High ice condition avoidance path planning method considering motion characteristics of ship in ice region

Abstract

The invention discloses a high-ice-condition avoidance path planning method considering motion characteristics of a ship in an ice region, and relates to the technical field of safe navigation of ships, and the method comprises the steps: constructing a coarse-scale grid map and a fine-scale grid map, and determining a high-ice-condition grid; determining an initial navigation path by using a global path planning algorithm based on the coarse scale grid map; determining an external exciting force and substituting the external exciting force into the three-degree-of-freedom manipulation motion equation to solve motion state parameters; based on the fine-scale grid map and the motion state parameters, taking each global path point of the initial navigation path as a sub-target point, generating a plurality of local paths, and combining the local paths to obtain candidate navigation paths; and repeatedly executing the step of generating the initial navigation path and the candidate navigation path until the candidate navigation path meets the navigation requirement, thereby obtaining an optimal navigation path. According to the method, global direction control and local motion capability analysis are both considered, and the accuracy and reliability of the optimal navigation path obtained through planning are effectively improved.

Description

Technical Field

[0001] The present application relates to the technical field of safe navigation of ships, and in particular to a method for planning a path to avoid high ice conditions that takes into account the motion characteristics of ships in ice areas. Background Art

[0002] Ice-bound vessels are leading equipment for implementing polar strategies and enhancing the region's voice. As core equipment for implementing polar strategies and developing resources, they undertake critical tasks such as polar scientific research, channel development, and emergency rescue. With global climate change and increasing polar activity, ice-bound vessels face significant challenges in navigating high-ice conditions in polar waters. High ice conditions not only increase the risk of collisions between target vessels and high-risk ice masses but also place higher demands on navigation path planning.

[0003] Currently, path planning methods for ice-travel vessels primarily rely on global path planning, treating sea ice as an obstacle to create a navigation path that avoids it. However, ice-travel vessels can actually break through ice and do not need to avoid all ice. Therefore, the navigation paths planned by this method are inaccurate and inefficient. Summary of the Invention

[0004] In response to the above-mentioned problems and technical needs, this application proposes a high ice avoidance path planning method that takes into account the motion characteristics of ships in ice areas. The technical solution of this application is as follows:

[0005] A high ice avoidance path planning method considering the motion characteristics of ships in ice areas comprises the following steps:

[0006] Constructing a coarse-scale grid map and a fine-scale grid map, and determining each high ice condition grid in the coarse-scale grid map and the fine-scale grid map respectively; the grid size of the coarse-scale grid map is larger than the grid size of the fine-scale grid map;

[0007] Based on the coarse-scale grid map, a global path planning algorithm is used to determine an initial navigation path of the target ship from a navigation starting point to a target point; the initial navigation path does not pass through a high ice condition grid of the coarse-scale grid map, and the initial navigation path includes a plurality of global path points from the navigation starting point to the target point;

[0008] Determine the external excitation force on the target ship, and substitute the external excitation force into the three-degree-of-freedom maneuvering motion equation to solve it, so as to obtain the real-time motion state parameters of the target ship;

[0009] Based on the fine-scale grid map and the real-time motion state parameters of the target ship, the global path points of the initial navigation path are used as sub-target points of the local path planning algorithm, and multiple local paths are generated using the local path planning algorithm; the local paths are combined to obtain a candidate navigation path for the target ship from the navigation starting point to the target point;

[0010] The steps of generating the initial navigation path and the candidate navigation paths are repeated until the generated candidate navigation paths meet the navigation requirements of the target ship, thereby obtaining the optimal navigation path.

[0011] A further technical solution is that each partial path corresponds to a sub-target point, the partial path from the navigation starting point to the sub-target point closest to the navigation starting point is the first partial path constituting the candidate navigation path, and the partial path from the sub-target point closest to the target point to the target point is the last partial path constituting the candidate navigation path. The partial path planning algorithm is used to generate any i-th partial path, including:

[0012] Generate multiple candidate paths from the end point of the i-1th local path to the i-th sub-target point using the local path planning algorithm;

[0013] Based on the real-time position and real-time motion state parameters of the target ship, the candidate state of each candidate path is determined, and the cost function of each candidate path with the candidate state as the selection state is calculated; the candidate path with the smallest cost function is selected as the i-th local path; the candidate state includes the selection state and the exclusion state, and the candidate path with the candidate state as the exclusion state is directly eliminated; the integer parameter i ≥ 1, the 0th sub-target point is the navigation starting point of the target ship, and the last sub-target point is the target point of the target ship.

[0014] A further technical solution is that the external excitation force on the target ship includes ice force and propeller thrust, any τ-th candidate path includes N local path points, and determining the candidate state of the τ-th candidate path includes:

[0015] According to the position of each local path point and the position of each high ice condition grid in the fine-scale grid map, the minimum distance between each local path point and the high ice condition grid is determined;

[0016] When the minimum distance between a local path point and the high ice condition grid is less than the safety distance threshold, or the motion state parameter of the target ship at the local path point does not meet the motion state parameter threshold, or the ice force acting on the target ship at the local path point is greater than the propeller thrust, the candidate state of the τth candidate path is determined to be the exclusion state; otherwise, the candidate state of the τth candidate path is determined to be the selection state.

[0017] A further technical solution is to determine the cost function C(τ) of any τ-th candidate path whose candidate state is the selected state, including:

[0018] Determine the safety cost f of the τth candidate path based on the minimum distance between each local path point and the high ice condition grid safe (τ);

[0019] Determine the dynamic feasibility cost f of the τth candidate path based on the motion state parameters of the target ship at each local path point dyn (τ);

[0020] The energy efficiency cost f of the τth candidate path is determined based on the motion state parameters of the target ship at each local path point and the ice force on the target ship. energy (τ);

[0021] The cost function for determining the τth candidate path is:

[0022] C(τ)=ω1f safe (τ)+ω2f dyn (τ)+ω3f energy (τ)

[0023] Among them, ω1 is the safety cost weight, ω2 is the dynamic feasibility cost weight, and ω3 is the energy efficiency cost weight.

[0024] Its further technical solution is to determine the security cost f of the τth candidate path safe (τ) includes:

[0025] Determine the minimum distance d between any k-th local path point of the τ-th candidate path and the high ice condition grid k ;

[0026] According to the minimum distance d between the kth local path point and the high ice condition grid k , determine the security cost of the τth candidate path

[0027] Its further technical solution is that the motion state parameters include the ship speed, and the energy efficiency cost f of the τth candidate path is determined. energy (τ) includes:

[0028] According to the ship speed V of the target ship at any k-th local path point of the τ-th candidate path k and the ice force on the target ship Determine the energy efficiency cost of the τth candidate path

[0029] Its further technical solution is that the motion state parameters include the ship speed and bow turning speed, and the dynamic feasibility cost f of the τth candidate path is determined. dyn (τ) includes:

[0030] According to the ship speed V of the target ship at any k-th local path point of the τ-th candidate path k and the ship speed V at the k-1th local path point k-1 , determine the speed change ΔV at the kth local path point k ;

[0031] According to the bow turning speed ω of the target ship at any k-th local path point of the τ-th candidate path k and the bow speed ω at the k-1th local path point k-1 , determine the bow speed change Δω at the kth local path point k ;

[0032] Determine the kinetic feasibility cost of the τth candidate path Among them, V max is the maximum speed of the ship, ω max is the maximum bow turning speed.

[0033] A further technical solution is that the three-degree-of-freedom control motion equation is constructed based on the ship-ice-water-engine-propeller-rudder multi-system coupled motion model; the constructed three-degree-of-freedom control motion equation is:

[0034]

[0035] Calculation of the target ship's real-time motion state parameters includes:

[0036] According to the target ship's longitudinal velocity u, the target ship's transverse velocity v and the target ship's heading angle ψ, the target ship's velocity component V along the x-axis of the geodetic coordinate system is determined. x =ucosψ-vsinψ, determine the velocity component V of the target ship along the y-axis of the geodetic coordinate system y = usinψ + vcosψ, the velocity component V of the target ship along the x-axis of the geodetic coordinate system is x and the velocity component V of the target ship along the y-axis of the geodetic coordinate system y Perform vector summation to obtain the target ship's ship speed V; determine the target ship's bow speed ω=r based on the target ship's bow angular velocity r;

[0037] Where m is the target ship mass, m x is the additional mass in the surge direction, m y is the additional mass in the sway direction, I zis the moment of inertia of the target ship rotating around the vertical axis passing through the center of gravity of the target ship, J z is the additional inertia moment of the target ship rotating around the vertical axis passing through the target ship's center of gravity; u is the target ship's surge velocity, v is the target ship's sway velocity, r is the target ship's bow angular velocity, is the surge acceleration of the target ship, is the sway acceleration of the target ship, is the yaw angular acceleration of the target ship; X H is the component of the viscous hydrodynamic force in the longitudinal direction, Y H is the component of the viscous hydrodynamic force in the sway direction, N H is the bow rolling moment component of viscous hydrodynamics, X P is the component of propeller thrust in the surge direction, X R is the component of the rudder force in the longitudinal direction, Y R is the component of the rudder force in the sway direction, N R is the yawing moment component of the rudder force, X ice is the component of the ice force in the longitudinal direction, Y ice is the component of the ice force in the sway direction, N ice is the yaw moment component of the ice force.

[0038] Its further technical solution is to determine the navigation time of the generated candidate navigation path from the navigation starting point to the target point. When the navigation time meets the time threshold and the curvature of the candidate navigation path at any path point meets the curvature threshold, it is determined that the candidate navigation path meets the navigation requirements of the target ship.

[0039] A further technical solution is to construct a coarse-scale grid map and a fine-scale grid map, including:

[0040] Acquire an image of the ice-covered sea area where the target ship is sailing, and determine sea ice parameters of the ice-covered sea area where the target ship is sailing; determine the coarse-scale grid size and the fine-scale grid size;

[0041] The ice-covered sea area image is gridded according to the coarse-scale grid size to obtain multiple coarse-scale grids. The seawater and sea ice are identified using image recognition methods, and the high ice condition attributes of each coarse-scale grid are determined based on sea ice parameters. The locations and high ice condition attributes of all coarse-scale grids are stored in an array to obtain a coarse-scale grid map.

[0042] The ice-covered sea area image is gridded according to the fine-scale grid size to obtain multiple fine-scale grids. The seawater and sea ice are identified using image recognition methods, and the high ice condition attributes of each fine-scale grid are determined based on sea ice parameters. The positions and high ice condition attributes of all fine-scale grids are stored in array form to obtain a fine-scale grid map.

[0043] The beneficial technical effects of this application are:

[0044] This application proposes a high ice avoidance path planning method that takes into account the motion characteristics of ships in ice areas. By integrating the global direction control capability of the global path planning algorithm with the focus on local motion characteristics of the local path planning algorithm, it solves the problem of low accuracy of the traditional global path planning algorithm. At the same time, it avoids the disadvantage of the traditional local path planning algorithm that is prone to falling into the local optimum. Moreover, under the guidance of the initial navigation path generated by the global path planning algorithm, it can reduce unnecessary calculations and effectively improve the computational efficiency of local path planning. Considering the efficient computing requirements of the global path planning algorithm and the high-precision computing requirements of the local path planning algorithm, a dual-scale grid map is constructed to provide grid map inputs of different scales for the global path planning algorithm and the local path planning algorithm, which ensures the computational efficiency of the global path planning algorithm and the computational accuracy of the local path planning algorithm.

[0045] Taking into account the ability of ice-breaking ships, only the grids in the sea ice-covered area that meet the characteristics of high ice conditions are set as high ice condition grids as obstacles. Compared with the traditional method of setting all grids in the sea ice-covered area as obstacle grids, this method is more in line with the actual situation, and the accurate obstacle marking also ensures the accuracy of the subsequent path planning algorithm. At the same time, taking into account the multi-system coupling characteristics of the actual application scenario of the ship, the three-degree-of-freedom control motion equation of the ice-covered ship is constructed based on the ship-ice-water-engine-propeller-rudder multi-system coupling motion model. Fully combined with the actual application scenario, the characteristics of the ship-ice effect and the motion characteristics of the ice-covered ship in the entire system are more realistically and effectively considered, ensuring the accuracy of the calculated motion state parameters, thereby further ensuring the accuracy of the optimal navigation path generated by the local path planning algorithm.

[0046] By analyzing the target vessel's performance and actual ship maneuvering requirements, and based on the safety, dynamic feasibility, and energy efficiency requirements of ice-bound vessels during navigation, a cost function for the local path planning algorithm is constructed according to the target vessel's real-time motion state. This cost function ensures the reliability of the optimal navigation path generated, thereby generating correct and feasible maneuvering instructions based on the planned optimal navigation path to ensure safe and stable navigation of the target vessel. This approach will help improve the survivability of ice-bound vessels in extreme environments and lay the foundation for implementing polar strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a flow chart of the high ice avoidance path planning method.

[0048] Figure 2 This is a schematic diagram of a dual-scale raster map.

[0049] Figure 3It is a schematic diagram of multiple candidate paths generated by the local path planning algorithm. DETAILED DESCRIPTION

[0050] The specific implementation of this application will be further described below with reference to the accompanying drawings.

[0051] This application proposes a high ice avoidance path planning method that takes into account the motion characteristics of ships in ice areas. Please refer to Figure 1 The specific steps of the method are as follows:

[0052] Step 1: construct a coarse-scale grid map and a fine-scale grid map, and determine each high ice condition grid in the coarse-scale grid map and the fine-scale grid map respectively; the grid size of the coarse-scale grid map is larger than the grid size of the fine-scale grid map.

[0053] First, the actual environmental field of the target ship's icy waters needs to be converted into a grid map for path planning. Global path planning algorithms require efficient computation without requiring high precision, while local path planning algorithms require higher precision. The larger the grid size of the grid map, the faster the path planning algorithm's computational efficiency but lower its precision. Because the computational requirements of global and local path planning algorithms differ, it is necessary to construct grid maps of different scales. These grid maps then record sea ice information in the target waters and determine obstacle information.

[0054] In one embodiment, the specific method of constructing the coarse-scale grid map and the fine-scale grid map is:

[0055] Acquire ice-covered sea area images of the target ship sailing in the ice-covered sea area, and determine the sea ice parameters of the target ship sailing in the ice-covered sea area; determine the coarse-scale grid size and the fine-scale grid size.

[0056] Real-scale sea ice imagery, such as that of the Antarctic and Arctic waters, is collected using ship-based sea ice parameter observation systems or large-scale sea ice satellite remote sensing data. Images of the target ship's navigating ice-covered waters are then selected from these real-scale sea ice images and subjected to image preprocessing, including image denoising and coordinate system conversion between pixel coordinates and actual environmental coordinates, to obtain images of the ice-covered waters that meet image quality requirements.

[0057] Based on statistical data on the physical and mechanical characteristics of sea ice in typical navigation areas or prediction formulas for these characteristics, sea ice information that can characterize the ice-covered sea environment of the target ship is determined. This includes sea ice parameters related to its physical properties (such as humidity, salinity, density, thickness, etc.) and mechanical properties (such as strength, Poisson's ratio, and elastic modulus). The statistical data and prediction formulas for these characteristics are publicly available. By consulting relevant data, statistical values ​​for the various parameters of the physical and mechanical properties of sea ice in the simulated sea area can be obtained.

[0058] Based on the coverage area of ​​the target ship's navigation ice area and the target ship's movement range, the coarse-scale grid size of the coarse-scale grid map is set according to the calculation requirements of the global path planning algorithm; the fine-scale grid size of the fine-scale grid map is set according to the calculation requirements of the local path planning algorithm. The specific grid size is customized according to the actual application requirements. This application sets the coarse-scale grid size to 50 meters and the fine-scale grid size to 2 meters based on experience and experimental results.

[0059] The ice-covered sea area image is gridded according to the coarse-scale grid size to obtain multiple coarse-scale grids. The seawater and sea ice are identified using image recognition methods, and the high ice condition attributes of each coarse-scale grid are determined based on sea ice parameters. The positions and high ice condition attributes of all coarse-scale grids are stored in array form to obtain a coarse-scale grid map.

[0060] The ice-covered sea area image is gridded according to the fine-scale grid size to obtain multiple fine-scale grids. The seawater and sea ice are identified using image recognition methods, and the high ice condition attributes of each fine-scale grid are determined based on sea ice parameters. The positions and high ice condition attributes of all fine-scale grids are stored in array form to obtain a fine-scale grid map.

[0061] The gridded image is used to detect seawater and sea ice using an image recognition method, and the image recognition method can adopt an existing technical method. This application uses a threshold segmentation method to extract seawater and sea ice areas, and each sea ice area corresponds to a piece of sea ice. Then, an edge detection method is used to detect each sea ice area to obtain all the grids contained in each piece of sea ice, and then determine whether each grid is seawater or sea ice and the sea ice to which it belongs, and finally form a rasterized map in a binary form. The coarse-scale grid map and the fine-scale grid map are constructed in the same way, and the only difference is the size and number of the constructed grids. Figure 2 As shown, the grid map with a larger grid size is a coarse-scale grid map, and the grid map with a smaller grid size is a fine-scale grid map. In each grid map, the black grid represents sea ice and the white grid represents sea water. The combination of the coarse-scale grid map and the fine-scale grid map constitutes a dual-scale grid map.

[0062] Because the location of each grid cell in a grid map and the sea ice parameters at that location vary, it is necessary to assign a value to each grid cell to identify the sea ice corresponding to that grid cell. The sea ice parameters include the sea ice to which the grid cell belongs and its sea ice parameters. When the sea ice parameters indicate that a grid cell does not belong to any sea ice, it is considered to be seawater. The specific method for assigning values ​​to each grid cell using sea ice parameters is as follows: for each grid cell in the sea ice area, a random value is assigned within a specified range based on the statistical value of each sea ice parameter. The high ice condition attribute of each grid cell is then determined based on the sea ice thickness statistics for high ice areas, such as ridges, ridges, and icebergs, combined with the maximum ice thickness that the target ship can break through. The high ice condition attribute indicates whether the grid cell is in high ice conditions or not. If the ice thickness parameter in the grid cell's sea ice parameters is greater than the sea ice thickness statistics for high ice areas or greater than the maximum ice thickness that the target ship can break through, the grid cell is considered to be in high ice conditions. The high ice grid serves as an obstacle grid for both the global and local path planning algorithms, enabling path planning to avoid high ice conditions. Finally, a two-dimensional array is used to store the location, sea ice parameters, and high ice properties of each grid cell in the coarse-scale and fine-scale grid maps.

[0063] Step 2: Based on the coarse-scale grid map, a global path planning algorithm is used to determine an initial navigation path for the target ship from the navigation starting point to the target point; the initial navigation path does not pass through the high ice grid of the coarse-scale grid map, and the initial navigation path includes multiple global path points from the navigation starting point to the target point.

[0064] First, the coarse-scale grid map is preprocessed. Edge dilation is used to appropriately magnify the high-ice areas in the coarse-scale environment map to increase the distance to obstacles and improve safety. Specifically, any existing edge dilation method can be used. This application uses an edge dilation method that expands the area consisting of all high-ice grids in the coarse-scale grid map outward by a circle.

[0065] A global path planning algorithm is then used to plan a path to avoid high ice conditions, generating a global path from the navigation starting point to the destination as the initial navigation path. This global path planning algorithm can employ any existing algorithm, such as an artificial potential field algorithm or an ant colony algorithm. This initial navigation path provides global guidance for subsequent local path planning algorithms, controlling their global direction and preventing them from falling into local optima. Furthermore, the initial navigation path effectively limits the computational area of ​​the local path planning algorithms to a certain extent, thereby reducing the computational effort and improving their efficiency.

[0066] Step 3: Determine the external excitation force on the target ship, and substitute the external excitation force into the three-degree-of-freedom maneuvering motion equation to solve it, so as to obtain the real-time motion state parameters of the target ship.

[0067] When navigating ice-covered waters, a target vessel is driven by a multi-system system: ship-ice-water-engine-propeller-rudder. Therefore, the target vessel is subject to external excitation forces including ice force, propeller thrust, rudder force, and the fluid force between the target vessel and the seawater. Constructing the target vessel's three-degree-of-freedom maneuvering equations based on this multi-system, multi-medium coupling allows for a more accurate reflection of the ice-induced effects, resulting in more precise parameters for the target vessel's motion state.

[0068] In one embodiment, based on the ship-ice-water-engine-propeller-rudder multi-system coupled motion model, the three-degree-of-freedom maneuvering motion equation is constructed by calculating the various external excitation forces on the target ship and superimposing them on the center of gravity of the ship for analysis. The constructed three-degree-of-freedom maneuvering motion equation is:

[0069]

[0070] Where m is the target ship mass, m x is the additional mass in the surge direction, m y is the additional mass in the sway direction, I z is the moment of inertia of the target ship rotating around the vertical axis passing through the center of gravity of the target ship, J z is the additional inertia moment of the target ship rotating around the vertical axis passing through the target ship's center of gravity; u is the target ship's surge velocity, v is the target ship's sway velocity, r is the target ship's bow angular velocity, is the surge acceleration of the target ship, is the sway acceleration of the target ship, is the yaw angular acceleration of the target ship; X H is the component of the viscous hydrodynamic force in the longitudinal direction, Y H is the component of the viscous hydrodynamic force in the sway direction, N H is the bow rolling moment component of viscous hydrodynamics, X P is the component of propeller thrust in the surge direction, X R is the component of the rudder force in the longitudinal direction, Y R is the component of the rudder force in the sway direction, N R is the yawing moment component of the rudder force, X ice is the component of the ice force in the longitudinal direction, Y ice is the component of the ice force in the sway direction, N ice is the yaw moment component of the ice force.

[0071] Viscous hydrodynamic forces, propeller thrust, and rudder forces are calculated based on the target vessel's ship parameters, including vessel length, draft, propeller pitch ratio, rudder aspect ratio, and other vessel attributes. Ice forces are calculated based on the sea ice parameters of the grid of sea ice in contact with the target vessel. The specific calculation methods for viscous hydrodynamic forces, propeller thrust, rudder forces, and ice forces can be solved using existing empirical formulas or time-domain methods. This is not the focus of this application and will not be further elaborated.

[0072] Substituting all external excitation forces into the three-degree-of-freedom maneuvering motion equations and solving them using the fourth-order Runge-Kutta method, the target ship's motion state parameters at the current simulation moment are obtained. The motion state parameters include the target ship's velocity and angular velocity in each degree of freedom direction. The target ship's ship speed and bow speed can be obtained by converting the velocity and angular velocity in each degree of freedom direction. The specific calculation method is: Based on the target ship's longitudinal and transverse velocity u and the target ship's heading angle ψ, the target ship's velocity component V along the x-axis of the geodetic coordinate system is determined. x =ucosψ-vsinψ, determine the velocity component V of the target ship along the y-axis of the geodetic coordinate system y = usinψ + vcosψ, the velocity component V of the target ship along the x-axis of the geodetic coordinate system is x and the velocity component V of the target ship along the y-axis of the geodetic coordinate system y Perform vector summation to obtain the target ship's speed V. Determine the target ship's bowing velocity ω = r based on the target ship's yaw angular velocity r. The angle between the x-axis of the ship's coordinate system and the x-axis of the geodetic coordinate system is the target ship's heading angle ψ. The positive x-axis of the ship's coordinate system points toward the bow, and the positive y-axis of the ship's coordinate system points toward the starboard side. The positive x-axis of the geodetic coordinate system points horizontally to the right, and the positive y-axis points vertically downward.

[0073] Step 4: Based on the fine-scale grid map and the real-time motion state parameters of the target ship, the global path points of the initial navigation path are used as sub-target points of the local path planning algorithm, and multiple local paths are generated using the local path planning algorithm; the local paths are combined to obtain the candidate navigation path of the target ship from the navigation starting point to the target point.

[0074] The initial navigation path includes multiple global path points from the navigation starting point to the target point. There is a sequence between each global path point. The global path points are used as sub-target points of the local path planning algorithm. The local path planning algorithm is guided to generate each local path in sequence according to the sequence of the sub-target points. Then, the candidate navigation path of the target ship from the navigation starting point to the target point can be obtained by combining each local path in sequence. The local path combination method can adopt the head-to-tail splicing method, or the partial overlapping local splicing method, etc., which can be selected according to the actual calculation requirements. Among them, the selection of sub-target points can also be partially deleted, and a specified number of global path points can be selected as sub-target points to improve the calculation efficiency. For example, dense sub-target points can be selected near high ice obstacle areas, and the number of sub-target points can be reduced in flat straight paths. Due to the complex navigation environment of ships sailing in ice areas, in order to ensure sufficient safety, the local path planning algorithm needs to be as accurate as possible. Therefore, this application uses all global path points as sub-target points.

[0075] The goal of the local path planning algorithm is to safely avoid high ice grids while adhering as closely as possible to the macro-guidance of the initial navigation path generated by the global path planning algorithm. Each local path corresponds to a sub-target point. The local path from the navigation starting point to the sub-target point closest to the navigation starting point is the first local path that constitutes the candidate navigation path, and the local path from the sub-target point closest to the target point to the target point is the last local path that constitutes the candidate navigation path. In one embodiment, the specific steps for generating any i-th local path using the local path planning algorithm are:

[0076] (1) Use the local path planning algorithm to generate multiple candidate paths from the end point of the i-1th local path to the i-th sub-target point.

[0077] Guided by the sub-target points, the system generates local paths based on a local path planning algorithm, taking into account the target vessel's motion and maneuverability. This algorithm can employ any existing technique, such as the dynamic window method or neural networks. For each sub-target point, the algorithm generates multiple candidate paths, requiring evaluation of each candidate path and selection of the optimal one.

[0078] (2) Based on the real-time position and real-time motion state parameters of the target ship, the candidate state of each candidate path is determined, and the cost function of each candidate path with the candidate state as the selection state is calculated; the candidate path with the smallest cost function is selected as the i-th local path; the candidate state includes the selection state and the exclusion state, and the candidate path with the candidate state as the exclusion state is directly eliminated; the integer parameter i ≥ 1, the 0th sub-target point is the navigation starting point of the target ship, and the last sub-target point is the target point of the target ship.

[0079] The method for evaluating candidate paths involves determining the candidate status of each candidate path and directly eliminating any candidate paths with an excluded status from the calculation. This reduces computational effort and improves efficiency. The quality of each candidate path is then quantified using a cost function. A higher cost function value indicates a higher cost and a lower quality candidate path. Therefore, the candidate path with the lowest cost function is selected as the optimal path when evaluating its quality.

[0080] Considering the complex ice conditions encountered by the target vessel, safely avoiding high ice conditions is a primary concern. Icebreaking capability is also crucial; insufficient propulsion will render the target vessel unable to proceed. Furthermore, the target vessel's real-time motion state is crucial for ensuring its dynamic performance and determining its maneuverability. Excessive fluctuations in motion between moments will exceed controllability and affect the target vessel's stability.

[0081] The external excitation forces on the target ship include ice force and propeller thrust. Any τ-th candidate path includes N local path points. In one embodiment, determining the candidate state of the τ-th candidate path includes:

[0082] According to the position of each local path point and the position of each high ice grid in the fine-scale grid map, the minimum distance between each local path point and the high ice grid is determined; when the minimum distance between a local path point and the high ice grid is less than the safety distance threshold, or the motion state parameter of the target ship at the local path point does not meet the motion state parameter threshold, or the ice force exerted on the target ship at the local path point is greater than the propeller thrust, the candidate state of the τth candidate path is determined to be the exclusion state; otherwise, the candidate state of the τth candidate path is determined to be the selection state.

[0083] Among them, considering the obstructive effect of ice force on the target ship, it is necessary to compare the propeller thrust and ice force to determine whether the propulsion capacity of the target ship is large enough, so as to determine the feasibility of the candidate path. Due to the time lag problem that the ship cannot respond immediately after adjusting the steering command when maneuvering the ship, the safety distance threshold needs to be set in combination with the actual application requirements and the target ship's motion response delay, and the motion state parameter threshold is set according to the target ship's own motion capacity. Figure 3As shown in the figure, the five-pointed stars represent sub-target points, the circular areas represent high ice areas, and the multiple curves represent the candidate paths generated by the local path planning algorithm. d is the minimum distance between each local path point and all high ice grids in the high ice area. The minimum distance d is also the minimum distance between the target ship and the candidate paths. The solid lines in the figure have a minimum distance from the high ice area that is no less than the safety distance threshold, and their corresponding candidate states are selected. The dashed lines have a minimum distance from the high ice area that is less than the safety distance threshold, and their corresponding candidate states are excluded.

[0084] For any τ-th candidate path whose candidate state is the selected state, based on the safety, dynamic feasibility, and feasibility of ship maneuverability of the target ship in ice breaking, in one embodiment, the cost function C(τ) for determining any τ-th candidate path whose candidate state is the selected state includes the following parts:

[0085] (1) Determine the safety cost f of the τth candidate path based on the minimum distance between each local path point and the high ice condition grid safe (τ).

[0086] The safety cost is used to evaluate the proximity of the candidate path to the high ice area, and is evaluated by comprehensively evaluating the distance between each local path point of the candidate path and the high ice grid. In one embodiment, the safety cost f of the τth candidate path is determined. safe The specific method of (τ) is:

[0087] The location of each high ice condition grid is determined by querying the two-dimensional array of the fine-scale grid map. The distance between any k-th local path point of the τ-th candidate path and all high ice condition grids is then calculated. The minimum distance d between any k-th local path point of the τ-th candidate path and a high ice condition grid is also determined. k ;

[0088] According to the minimum distance d between the kth local path point and the high ice condition grid k , determine the security cost of the τth candidate path The minimum distance d between the kth local path point and the high ice condition grid k The larger it is, the higher the security at the kth local path point, and the lower the security cost contribution of the kth local path point; integer parameter k≥1, integer parameter N≥2.

[0089] (2) Determine the dynamic feasibility cost f of the τth candidate path based on the motion state parameters of the target ship at each local path point dyn (τ).

[0090] The dynamic feasibility cost is used to ensure that the motion state of the target ship meets the ship motion capability constraint. It is comprehensively evaluated by the change of the ship speed and bow turning speed at each local path point of the candidate path relative to the ship speed and bow turning speed at the previous local path point. In one embodiment, the dynamic feasibility cost f of the τth candidate path is determined. dyn The specific method of (τ) is:

[0091] According to the ship speed V of the target ship at any k-th local path point of the τ-th candidate path k and the ship speed V at the k-1th local path point k-1 , determine the speed change ΔV at the kth local path point k ; According to the bow turning speed ω of the target ship at any k-th local path point on the τ-th candidate path k and the bow speed ω at the k-1th local path point k-1 , determine the bow speed change Δω at the kth local path point k ;

[0092] According to the speed change ΔV at the kth local path point k and the bow speed change Δω k , determine the kinetic feasibility cost of the τth candidate path Among them, V max is the maximum speed of the ship, ω max is the maximum bow turning speed. The 0th local path point represents the end point of the i-1th local path.

[0093] (3) Determine the energy efficiency cost f of the τth candidate path based on the motion state parameters of the target ship at each local path point and the ice force on the target ship. energy (τ).

[0094] The energy efficiency cost is used to evaluate the resistance energy efficiency of the target ship. By calculating the resistance energy efficiency at each local path point that constitutes the candidate path, it is ensured that the propulsion energy efficiency provided by the target ship's propulsion system meets the requirements for normal navigation. In one embodiment, the energy efficiency cost f of the τth candidate path is determined. energy The specific method of (τ) is:

[0095] According to the ship speed V of the target ship at any k-th local path point of the τ-th candidate path k and the ice force on the target ship Determine the energy efficiency cost of the τth candidate path Among them, the ice force on the target ship is The main consideration is the component of the ice force in the longitudinal direction calculated in step 3.

[0096] Taking into account the avoidance of high ice conditions, movement ability, and control ability, the cost function of the τth candidate path is determined based on the safety cost, dynamic feasibility cost, and energy efficiency cost:

[0097] C(τ)=ω1f safe (τ)+ω2f dyn (τ)+ω3f energy (τ)

[0098] Where ω1 is the safety cost weight, ω2 is the dynamic feasibility cost weight, and ω3 is the energy efficiency cost weight. Each cost weight is customized according to the priority. In this application, safety is the highest priority, and ω1 = 0.4, ω2 = 0.3, and ω3 = 0.3 are set.

[0099] Step 5: Repeat the steps of generating the initial navigation path and candidate navigation paths until the generated candidate navigation paths meet the navigation requirements of the target ship, and the optimal navigation path is obtained.

[0100] Further analysis of the global characteristics of the candidate navigation path, taking into account the target vessel's navigation time requirements and the need for smoothness during navigation, requires ensuring the smoothness of the candidate navigation path. In one embodiment, the navigation time of the generated candidate navigation path from the navigation starting point to the destination point is determined. When the navigation time meets a duration threshold and the curvature of the candidate navigation path at any point meets a curvature threshold, the candidate navigation path is determined to meet the navigation requirements of the target vessel.

[0101] The purpose of limiting the navigation time is to ensure fast navigation in actual applications, and the time threshold is customized according to the actual application situation. The purpose of limiting the curvature of the navigation path is to avoid frequent turns caused by sudden changes in the trajectory curvature, which is not only not conducive to ship control but also affects the stability of the ship. Among them, the calculation of the curvature of the candidate navigation path at any path point can adopt the curvature calculation method of the existing technology, such as numerical method or geometric method, and the curvature threshold is customized according to the actual application situation. The method of this application has great technical advantages in the overall evaluation of the ship's icebreaking capability, high ice avoidance capability, navigation speed and navigation safety capability.

[0102] The above description is only a preferred embodiment of the present application, and the present application is not limited to the above embodiments. It is understood that other improvements and variations directly derived or imagined by those skilled in the art without departing from the spirit and concept of the present application should be considered to be included in the scope of protection of the present application.

Claims

1. A high ice avoidance path planning method considering the motion characteristics of ships in ice areas, characterized by: The high ice avoidance path planning method includes: Constructing a coarse-scale grid map and a fine-scale grid map, and determining each high ice condition grid in the coarse-scale grid map and the fine-scale grid map respectively; the grid size of the coarse-scale grid map is larger than the grid size of the fine-scale grid map; Based on the coarse-scale grid map, a global path planning algorithm is used to determine an initial navigation path of the target ship from a navigation starting point to a target point; the initial navigation path does not pass through a high ice condition grid of the coarse-scale grid map, and the initial navigation path includes a plurality of global path points from the navigation starting point to the target point; Determine the external excitation force on the target ship, and substitute the external excitation force into the three-degree-of-freedom maneuvering motion equation to solve it, so as to obtain the real-time motion state parameters of the target ship; Based on the fine-scale grid map and the real-time motion state parameters of the target ship, the global path points of the initial navigation path are used as sub-target points of the local path planning algorithm, and multiple local paths are generated using the local path planning algorithm; the local paths are combined to obtain a candidate navigation path for the target ship from the navigation starting point to the target point; The steps of generating the initial navigation path and the candidate navigation paths are repeated until the generated candidate navigation paths meet the navigation requirements of the target ship, thereby obtaining the optimal navigation path.

2. The high ice avoidance path planning method according to claim 1, characterized in that: Each local path corresponds to a sub-target point. The local path from the navigation starting point to the sub-target point closest to the navigation starting point is the first local path constituting the candidate navigation path. The local path from the sub-target point closest to the target point to the target point is the last local path constituting the candidate navigation path. The method of generating any i-th local path using the local path planning algorithm includes: Generate multiple candidate paths from the end point of the i-1th local path to the i-th sub-target point using the local path planning algorithm; Based on the real-time position and real-time motion state parameters of the target ship, the candidate state of each candidate path is determined, and the cost function of each candidate path with the candidate state as the selection state is calculated; the candidate path with the smallest cost function is selected as the i-th local path; the candidate state includes the selection state and the exclusion state, and the candidate path with the candidate state as the exclusion state is directly eliminated; the integer parameter i ≥ 1, the 0th sub-target point is the navigation starting point of the target ship, and the last sub-target point is the target point of the target ship.

3. The high ice avoidance path planning method according to claim 2, characterized in that: The external excitation forces on the target ship include ice force and propeller thrust. Any τ-th candidate path includes N local path points. The candidate states for determining the τ-th candidate path include: Determining a minimum distance between each local path point and the high ice condition grid according to the position of each local path point and the position of each high ice condition grid in the fine-scale grid map; When the minimum distance between a local path point and the high ice condition grid is less than the safety distance threshold, or the motion state parameter of the target ship at the local path point does not meet the motion state parameter threshold, or the ice force acting on the target ship at the local path point is greater than the propeller thrust, the candidate state of the τth candidate path is determined to be an exclusion state; otherwise, the candidate state of the τth candidate path is determined to be a selection state.

4. The high ice avoidance path planning method according to claim 3, characterized in that: The cost function C(τ) for determining any τ-th candidate path that is the selected state includes: Determine the safety cost f of the τth candidate path based on the minimum distance between each local path point and the high ice condition grid safe (τ); Determine the dynamic feasibility cost f of the τth candidate path based on the motion state parameters of the target ship at each local path point dyn (τ); The energy efficiency cost f of the τth candidate path is determined based on the motion state parameters of the target ship at each local path point and the ice force on the target ship. energy (τ); The cost function for determining the τth candidate path is: C(τ)=ω1f safe (τ)+ω2f dyn (τ)+ω3f energy (t) Among them, ω1 is the safety cost weight, ω2 is the dynamic feasibility cost weight, and ω3 is the energy efficiency cost weight.

5. The high ice avoidance path planning method according to claim 4, characterized in that: Determine the security cost f of the τth candidate path safe (τ) includes: Determine the minimum distance d between any k-th local path point of the τ-th candidate path and the high ice condition grid k ; According to the minimum distance d between the kth local path point and the high ice condition grid k , determine the security cost of the τth candidate path 6. The high ice avoidance path planning method according to claim 4, characterized in that: The motion state parameters include the ship speed, and the energy efficiency cost f of the τth candidate path is determined. energy (τ) includes: According to the ship speed V of the target ship at any k-th local path point of the τ-th candidate path k and the ice force on the target ship Determine the energy efficiency cost of the τth candidate path 7. The high ice avoidance path planning method according to claim 4, characterized in that: The motion state parameters include the ship speed and bow turning speed, and the dynamic feasibility cost f of the τth candidate path is determined. dyn (τ) includes: According to the ship speed V of the target ship at any k-th local path point of the τ-th candidate path k and the ship speed V at the k-1th local path point k-1 , determine the speed change ΔV at the kth local path point k ; According to the bow turning speed ω of the target ship at any k-th local path point of the τ-th candidate path k and the bow speed ω at the k-1th local path point k-1 , determine the bow speed change Δω at the kth local path point k ; Determine the kinetic feasibility cost of the τth candidate path Among them, V max is the maximum speed of the ship, ω max is the maximum bow turning speed.

8. The high ice avoidance path planning method according to claim 1, characterized in that: The three-degree-of-freedom control motion equation is constructed based on the ship-ice-water-engine-propeller-rudder multi-system coupled motion model; the constructed three-degree-of-freedom control motion equation is: Calculation of the target ship's real-time motion state parameters includes: According to the target ship's longitudinal velocity u, the target ship's transverse velocity v and the target ship's heading angle ψ, the target ship's velocity component V along the x-axis of the geodetic coordinate system is determined. x =ucosψ-vsinψ, determine the velocity component V of the target ship along the y-axis of the geodetic coordinate system y = usinψ + vcosψ, the velocity component V of the target ship along the x-axis of the geodetic coordinate system is x and the velocity component V of the target ship along the y-axis of the geodetic coordinate system y Perform vector summation to obtain the target ship's ship speed V; determine the target ship's bow speed ω=r based on the target ship's bow angular velocity r; Where m is the target ship mass, m x is the additional mass in the surge direction, m y is the additional mass in the sway direction, I z is the moment of inertia of the target ship rotating around the vertical axis passing through the center of gravity of the target ship, J z is the additional inertia moment of the target ship rotating around the vertical axis passing through the target ship's center of gravity; u is the target ship's surge velocity, v is the target ship's sway velocity, r is the target ship's bow angular velocity, is the surge acceleration of the target ship, is the sway acceleration of the target ship, is the yaw angular acceleration of the target ship; X H is the component of the viscous hydrodynamic force in the longitudinal direction, Y H is the component of the viscous hydrodynamic force in the sway direction, N H is the bow rolling moment component of viscous hydrodynamics, X P is the component of propeller thrust in the surge direction, X R is the component of the rudder force in the longitudinal direction, Y R is the component of the rudder force in the sway direction, N R is the yawing moment component of the rudder force, X ice is the component of the ice force in the longitudinal direction, Y ice is the component of the ice force in the sway direction, N ice is the yaw moment component of the ice force.

9. The high ice avoidance path planning method according to claim 1, characterized in that: Determine the navigation time of the generated candidate navigation path from the navigation starting point to the target point. When the navigation time meets the time threshold and the curvature of the candidate navigation path at any path point meets the curvature threshold, determine that the candidate navigation path meets the navigation requirements of the target ship.

10. The high ice avoidance path planning method according to claim 1, characterized in that: The constructing of the coarse-scale grid map and the fine-scale grid map includes: Acquire an image of the ice-covered sea area where the target ship is sailing, and determine sea ice parameters of the ice-covered sea area where the target ship is sailing; determine the coarse-scale grid size and the fine-scale grid size; Gridding the ice-covered sea area image according to the coarse-scale grid size to obtain a plurality of coarse-scale grids, identifying seawater and sea ice using an image recognition method and determining a high ice condition attribute of each coarse-scale grid based on the sea ice parameters; storing the positions and high ice condition attributes of all coarse-scale grids in an array form to obtain a coarse-scale grid map; The ice zone sea area image is gridded according to the fine-scale grid size to obtain multiple fine-scale grids, the sea water and sea ice are identified using an image recognition method, and the high ice condition attribute of each fine-scale grid is determined according to the sea ice parameters; the position and high ice condition attribute of all fine-scale grids are stored in an array form to obtain a fine-scale grid map.

Images