy consumption and the safety of the track; the constraint conditions include the maximum range and the limitation of the navigation speed:

[0023] (1) Optimization objective:

[0024] The length L of the track of the unmanned ship is:

[0025]

[0026] where i is the node where the unmanned ship is located, i+1 is the next node of the unmanned ship; m is the number of track segments.

[0027] The energy consumption E of the unmanned ship during navigation is entirely derived from the propulsion system, so:

[0028]

[0029] Assuming that the unmanned ship is uniformly navigated between node i and node i+1, the energy consumption E i,i+1 during this period is equal to the work done by the propulsion system to overcome the environmental disturbance resistance to the unmanned ship, so:

[0030] E i,i+1 = τ w |v usv |t

[0031] t = L i,i+1 |v|

[0032] where m is the number of track segments; v usv is the speed generated by the propulsion system of the unmanned surface vehicle; τ w is the resultant of the ocean environmental disturbance forces; and v is the actual sailing speed of the unmanned surface vehicle. It can be seen that the size of the sailing energy consumption of the unmanned surface vehicle is proportional to |v usv and inversely proportional to |v|. Therefore, when the thrust generated by the propulsion system is set to a fixed value, |v usv is also a fixed value. In order to obtain lower sailing energy consumption, the heading of the unmanned surface vehicle needs to be adjusted to fully utilize the ocean environmental forces to obtain a larger |v|.

[0033] To ensure the safety of the unmanned surface vehicle, the key is to realize the safest track from the starting point to the target point. The track safety P safe is:

[0034]

[0035] where d is the actual distance between the unmanned surface vehicle and the center of the obstacle; and d safe is the safety distance between the unmanned surface vehicle and the obstacle.

[0036] The unmanned surface vehicle sailing cost function composed of the track length L, the sailing energy consumption E, and the track safety P safe is given by the following formula:

[0037] min c = P safe (ε1L + ε2E)

[0038] where ε1 and ε2 are weight coefficients, and ε1 + ε2 = 1.

[0039] (2) Constraint conditions:

[0040]

[0041] where L max is the maximum allowable sailing distance of the unmanned surface vehicle; v i is the sailing speed of the unmanned surface vehicle in the i-th track segment; i is the node where the unmanned surface vehicle is located; and v max is the maximum allowable sailing speed of the unmanned surface vehicle.

[0042] Step three: solving the unmanned surface vehicle sailing cost by using the improved ant colony algorithm:

[0043] (1) Information and heuristic factor representation. In the traditional ant colony algorithm, i and j are positive integers, the information τ ij and the heuristic factor η ijAll only relate to track distance, and the influence of the constraint of unmanned ship dynamics and external disturbance on the motion performance of unmanned ship is not considered, so that the planned track cannot meet the actual application requirement of unmanned ship, and the effect of task execution is affected. ij The pheromone τ ij and the heuristic factor η ij are represented. The pheromone τ ij of the improved ant colony algorithm is represented as follows:

[0044] g≠h and at least one of g and h is equal to i or j

[0045] In the formula, g and h are positive integers, c gh represents the navigation cost of the unmanned ship between the track point g and the track point h; the pheromone τ of the track is inversely proportional to the navigation cost of the track, the smaller the navigation cost of the track is, the larger the pheromone on the track is, otherwise, the smaller the pheromone value on the track is. The representation of the heuristic factor η ij is as follows:

[0046]

[0047] In the formula, c ij and respectively represent the navigation cost between the track point i and the track point j and the navigation cost between the track point j and the terminal point P n .

[0048] (2) Initial pheromone setting. The initial value of the pheromone of the traditional ant colony algorithm is the same on each track, so that the ants have a certain blindness when searching the track in the early stage, and the convergence speed is slow. In order to speed up the algorithm, according to the navigation cost between the starting point P0, the current track point i, the next track point j and the terminal point P n , the initial value of the pheromone is unevenly distributed, so that the pheromone on the shorter track is higher, the blindness in the initial stage of the algorithm is avoided, and the search speed of the algorithm is improved. The specific implementation is as follows:

[0049]

[0050] In the formula, τ ij (0) represents the initial pheromone on the track (i, j); c ij and respectively represent the navigation cost between the starting point P0 and the terminal point P n and the navigation cost between the starting point P0 and the track point i. R F represents the deep sea area and the near sea area.

[0051] (3) Pheromone update rule. In the traditional ant colony algorithm, the update of pheromone on the path is mainly the residual pheromone part after volatilization with time and the pheromone increment after the passing of ant k Composition, the pheromone increment only related to ant k itself, without considering the interaction between the ant population, and due to the difference between the paths passed by each ant in each iteration process, the difference between the pheromones of each path is small, which leads to the algorithm being easily disturbed by the pheromone of the non-optimal path and falling into local optimum, wherein k is a positive integer. In order to solve this problem, the application fully considers the interaction between the ant population in the pheromone update process, and improves the update mechanism of the pheromone increment by introducing a weight factor h, thereby improving the global search ability of the algorithm.

[0052] The pheromone update rule is represented as:

[0053] τ ij (t+1)=(1-γ)τ ij (t)+Δτ ij (t,t+1)

[0054] In the formula, t represents time; τ ij (t+1) represents the pheromone on the path (i, j) at t+1 time; γ represents the pheromone volatilization coefficient; Δτ ij represents the pheromone increment on the path (i, j) at (t, t+1) time, which is represented as:

[0055]

[0056] Among them

[0057]

[0058] In the formula, represents the optimal global path planned by the ant population in the t+1 iteration optimization process; represents the total path length of the optimal global path planned in the t+1 iteration optimization process . Since the smaller the path navigation cost is, the greater the weight factor h is, and the greater the pheromone increment on the path is. When updating the pheromone, the pheromone on the path with smaller navigation cost will increase more, thereby enhancing the guiding effect of the better solution on subsequent iterations, the pheromone on the path with larger navigation cost will increase less, thereby reducing the misleading effect of the poor solution on subsequent iterations, thereby accelerating the convergence speed of the algorithm and improving the global optimization ability of the algorithm.

[0059] ​(4) pheromone update mechanism of deadlock point. In the traditional ant colony algorithm applied to path planning, only the ants trapped in the deadlock point are killed, and no treatment is performed on the pheromone of the deadlock point. In this way, on the one hand, the number of ants is reduced, the convergence speed is reduced, and the diversity of solutions is reduced, and on the other hand, the ants not trapped in the deadlock point still have a certain probability of being trapped in the deadlock point, so that the diversity of solutions in the later iteration optimization of the algorithm cannot be guaranteed. In order to solve this problem, the present application introduces a punishment factor μ to punish the pheromone of the deadlock point. The number of dead ants of the deadlock point N D_a , the greater the punishment degree of the pheromone of the point, that is, the greater the punishment factor μ, so that the number of subsequent ants trapped in the deadlock point is less, and the diversity of solutions is ensured. Therefore, after introducing the punishment factor μ, the pheromone update mechanism of the deadlock point is as follows:

[0060]

[0061] (5) path iteration. After the initial value of the pheromone is determined, the initial value of the pheromone and the heuristic factor related quantity is obtained, that is, the iteration is entered. In the tth iteration optimization process, that is, at t time, the next path point j selected by the ant from the path point i is determined according to the pheromone τ ij (t) and the heuristic factor η ij (t) determined by the pseudo-random state transition rule, that is, the ant directly transfers to the path point j which makes (τ ij (t)) α ·(η ij (t)) β max with a probability q0, otherwise the next path point j is selected by using the roulette method, and the selection rule is described as follows:

[0062]

[0063] Wherein

[0064]

[0065] In the formula, q0∈(0, 1) represents a threshold value of the ant selection rule; q∈(0, 1) represents a random number; represents a set of the next path point j selected by the ant; α represents a pheromone incentive factor; τ ij (t) represents the pheromone on the path (i, j) at t time; β represents an expected heuristic factor; η ij (t) represents the heuristic factor of the path (i, j) at t time; d ij , respectively represent the distance between the path point i and the path point j, and the distance between the path point j and the terminal point P nThe distance between the two points; R represents a path point selected by using a roulette wheel selection method, and the probability of transferring to each selectable path point is calculated by the following formula, and the path point obtained by using the roulette wheel selection method is selected.

[0066]

[0067] After iteration, a path about (tau(t)) α ·(eta(t)) β The path point corresponding to the path always maximizes the change over time, and the path composed of the path point is the target track.

[0068] Compared with the prior art, the unmanned ship track planning method has the advantages that:

[0069] (1) The present application establishes a relatively perfect global track planning optimization mathematical model for unmanned ships, wherein the optimization objectives include track length, navigation energy consumption and track safety; the constraint conditions include the maximum range and the limitation of navigation speed, so that the model itself has a relatively accurate description of the navigation process of the unmanned ship.

[0070] (2) The present application considers the navigation and power characteristics of the unmanned ship, introduces the navigation cost into the ant colony algorithm to represent the pheromone and the heuristic factor, and on this basis, proposes an initial pheromone setting method to speed up the calculation speed of the algorithm.

[0071] (3) The present application proposes a pheromone updating rule, solves the problem that the traditional ant colony algorithm is easy to fall into local optimum, and improves the global search ability of the algorithm.

[0072] (4) The present application proposes a pheromone updating mechanism for deadlocks to ensure the diversity of solutions in the later optimization process of the algorithm. BRIEF DESCRIPTION OF DRAWINGS

[0073] Figure 1 It is a flow chart of an unmanned ship global track planning method based on an ant colony algorithm;

[0074] Figure 2 It is a result graph of the original image down-sampling;

[0075] Figure 3 It is a fuzzy mean clustering segmentation result graph;

[0076] Figure 4 It is a result graph of the image up-sampling reconstruction after segmentation;

[0077] Figure 5 It is a result graph of extracting target information by using a mask method;

[0078] Figure 6A plot of the pheromone multiplied by the heuristic factor as a function of time;

[0079] Figure 7 A plot of the target track results. DETAILED DESCRIPTION

[0080] The specific embodiments of the present application will be further described below with reference to the accompanying drawings:

[0081] Referring to Figure 1 , the embodiment provides an unmanned ship global track planning method based on an ant colony algorithm, including the following steps:

[0082] Step 1: Establish a task environment model to complete the representation of feasible and infeasible areas for unmanned ship operations:

[0083] (1) Original image down-sampling. The original image is down-sampled, and the process is described as:

[0084] I s = D n×n [·]I o

[0085] In the formula, I o represents the original input image; I s represents the image after down-sampling; D n×n [·] represents an operation of reducing the original image by n times in length and width, 1 < n < N, N is the number of pixels in the length and width of the image itself. Here, n = 20 is used to reduce the original image by 20 times. A 100m*100m image is input, and after D 20×20 [·] operator processing, the output color map is as shown in Figure 2 .

[0086] (2) Fuzzy mean clustering segmentation. The fuzzy mean clustering segmentation algorithm is used to segment the sea area image. This method compares the value of each pixel in the RGB space with the center point value of each class, then assigns a fuzzy value between 0 and 1 to each pixel to indicate which class center point the pixel is closer to, thereby achieving segmentation. The fuzzy rule is that the sum of the values of all classes for this pixel is 1, and the process of fuzzy mean clustering segmentation is described as:

[0087] {I d1 ,I d2 ,I d3} = C[I d ]

[0088] In the formula, I d1 , I d2 , I d3 ​respectively represent the deep sea area, the near sea area and the land area; C[·] operator represents the fuzzy mean clustering segmentation algorithm. After C[·] operator processing, the output is shown in Figure 3 .

[0089] (3) Up-sampling reconstruction of the segmented image. The segmented deep sea area I d1 , the near sea area I d2 and the land area I d3 are up-sampled to reconstruct them to the same size as the original image, and the processing process is described as:

[0090] I o1 = U n×n [ I d1 ]

[0091] I o2 = U n×n [ I d2 ]

[0092] I o3 = U n×n [ I d3 ]

[0093] In the formula, I o1 , I o2 and I o3 represent the reconstructed deep sea area, the near sea area and the land area respectively; U[·] operator uses the common bilinear interpolation method to make the edges of the reconstructed image smoother. After U[·] operator processing, the output is shown in Figure 4 .

[0094] (4) Extracting target information by using the mask method, the specific process is: first, converting the original RGB image into a gray image, and then using the reconstructed deep sea area I o1 , the near sea area I o2 and the land area I o3 as a mask to segment the original image. Map the mask image to the original image I o , map the deep sea area and the near sea area identified as navigable to white, and map the land area identified as non-navigable to black. The specific mask method processing process is described as:

[0095]

[0096] In the formula, (x, y) represents the specific position of the pixel point, and I s (x, y) represents the sea-land segmentation image. After sea-land segmentation, the unmanned boat navigates in the black land area, the white deep sea area or the near sea area. The result is shown in Figure 5 . In the figure, the black area is the land area, and the white area is the deep sea area or the near sea area.

[0097] Step two: Establishing the unmanned vehicle navigation cost function, in which the optimization objective includes the length of the track, the energy consumption of navigation and the safety of the track; the constraint conditions include the maximum range and the limit of the navigation speed:

[0098] (1) Optimization objective:

[0099] The length of the track L of the unmanned vehicle is:

[0100]

[0101] In the formula, i is the node where the unmanned vehicle is located, i+1 is the next node of the unmanned vehicle; m is the number of track segments, which is set to 200.

[0102] The energy consumption E of the navigation route of the unmanned vehicle is entirely derived from the propulsion system.

[0103]

[0104] Assuming that the unmanned vehicle is uniformly navigated between the node i and the node i+1, the energy consumption E i,i+1 of the period is equal to the work done by the propulsion system to overcome the environmental disturbance resistance to the unmanned vehicle, so:

[0105] E i,i+1 = τ w |v usv |t

[0106] t = L i,i+1 / |v|

[0107] In the formula, m is the number of track segments; v usv is the speed generated by the propulsion system of the unmanned vehicle; τ w is the resultant force of the marine environmental disturbance; v is the actual navigation speed of the unmanned vehicle. It can be known that the size of the energy consumption of the navigation of the unmanned vehicle is proportional to |v usv | and inversely proportional to |v|. Therefore, when the thrust generated by the propulsion system is set to a fixed value, |v usv | is also a fixed value. In order to obtain lower energy consumption of navigation, the heading of the unmanned vehicle needs to be adjusted to fully utilize the marine environmental force to obtain a larger |v|.

[0108] In order to ensure the safety of the unmanned vehicle, the key is to realize the most safe track from the starting point to the target point. The safety P safe of the track is:

[0109]

[0110] In the formula, d is the actual distance between the unmanned vehicle and the center of the obstacle; d safe is the safety distance between the unmanned vehicle and the obstacle, which is set to 10 m.

[0111] The trajectory length L, the energy consumption E, and the safety P of the route safe The unmanned vehicle navigation cost function is given by the following formula:

[0112] min c = P safe (ε1L + ε2E)

[0113] In the formula, ε1 = 0.5, ε2 = 0.5 are weight coefficients, and ε1 + ε2 = 1

[0114] (2) Constraint conditions:

[0115]

[0116] In the formula, L max is the maximum allowable range of the unmanned vehicle, which is set to 20000km; v i is the navigation speed of the unmanned vehicle in the i-th trajectory segment; i is the node where the unmanned vehicle is located; v max is the maximum allowable navigation speed of the unmanned vehicle, which is set to 15m / s.

[0117] Step three: solving the unmanned vehicle navigation cost using the improved ant colony algorithm:

[0118] (1) Information pheromone and heuristic factor representation. Considering the length of the trajectory, the navigation energy consumption and the safety of the route, the constructed trajectory navigation cost c ij is used to represent the information pheromone τ ij and the heuristic factor η ij , i and j are both positive integers.

[0119] The information pheromone τ ij is represented as:

[0120] g≠h and at least one of g, h is equal to i or j

[0121] In the formula, g and h are positive integers, c gh represents the navigation cost between the trajectory point g and the trajectory point h; the information pheromone of the trajectory is inversely proportional to the navigation cost of the trajectory, the smaller the navigation cost of the trajectory, the larger the information pheromone on the trajectory, otherwise, the smaller the information pheromone value on the trajectory.

[0122] The heuristic factor η ij is represented as:

[0123]

[0124] In the formula, c ij and c jPn represent the navigation cost between the trajectory point i and the trajectory point j, and the navigation cost between the trajectory point j and the terminal P nThe navigation cost between the two.

[0125] (2) Initial pheromone setting. The initial pheromone of each path is calculated. According to the navigation cost between the starting point P0=(0, 0), the current path point i, the next path point j and the end point P n =(300, 300), the initial value of the pheromone is unevenly distributed, so that the pheromone on the shorter path is higher, avoiding the blindness of the algorithm in the initial stage, and improving the search speed of the algorithm. The specific implementation is as follows:

[0126]

[0127] Wherein

[0128]

[0129] In the formula, τ ij (0) represents the initial pheromone on the path (i, j); c ij and respectively represent the navigation cost between the starting point P0and the end point P n , the starting point P0and the path point i.

[0130] (3) Pheromone update rule. When the ants are searching for the path by following the pheromone, new pheromone will be left, so the pheromone on the path will volatilize, remain and superimpose. At t+1 time, the pheromone update rule on the path (i, j) is:

[0131] τ ij (t+1) = (1-γ)τ ij (t) + Δτ ij (t, t+1)

[0132] In the formula, t represents time; γ represents the pheromone volatilization coefficient, which is set to γ=0.2; Δτ ij represents the pheromone increment on the path (i, j) at (t, t+1) time, and its expression is:

[0133]

[0134] Wherein

[0135]

[0136] In the formula, k is a positive integer; represents the optimal global path planned by the ant colony in the t+1 iteration optimization process; represents the optimal global path planned by the ant colony in the t+1 iteration optimization process The total track length. Since a smaller track travel cost results in a larger weighting factor h, and a larger pheromone increment on the track. When updating pheromones, tracks with smaller travel costs will see a greater increase in pheromone, enhancing the guiding effect of better solutions on subsequent iterations. Tracks with larger travel costs will see a smaller increase in pheromone, reducing the misleading effect of worse solutions on subsequent iterations, thus accelerating the algorithm's convergence speed and improving its global optimization ability.

[0137] (4) Pheromone update mechanism at deadlock points. This invention introduces a penalty factor μ to punish the pheromones at deadlock points. The number N of deadlock-point ants is... D_a The more pheromones there are, the greater the penalty for that point, i.e., the larger the penalty factor μ, the fewer ants will subsequently get stuck at this deadlock point, ensuring solution diversity. Therefore, after introducing the penalty factor μ, the pheromone update mechanism for deadlock points is as follows:

[0138]

[0139] (5) The improved ant colony algorithm described above was applied to USV trajectory planning, and the trajectory planning results are as follows: Figure 5 As shown. The specific process is as follows: During the t-th iteration of optimization, i.e., at time t, the ant selects the next path point j from path point i based on the pheromone τ. ij (t) and heuristic factor η ij (t) is determined using a pseudo-random state transition rule, that is, the ant directly transitions with probability q0 to a state that makes (τ) ij (t)) α ·(η ij (t)) β If the path point j is the largest, then the next path point j is selected using a roulette wheel method, with the following path selection rules:

[0140]

[0141] in

[0142]

[0143] In the formula, q0∈(0,1) represents the threshold of the ant path selection rule, set as q0=0.5; q∈(0,1) represents a random number; Let represent the set of next path points j that the ant can choose; α represents the pheromone activation factor, set as α = 1.5; τ ij (t) represents the pheromone on the path (i,j) at time t; β represents the expected heuristic factor, set as β = 6.0; η ij (t) represents the heuristic factor of the path (i,j) at time t; d ij , These represent the distance between path point i and path point j, and the distance between path point j and the destination P, respectively. n The distance between; R represents the path point selected using the roulette wheel selection method, which is obtained by calculating the probability of moving to each available path point using the following formula, and then selecting it using the roulette wheel selection method.

[0144]

[0145] The product of pheromone and heuristic factor (τ(t)) α ·(η(t)) β The curve that changes iteratively over time is as follows Figure 6 As shown. By Figure 6 As shown, the algorithm proposed in this patent converges quickly, and the product of pheromone and heuristic factor is always maximized. The path formed by the path points corresponding to the product of the pheromone and heuristic factor is the target trajectory, as shown below. Figure 7 As shown.

[0146] It should be recognized that embodiments of the present invention can also be implemented or carried out by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer-readable storage medium. The method can be implemented using standard programming techniques—including implementation in a computer program on a non-transitory computer-readable storage medium configured to allow the computer to operate in a specific and predefined manner—according to the methods and drawings described in the specific embodiments. Each program can be implemented in a high-level procedural or object-oriented programming language to communicate with the computer system. However, if desired, the program can be implemented in assembly or machine language. In any case, the language can be a compiled or interpreted language. Furthermore, for this purpose, the program can run on a programmed application-specific integrated circuit (ASIC).

[0147] Further, the methods can be implemented in any type of computing platform operably connected to a suitable computing platform, including but not limited to a personal computer, mini-computer, mainframe, workstation, network or distributed computing environment, separate or integrated computer platforms, or in communication with charged particle tools or other imaging devices, and the like. Aspects of the present application can be implemented in machine readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard disk, optical read and / or write storage media, RAM, ROM, and the like, such that it can be read by a programmable computer to configure and operate the computer to perform the processes described herein when the storage medium or device is read by the computer. In addition, the machine readable code, or portions thereof, can be transmitted over wired or wireless networks. The present application described herein includes these and other different types of non-transitory computer readable storage media when such media include instructions or programs that implement the steps described above in conjunction with a microprocessor or other data processor. The present application also includes the computer itself when programmed according to the methods and techniques described herein.

[0148] The above description is merely that of the preferred embodiments of the present application and is not intended to limit the present application. The present application can be variously changed and modified by those skilled in the art without departing from the spirit and scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of the present application.

Claims

1. A global path planning method for unmanned surface vehicle based on ant colony algorithm, characterized in that, a cost function of unmanned surface vehicle navigation is established; an improved ant colony algorithm is used to solve the cost of unmanned surface vehicle navigation, including steps of pheromone, heuristic factor representation; initial pheromone setting; pheromone update rule; pheromone update mechanism of deadlock point; path iteration, the pheromone update rule includes: ； where t represents time; represents pheromone at time t+1 on the path represents the pheromone evaporation coefficient; represents time t+1 on the path the pheromone increment on the path​ ； wherein ； In the formula, represents the optimal global path planned by the ant colony in the first iteration optimization process; represents the optimal global path planned by the ant colony in the first iteration optimization process total path length of the optimal global path 2. The ant colony algorithm-based global path planning method for an unmanned surface vehicle according to claim 1, characterized in that, the cost function of unmanned surface vehicle navigation includes steps of: establishing an optimization goal, including path length, navigation energy consumption and path safety; Length of trajectory of an unmanned surface vehicle is: In the formula, i is the node where the unmanned ship is located, i+1 is the next node of the unmanned ship; is the number of track segments; The energy consumption of the unmanned ship during the voyage is all derived from the propulsion system, and the route energy consumption is: : Assuming the UUV is cruising at a constant speed between nodes and nodes , the energy consumption during this period is equal to the work done by the propulsion system to overcome the environmental disturbance drag on the UUV, so: wherein is the number of track segments; is the speed generated by the USV propulsion system; is the resultant of the ocean environmental disturbance forces; is the actual speed of the USV; Flight path safety To: ； wherein is the actual distance between the unmanned surface vehicle and the center of the obstacle; is the safety distance between the unmanned surface vehicle and the obstacle; The trajectory length , the energy consumption of the route , and the safety of the track of the unmanned ship navigation cost function is given by the following formula: ； wherein , are weight coefficients, and ; establishing a constraint condition, including maximum range and navigation speed limit; ； wherein is the maximum allowed range for the unmanned boat; is the sailing speed of the unmanned boat in the i-th segment of the route; is the sailing speed of the unmanned boat in the i-th segment of the route; i is the node where the unmanned boat is located; is the maximum allowed sailing speed for the unmanned boat.

3. The method of claim 1, wherein, the pheromone, heuristic factor representation includes: Pheromones characterized as: ； wherein i, j, g, h are positive integers, represents a cost of unmanned surface vehicle navigation between the track point and the track point ; Heuristics characterized by: wherein and represent the cost of travel between waypoints and waypoints and between waypoints and the end point respectively.

4. The ant colony algorithm-based global path planning method for an unmanned surface vehicle according to claim 1, characterized in that, the initial pheromone setting includes: ； wherein represents initial pheromone on a track; , , and represent sailing cost between start point and end point , start point and track point , respectively; represents deep sea area and near sea area.​ 5. The ant colony algorithm-based global path planning method for an unmanned surface vehicle according to claim 1, characterized in that, the pheromone update mechanism of deadlock point includes: Introducing a punishment factor After that, the pheromone updating mechanism of the deadlock point is as follows: 。 6. The ant colony algorithm-based global path planning method for an unmanned surface vehicle according to claim 1, characterized in that, the process of path iteration includes: Once the initial pheromone value is determined, the initial values ​​of the correlation between pheromone and heuristic factor are obtained, and the iteration begins; in the... In the next iteration of the optimization process, that is At that moment, the ant moved from the path point Select the next waypoint It is based on pheromones and heuristic factors The state transition rule is determined using pseudo-random rules, meaning the ants transition by probability. Directly transfer to make Largest path point Otherwise, use the roulette wheel method to select the next path point. The routing rules are described as follows: ； wherein wherein denotes a threshold value representing the ant routing rule; denotes a random number; denotes a set of next path points selected by the ant ; denotes a pheromone excitation factor; denotes the pheromone on the path at the moment; denotes an expected heuristic factor; denotes the heuristic on the path at the moment; , denote the distance between path point and path point , and between path point and the end point , respectively; denotes the path point selected by using the roulette wheel selection method, and the probability of transition to each selectable path point is calculated by the following formula, on the basis of which the path point is obtained by using the roulette wheel method.​​ ； After iteration success, a line about The path composed of the corresponding path points is always the maximum with time change, which is the target track.

7. A computer device, comprising: including: a memory for storing a computer program; a processor for executing the computer program to realize the method of any one of claims 1 to 6.

8. A readable storage medium, characterized by, the computer program is stored on the readable storage medium, and the computer program is executed by the processor to realize the method of any one of claims 1 to 6.

Images