A multi-port ship path planning method based on swarm intelligence computing

Abstract

The present invention discloses a multi-port ship path planning method based on swarm intelligence calculation, which is particularly suitable for the effective handling of dynamic obstacles in a multi-port environment. The method represents the chart information, including port locations, obstacles and channel restrictions, by constructing a two-dimensional grid map. An artificial potential field method is introduced to calculate the forces of each grid, including gravity and repulsion, to guide ants to move to the target port while avoiding obstacles. The ant colony algorithm uses pheromones to record and update each path, and enhances the real-time and adaptability of path planning through an improved pheromone update mechanism and dynamic potential field adjustment. The innovation of the present invention lies in the introduction of an adjustment factor q to balance global search and local utilization, and to adjust the strategy in real time to adapt to the dynamic changes of the marine environment, thereby optimizing route planning, reducing fuel consumption and improving navigation safety, and has important application value for commercial shipping, maritime search and rescue, and marine scientific research.

Description

Technical field:

[0001] The present invention relates to the technical field of port ship path planning, and in particular to a multi-port ship path planning method based on swarm intelligent computing. Background technology:

[0002] In the context of globalized economy, maritime transport plays a vital role as an important part of international trade. Ship path planning technology is one of the key technologies to ensure the efficiency and safety of maritime transport. This technology involves how to find the optimal or approximately optimal path for a ship to sail from its starting point to its destination in order to reduce travel time, fuel consumption, and avoid various maritime risks.

[0003] Traditional ship path planning methods mainly include the A* algorithm based on graph theory, the Dijkstra algorithm, and algorithms based on swarm intelligence such as the ant colony algorithm and the particle swarm optimization algorithm. These methods perform well in static environments and can handle simple path selection problems. A* algorithm: The A* algorithm is a widely used path search algorithm that estimates the best path from the current node to the target node through a heuristic evaluation function. However, this method is inefficient when dealing with dynamically changing environments because it cannot adjust to environmental conditions that have changed in real time. Ant colony algorithm: The ant colony algorithm imitates the foraging behavior of ants in nature and discovers paths through the pheromone communication mechanism between ants. Although this algorithm can find a globally optimized path, it converges slowly under multi-objective and dynamic obstacle conditions.

[0004] Although existing path planning technologies have made progress in many aspects, there are still some key deficiencies: Inadequate handling of dynamic obstacles: Most existing algorithms are good at handling static obstacles, but in dynamic environments, such as encountering other ships or sudden changes in obstacles, these algorithms often need to recalculate the path, resulting in low efficiency. Computational efficiency and real-time issues: For long-distance ship path planning, existing algorithms often require a long calculation time, which is not conducive to navigation operations that require fast response. Local optimality problem: Many heuristic-based algorithms are prone to falling into local optimal solutions, especially under complex nautical charts and multiple constraints. Summary of the invention:

[0005] The purpose of the present invention is to provide a multi-port ship path planning method based on swarm intelligence computing to address the defects of the prior art and solve the problems raised by the above-mentioned background technology.

[0006] The present invention proposes a multi-port ship path planning method based on swarm intelligence computing, which is particularly optimized for dynamic obstacles and multiple targets. By introducing a constant q to improve the state transition rule, the present invention balances the relationship between global search and local search and accelerates the convergence speed of the algorithm. At the same time, through the improved pheromone update mechanism and repulsion function, the adaptability and real-time performance of the algorithm in a dynamic environment are improved.

[0007] The present invention aims to solve the deficiencies of the prior art, and the proposed method comprises the following steps:

[0008] S1: Start, start the ship path planning method;

[0009] S2: Establish an environmental model to construct an environmental model for path planning, including nautical charts, port locations, obstacles and channel restrictions;

[0010] S3: Initialize related parameters, including the number of ant colonies, pheromone concentration and the upper limit of the number of iterations;

[0011] S4: Calculate the heuristic function and the value of the heuristic function at each step to evaluate the pros and cons of different paths and guide the search direction of the ants;

[0012] S5: Select the next node according to the state transfer rule;

[0013] S6: Determine whether the iteration is finished, whether the current number of iterations reaches the preset maximum number of iterations, or whether other termination conditions are met. If the iteration is finished, proceed to step S7; if the iteration is not finished, return to step S4;

[0014] S7: Update pheromone;

[0015] S8: Determine whether the ship has returned to the initial port or reached the target port. If the determination is yes, proceed to step S9; if the determination is no, proceed to step S4;

[0016] S9: Determine whether the maximum number of iterations has been reached, and again determine whether the number of iterations has reached the preset maximum value. If the determination is yes, proceed to step S8; if the determination is no, return to step S3;

[0017] S10: Output the optimal path. According to the optimal path found in the iterative process, the navigation path of the ship from the initial port to the target port is output.

[0018] As a technical preferred solution of the present invention, the specific operation of initializing the relevant parameters in step S3 is: initializing the grid map and the dynamic obstacle position, abstracting the environment into a two-dimensional grid map; initializing the obstacles, randomly placing dynamic obstacles in the map, and recording the positions of the dynamic obstacles.

[0019] As a technical preferred solution of the present invention, the specific operation of selecting the next node in step S5 is: initializing the ant colony algorithm parameters, calculating the probability of the path according to the current pheromone concentration and the heuristic function value, selecting the next moving position, and selecting the probability The specific calculation formula is:

[0020]

[0021] Among them, τ ij is the pheromone concentration from node i to node j, η ij is the heuristic function from node i to node j. A constant q is introduced when calculating the selection probability to control the balance between exploration and utilization. The specific calculation formula is:

[0022]

[0023] Here, q represents a random number in the range [0,1].

[0024] As a technical preferred solution of the present invention, the specific operation of updating the pheromone in step S7 is: when all ants complete a complete movement, the pheromone will be updated according to the movement path of the ants, and the specific calculation formula is:

[0025]

[0026] Among them, ρ is the pheromone volatility factor, is the pheromone increment left by ant k from node i to node j, and n is the number of ants.

[0027]

[0028] Where Q is the pheromone constant, L K is the length of the path completed by ant k.

[0029] As a technical preferred solution of the present invention, the updating of the pheromone specifically comprises the following steps:

[0030] T1: In each iteration, the ant moves from one node to another according to the current pheromone concentration and the heuristic function value;

[0031] T2: After completing the movement, update the pheromone according to formula (4);

[0032] T3: After completing the pheromone update, the ant's position is reset and a new round of iterative movement begins;

[0033] T4: Repeat steps T1 to T3 until the preset maximum number of iterations is reached or the optimal path from the starting point to the target port is found.

[0034] As a technical preferred solution of the present invention, the calculation of the pheromone function also includes the following features: the pheromone function is improved by an artificial potential field method, and the specific calculation formula is:

[0035]

[0036] Among them, τ' ij (m) is the potential field pheromone introduced by the artificial potential field method in the mth generation, τ″ ij (m) is the potential field ant colony pheromone concentration obtained using the mth generation potential field pheromone, m0 is the maximum number of iterations, and δ is the potential field pheromone coefficient.

[0037] As a technical preferred solution of the present invention, the movement of the ship in the total potential field is a synthesis of the gravitational field, the repulsive field and the pheromone concentration.

[0038] As a technical preferred solution of the present invention, the calculation formula of the gravitational field is:

[0039]

[0040] Among them, U att (P) is the gravitational field strength at point P, k att is the gravitational constant, |PP goal | is point P and point P goal The Euclidean distance between .

[0041] As a technical preferred solution of the present invention, the calculation formula of the repulsive field is:

[0042]

[0043] Among them, U rep (P) is the repulsive field strength at point P, k rep is the repulsive force constant, |PP oi | is the current position P and P oi The Euclidean distance between them, D0 is the safe distance, where in order to avoid the unreachable target and local minimum problems, the improved repulsion function is:

[0044]

[0045] When the distance between the ship and the obstacle |PP oi | is less than r, the repulsive force will become very large, thus avoiding the collision between the ship and the obstacle; when the distance between the ship and the obstacle is greater than r, the repulsive force will gradually decrease, allowing the ship to bypass the obstacle and reach the target point; when the distance between the ship and the obstacle is 0, the repulsive force exerted on the ship by the obstacle is also 0, solving the problems of falling into local optimality and unreachable targets.

[0046] As a technical preferred solution of the present invention, the calculation formula of the total potential field U(P) is:

[0047] U(P)=U att (P)+U rep (P)+U aco (P) (9)

[0048] When dealing with dynamic obstacles, assuming that the dynamic obstacle moves to the right at a certain speed, its position can be expressed as:

[0049]

[0050] in, is the new position of the obstacle, is the old position of the obstacle, and △x is the distance the obstacle moves to the right.

[0051] When updating the position of dynamic obstacles in the grid map, the position of dynamic obstacles needs to be considered when recalculating the path in each iteration. Combining the ACO and APF methods, the calculation formula of the total potential field U(P) is:

[0052] U(P)=U att (P)+U rep (P)+U aco (P)+U dyn (P) (11)

[0053]

[0054] Among them, k dyn is the dynamic obstacle repulsion constant.

[0055] Compared with the related prior art, the present invention has the following main technical advantages and the beneficial effects of the present invention are:

[0056] Enhanced adaptability: The algorithm of the present invention is able to adapt to rapidly changing environmental conditions, including various types of dynamic obstacles, thereby providing more reliable route planning.

[0057] Improve computational efficiency: By optimizing the pheromone update rules and introducing the artificial potential field method, the present invention significantly accelerates the convergence speed of the algorithm, shortens the computation time, and makes the algorithm more suitable for real-time or near real-time navigation planning.

[0058] Reduced energy consumption and improved safety: Through more accurate and efficient path planning, ships’ fuel consumption can be reduced and navigation safety can be increased, especially when avoiding collisions and dealing with complex waterway conditions.

[0059] Avoiding the local optimal trap: The improved exploration and exploitation balance mechanism effectively prevents the algorithm from falling into the local optimal solution and increases the possibility of finding the global optimal solution. Description of the drawings:

[0060] Figure 1 It is a flow chart of a multi-port ship path planning method based on swarm intelligence computing according to the present invention;

[0061] Figure 2 is a static environment path planning diagram of an embodiment of the present invention;

[0062] (a): Traditional ACO;

[0063] (b): Improved ACO-APF;

[0064] Figure 3 is a static environment iteration diagram of an embodiment of the present invention;

[0065] (a): Traditional ACO;

[0066] (b): Improved ACO-APF;

[0067] Figure 4 It is a b-spline static environment path planning diagram of an embodiment of the present invention;

[0068] (a): Traditional ACO;

[0069] (b): Improved ACO-APF;

[0070] Figure 5 It is a ship path planning diagram in a dynamic environment according to an embodiment of the present invention. Specific implementation method:

[0071] The present invention is further described below in conjunction with the accompanying drawings and examples. However, the present invention can be implemented in many different ways and should not be construed as being limited to the embodiments shown; on the contrary, these embodiments provide those skilled in the art with implementation methods that meet applicable legal requirements.

[0072] Example 1: According to Figure 1 As shown, this embodiment details a multi-port ship path planning method based on swarm intelligence computing to solve the problems of insufficient adaptability to dynamic environments, low computational efficiency, and easy to fall into local optimal solutions in the prior art. This embodiment will show how to achieve optimal navigation path planning from the starting port to the target port through a specific algorithm flow, while effectively dealing with dynamic obstacles at sea, such as moving ships, marine debris, etc.

[0073] 1. System architecture and initial setup;

[0074] The system first initializes the chart data and related navigation parameters. The chart data includes the location of ports, known static obstacles (such as reefs, shipwrecks, etc.), and common channel restrictions. The system uses a two-dimensional grid map to represent this information, and each grid cell represents a specific area on the sea. The status of each grid cell can be passable, obstructed, or restricted.

[0075] The initialization parameters of the ant colony algorithm include:

[0076] Number of ants: This value is set based on the size and complexity of the problem, usually between 50 and 100.

[0077] Pheromone concentration: The initial pheromone concentration on each path is set to a small positive number.

[0078] Heuristic factor (η): Calculated based on chart data, reflecting the straight-line distance from the current position to the target position.

[0079] Pheromone volatility (ρ): Usually set between 0.1 and 0.5 to simulate the natural phenomenon of pheromone volatilization over time.

[0080] 2. The iterative process of the ant colony algorithm;

[0081] Initialize the ant colony algorithm parameters, calculate the probability of the path based on the current pheromone concentration and the heuristic function value, and select the next moving position. Then, select the probability It can be calculated according to the following formula:

[0082]

[0083] In formula (1), τ ij is the pheromone concentration from node i to node j, η ij is the heuristic function from node i to node j.

[0084] When calculating the selection probability, a constant q is usually introduced to control the balance between exploration and exploitation. The new formula is as follows:

[0085]

[0086] Where q can be a random value, such as a random number in the range of [0,1], to increase the randomness and diversity of the algorithm. The pheromone and path are continuously updated iteratively until the maximum number of iterations is reached or the optimal path is found.

[0087] In the ant colony algorithm, the update of pheromone and path is achieved through the movement and feedback of ants. When all ants complete a complete movement, the pheromone will be updated according to the movement path of the ants. The process of pheromone update can be expressed as the following formula:

[0088]

[0089] In formula (3), ρ is the pheromone volatility factor, is the pheromone increment left by ant k from node i to node j, and n is the number of ants.

[0090] in:

[0091]

[0092] In formula (4), Q is the pheromone constant, L K is the length of the path completed by ant k.

[0093] In each iteration, after the ant completes a full move, the pheromone is updated according to the above formula. Then, the ant will reset and restart the next round of movement. This process will be repeated until the predetermined maximum number of iterations is reached or the optimal path is found.

[0094] In the ant colony algorithm, the concentration of pheromones determines the direction of the ants, which is an important link. At the beginning of the algorithm, the ants randomly select paths to search. Some ants may choose paths that are longer from the target point to update pheromones, which may mislead the path selection of subsequent ants and cause the optimal path to converge slowly. Since the artificial potential field method has an optimization effect on the ant colony algorithm, the pheromone function is improved by the artificial potential field method:

[0095]

[0096] In formula (5), τ' ij (m) is the potential field pheromone introduced by the artificial potential field method in the mth generation, τ″ ij (m) is the potential field ant colony pheromone concentration obtained using the mth generation potential field pheromone, m0 is the maximum number of iterations, and δ is the potential field pheromone coefficient.

[0097] 3. Application and improvement of potential field method;

[0098] Initialize the gravitational field repulsive field function. The gravitational field can be expressed as:

[0099]

[0100] In formula (6), U att (P) is the gravitational field strength at point P, k att is the gravitational constant, |PP goal | is point P and point P goal The Euclidean distance between .

[0101] The repulsive field can be expressed as:

[0102]

[0103] In formula (7), U rep (P) is the repulsive field strength at point P, k rep is the repulsive force constant, |PP oi | is the current position P and P oi The Euclidean distance between them, D0 is the safe distance.

[0104] In order to avoid the problem of unreachable target and local minimum, the improved repulsion function can be expressed as:

[0105]

[0106] In this improved repulsion function, when the distance between the ship and the obstacle |PP oi | is less than r, the repulsive force will become very large, thus avoiding the collision between the ship and the obstacle; when the distance between the ship and the obstacle is greater than r, the repulsive force will gradually decrease, allowing the ship to bypass the obstacle and reach the target point; at the same time, when the distance between the ship and the obstacle is 0, the repulsive force on the ship from the obstacle is also 0, solving the problems of falling into local optimality and unreachable targets.

[0107] In the combined method, the movement of the ship in the total potential field is the synthesis of the gravitational field, the repulsive field and the pheromone concentration. The total potential field U(P) can be expressed as:

[0108] U(P)=U att (P)+U rep (P)+U aco (P) (9)

[0109] When dealing with dynamic obstacles, assuming that the dynamic obstacle moves to the right at a certain speed, its position can be expressed as:

[0110]

[0111] In formula (10), is the new position of the obstacle, is the old position of the obstacle, and △x is the distance the obstacle moves to the right.

[0112] When updating the position of dynamic obstacles in the grid map, the position of dynamic obstacles needs to be considered when recalculating the path in each iteration. Combining the ACO and APF methods, the total potential field U(P) can be modified as follows:

[0113] U(P)=U att (P)+U rep (P)+U aco (P)+U dyn (P) (11)

[0114]

[0115] In formula (12), k dyn is the dynamic obstacle repulsion constant.

[0116] 4. Result evaluation and optimization;

[0117] After each iteration, the path efficiency of all ants is evaluated, including path length, number of obstacles encountered, etc. The system adjusts parameters based on the evaluation results, such as increasing or decreasing the number of ants, adjusting the pheromone volatilization rate, etc., to optimize the overall path search effect.

[0118] Example 2: To verify the effectiveness of the method in static and dynamic environments, two simulation environments are proposed, namely, multi-port ship path planning in a static obstacle environment and ship path planning in a dynamic obstacle environment.

[0119] In a static obstacle environment, simulation experiments were conducted in a 20×20 grid map, using traditional ACO and the improved ACO-APF of the present invention to plan the path. The number of ants was set to 50, the maximum number of iterations was set to 50, α was set to 0.8, β was set to 15, Q was set to 5, and ρ was set to 0.6. The route planning diagrams of the two algorithms are shown in Figure 2. Figure 2 As shown, the convergence curve changes as Figure 3 As shown in the figure, the path planning diagram after b-spline curve is as follows Figure 4 shown.

[0120] In the dynamic obstacle environment, a simulation experiment was conducted in a 15×15 grid map. The collision avoidance process was performed in combination with the planned path of the ship and the position of the dynamic obstacles to achieve the one-way path planning of the ship. Three dynamic obstacles were set in the grid map. They moved to the right at a uniform speed. The blue dots are moving tracks, and the parameter settings are the same as those in the static environment. Figure 5 shown.

[0121] Table 1: Comparative test results

[0122]

[0123] Table 1 shows the ship path planning in static and dynamic environments. In the static environment, the improved ACO-APF has only 6 iterations, which is 40% lower than the traditional ACO iterations, and the planned path length is 69.77, which is 9.2% lower than the traditional ACO path length. This shows that the improved ACO-APF algorithm is superior to the traditional ACO algorithm in terms of path planning efficiency and path length optimization. Figure 4It can also be seen that the curve after B-spline is smoother and the path length is shorter, which further improves the efficiency and quality of path planning.

[0124] It can be seen from the above embodiments that in a dynamic environment, Figure 5 It can be seen that the ship successfully avoids obstacles during the driving process, indicating that the path obtained by the method of the present invention after collision avoidance processing is more effective and has certain practical significance.

[0125] The above embodiments only express several implementation methods of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A multi-port ship path planning method based on swarm intelligence computing, the method comprising the following steps: S1: Start, start the ship path planning method; S2: Establish an environmental model to build an environmental model for path planning, and the nautical chart includes port locations, obstacles and channel restrictions; S3: Initialize related parameters, including the number of ant colonies, pheromone concentration and the upper limit of the number of iterations; The specific operations for initializing relevant parameters are as follows: initializing the grid map and the position of dynamic obstacles, abstracting the environment into a two-dimensional grid map; initializing obstacles, randomly placing dynamic obstacles in the map, and recording the positions of dynamic obstacles; S4: Calculate the heuristic function and the value of the heuristic function at each step to evaluate the pros and cons of different paths and guide the search direction of the ants; S5: Select the next node according to the state transfer rule; The specific operation of selecting the next node is: initializing the ant colony algorithm parameters, calculating the probability of the path according to the current pheromone concentration and the heuristic function value, and selecting the next moving position. The specific calculation formula is: Among them, τ ij is the pheromone concentration from node i to node j, η ij is the heuristic function from node i to node j. A constant q is introduced when calculating the selection probability to control the balance between exploration and utilization. The specific calculation formula is: Where q represents a random number in the range of [0,1], α is the pheromone importance factor, and β is the heuristic function importance factor; S6: Determine whether the iteration is finished, whether the current number of iterations reaches the preset maximum number of iterations, or whether other termination conditions are met. If the iteration is finished, proceed to step S7; if the iteration is not finished, return to step S4; S7: Update pheromone. The specific operation is: when all ants complete a complete movement, the pheromone will be updated according to the movement path of the ants. The specific calculation formula is: Among them, ρ is the pheromone volatility factor, is the pheromone increment left by ant k from node i to node j, and n is the number of ants; Where Q is the pheromone constant, L K is the length of the path completed by ant k; The updating of pheromones specifically includes the following steps: T1: In each iteration, the ant moves from one node to another according to the current pheromone concentration and the heuristic function value; T2: After completing the movement, update the pheromone according to formula (4); T3: After completing the pheromone update, the ant's position is reset and a new round of iterative movement begins; T4: Repeat steps T1 to T3 until the preset maximum number of iterations is reached or the optimal path from the starting point to the target port is found; The calculation of pheromone also includes the following features: the pheromone function is improved by artificial potential field method, and the specific calculation formula is: Among them, τ' ij m is the mth generation of artificial potential field pheromone, τ' i ' j m is the potential field ant colony pheromone concentration obtained using the mth generation potential field pheromone, m0 is the maximum number of iterations, and δ is the potential field pheromone coefficient; S8: Determine whether the ship has returned to the initial port or reached the target port. If the determination is yes, proceed to step S9; if the determination is no, proceed to step S4; S9: Determine whether the maximum number of iterations has been reached, and again determine whether the number of iterations has reached the preset maximum value. If the determination is yes, proceed to step S8; if the determination is no, return to step S3; S10: Output the optimal path. According to the optimal path found in the iterative process, the navigation path of the ship from the initial port to the target port is output.

Images