A dynamic route planning method suitable for polar ships

Abstract

The present invention proposes a dynamic route planning method suitable for polar ships, improves the original D*Lite algorithm, modifies the cost value update rule and path extraction rule of the node, limits the node diffusion update constraint within the local area, correctly identifies the real successor node of the current node, and completes the correct path extraction; the method of the present invention limits the node diffusion update constraint within the local area by modifying the node cost value update rule, uses the method of modifying the underlying local ice condition data to perform dynamic planning, effectively reduces the number of updated nodes during dynamic planning, so that the D*Lite algorithm improved by the method of the present invention can adapt to the large-scale changes in polar ice conditions, and has efficient ship dynamic route planning capabilities in the polar regions; and also ensures that the interference at the junction of the updated area and the non-updated area is excluded during the path extraction process by modifying the path extraction algorithm rule, correctly identifies the real successor node of the current node, and completes the correct path extraction.

Description

Technical Field

[0001] The present invention relates to the technical field of ship navigation planning, and in particular to a dynamic route planning method suitable for polar ships. Background Art

[0002] Since polar ice data are generally provided in gridded form, most studies on ship route planning in polar waters are mainly based on gridded maps. For this reason, graph search algorithms based on node traversal are widely used in studies on ship route planning in polar waters. At the same time, considering the limitations of the search direction of the graph search algorithm, the initial route planned is jagged and does not meet actual requirements. Therefore, smoothing the initial path is a common practice for path planning using graph search algorithms.

[0003] With the development of ice detection technology in polar waters, the geographical accuracy of ice data has been continuously improved. At the same time, the research on ship ice resistance has gradually deepened, and the ship icebreaking performance model has become more complicated. In summary, the ice data publishing format, the complex and changeable environment, and the complexity of node value calculation have made it an inevitable trend to apply dynamic graph search path planning algorithms to polar ship route planning.

[0004] However, the current mainstream dynamic graph search path planning algorithms still have some defects in the application of ship route planning in polar waters. For example, the D* algorithm only triggers updates when new obstacles are added to the current path, and cannot cope with the scenario where the navigable area for ships increases during the polar ablation period. The original D*Lite algorithm is not suitable for scenarios where the map environment changes on a large scale. When the environment changes on a large scale, the update of the original D*Lite algorithm is close to global re-planning and there may be a possibility of path finding failure. Summary of the invention

[0005] The purpose of the present invention is to solve the above problems and to propose a navigation path planning method with high accuracy and applicable to large-scale overall changes in polar ice conditions.

[0006] To achieve the above object, the present invention proposes a dynamic route planning method suitable for polar ships. The method improves the original D*Lite algorithm based on the characteristics of polar ice condition changes, which are holistic and convergent, and re-formulates the local node update rules. The improved D*Lite algorithm can adapt to the characteristics of polar ice condition changes by using modified underlying local ice condition data instead of completely using the latest ice condition data for dynamic planning. The method comprises the following steps:

[0007] S1: Read environmental data, build a grid environment map, calculate the ship navigation risk value of each grid, and obtain the risk threshold , determine the non-navigable area;

[0008] S2: Determine the starting and ending positions, minimize the estimated cost values ​​of all nodes according to the modified edge cost function and heuristic function, and obtain the initial route through the modified path extraction method;

[0009] S3: Optimizing the initial route through a path optimization algorithm to obtain an optimal route;

[0010] S4: Update the unnavigable area grid set based on the latest ice data , and the non-navigable area grid set of the previous round Compare; update the cost values ​​of the nodes in the corresponding area according to the modified node cost value update rules;

[0011] S5: Replan the route and perform optimization processing to obtain the latest optimal route, and repeat step S4 until the ship arrives at the grid where the destination port is located.

[0012] Furthermore, step S1 is specifically as follows:

[0013] S1.1: Read environmental data, including average sea ice thickness, sea ice density and corresponding geographic longitude and latitude data, and convert them into rasterized form; read the current ship's highest continuous ice breaking thickness and the current voyage risk threshold ;

[0014] S1.2: According to the geographical location of the starting port and the destination port, the data is trimmed after the area to which the waterway belongs is determined. Specifically, it can be divided into the Northeast Passage and the Northwest Passage, and only the environmental data of the area to which the waterway belongs is retained;

[0015] S1.3: Calculate all grid navigation risk values: Calculate the navigation risk value r according to the environmental data and the highest continuous ice breaking thickness;

[0016] S1.4: Determine the set of non-navigable areas: Identify the grid set marked as land and islands in the environmental data , and include it in the unnavigable area grid set; according to the highest continuous icebreaking thickness of the ship, the grid set with average sea ice thickness greater than this value Include the unnavigable area grid set; the calculated result is greater than the risk threshold A collection of grids Incorporating the non-navigable area grid set, thereby obtaining the non-navigable area grid set in the grid environment map for:

[0017] ,

[0018] in, Indicates land and islands, Indicates sea areas where the average sea ice thickness exceeds the maximum continuous icebreaking thickness of ships. Indicates sea areas where the navigation risk is higher than the risk threshold.

[0019] Furthermore, in step S1.3, the navigation risk value r is calculated as follows:

[0020] ,

[0021] in, is the average sea ice thickness, is the sea ice density, It is the maximum continuous ice breaking thickness of the ship.

[0022] The calculation formula of the navigation analysis value r mainly considers the maximum continuous ice breaking thickness of the ship. , average sea ice thickness , sea ice density Three factors make all average sea ice thickness The calculated result of the navigation risk in the navigable area that does not exceed the maximum continuous icebreaking thickness of the ship is between [0,1]. The closer the calculated value r is to 0, the lower the navigation risk is; conversely, the closer the value is to 1, the higher the navigation risk is.

[0023] Furthermore, step S2 is specifically as follows:

[0024] S2.1: Set the grid where the current ship position is located as the starting point and the grid where the destination port is located as the end point; calculate the estimated cost value rhs(s) and the actual minimum cost value of all grid points Initialize to infinity; set the estimated cost value rhs(s) of the endpoint to 0 and insert it into the node update queue to complete the initialization.

[0025] S2.2: According to the modified edge cost function and the heuristic function , minimize the estimated cost rhs(s) of all nodes;

[0026] S2.3: Obtain the initial route according to the modified path extraction algorithm. The logic of the modified path extraction algorithm is: select the node with the lowest sum of the actual minimum cost value and the cost value to the current path node from the unexplored nodes around the current path point as the next path point. The specific formula is as follows:

[0027] ,

[0028] in, Indicates the successor node of the current node. Represents the current node, represents the set of unexplored nodes, Represents unexplored nodes around the current node.

[0029] Furthermore, in step S2.2, the modified edge cost function for:

[0030] ,

[0031] The modified heuristic function for:

[0032] ,

[0033] in, , are the risk factor and the weight of the voyage, and They are and The risk value of the node, and Points The horizontal and vertical coordinates of and Points The horizontal and vertical coordinates of and Indicates the horizontal and vertical coordinates of the starting point.

[0034] Furthermore, in step S3, the specific method of optimizing the initial route by using the optimization algorithm is:

[0035] S3.1: Calculate the cumulative cost value cost1 of the initial point and subsequent target node in the current path point list under the initial path. The calculation formula is:

[0036] ,

[0037] in, Represents the i-th node and the i+1th point The cost between

[0038] S3.2: Calculate the cumulative cost value cost2 of the direct connection between the initial point of the current path point list and the subsequent target node. The calculation formula is:

[0039] ,

[0040] in, Indicate point The cost value per unit distance, Indicates that the line is at a grid point The length of the inner

[0041] S3.3: By comparing cost1 and cost2, determine whether there is room for optimization of the route. The judgment method is:

[0042] if When the line between two points does not pass through an obstacle, it means that the cost value of the direct connection between the two nodes is lower than the cost value under the initial path, and there is room for optimization;

[0043] Directly connect the two path nodes and use the next node of the subsequent target node as the new current point;

[0044] Repeat steps S3.1-S3.3, loop through to find nodes with optimization space, and directly connect them for optimization until there are no nodes that can be optimized. The optimized optimal path can be obtained and the route optimization is completed.

[0045] Furthermore, the i-th node and the i+1th point The cost between The calculation formula is:

[0046] ,

[0047] in, Represents the cost value per unit distance of a single node, Represents the cost value per unit distance of the i-th path node; Represents the cost value per unit distance of the i+1th path node.

[0048] Furthermore, step S4 is specifically as follows:

[0049] S4.1: Determine the latest unnavigable area grid set based on the latest ice data And compare it with the unnavigable area grid set in the last round of route update By comparison, a set of all navigable change areas is obtained;

[0050] S4.2: Update the cost value of the corresponding node according to the modified node cost value update rule. The modified node cost value update rule is: when the node that triggers the local update is a newly added obstacle point, the node and the nodes affected by it use the old environment data to update the cost value; if the node that triggers the local update is a newly added navigable point, the cost value of the node is updated using the new environment data, and the nodes affected by it use the old environment data to update the cost value. The specific update rules are as follows:

[0051] ,

[0052] in, Indicates the surrounding nodes affected by the change of the triggering update node, Indicates the latest ice data. Represents the initial historical data after partial modification.

[0053] Furthermore, step S5 is specifically as follows:

[0054] S5.1: Calculate the historical movement cost k based on the current position of the ship and the last updated position m The key value of the node with the highest priority in the current list to be updated is updated locally until the exit condition of the cyclic update is met again. Greater than or equal to the starting point key value and the estimated cost value rhs(s) of the starting point is equal to its actual minimum cost value value;

[0055] S5.2: Use the modified path extraction algorithm to search for a new initial route and optimize it to obtain the latest optimal route, and repeat step S4 until the ship arrives at the grid where the destination port is located.

[0056] Compared with the prior art, the advantages of the present invention are:

[0057] 1. The present invention utilizes the characteristics of polar ice conditions that are holistic and convergent to improve the original D*Lite algorithm. Compared with the original method that uses the latest ice data for dynamic planning, the method of the present invention limits the node diffusion update constraints to the local area by modifying the node cost value update rule, and uses the method of modifying the underlying local ice data for dynamic planning, which effectively reduces the number of nodes updated during dynamic planning, so that the D*Lite algorithm improved by the method of the present invention can adapt to the large-scale changes in polar ice conditions and has efficient ship dynamic route planning capabilities in the polar regions.

[0058] 2. The present invention ensures that the interference at the junction of the updated area and the unupdated area is eliminated during the path extraction process by modifying the path extraction algorithm rules, correctly identifies the real successor node of the current node, and completes the correct path extraction.

[0059] 3. The present invention customizes the calculation formula for the polar navigation risk of ships by taking into account the three major factors of the maximum continuous icebreaking thickness of ships, the average sea ice thickness, and the sea ice density, so that the calculation results of the navigation risk in the navigable areas where the average sea ice thickness does not exceed the maximum continuous icebreaking thickness of ships are between [0,1]. The closer to 0, the lower the navigation risk; conversely, the closer the value is to 1, the higher the navigation risk. The calculation is simple and the result is accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A schematic flow chart of a dynamic route planning method applicable to polar ships proposed in an embodiment of the present invention;

[0061] Figure 2 A schematic diagram comparing the modified path extraction algorithm and the original path extraction algorithm in the method of the embodiment of the present invention;

[0062] Figure 3 This is a flow chart of a modified path extraction algorithm in the method of an embodiment of the present invention;

[0063] Figure 4 A schematic diagram of optimizing initial path cost calculation in a method according to an embodiment of the present invention;

[0064] Figure 5 This is a schematic diagram of the initial path optimization result in the method of the embodiment of the present invention;

[0065] Figure 6 A schematic diagram of the cost of each node of the original map and two maps after changes are simulated and verified by the present invention;

[0066] Figure 7 This is a comparison diagram of the path planning results of the simulated ablation period before and after the modification of the algorithm of the present invention;

[0067] Figure 8 This is a comparison diagram of the path planning results of the simulated freezing period before and after the modification of the algorithm of the present invention;

[0068] Fig. 9 This is a comparison chart of the number of updated nodes in the path after two changes before and after the modification of the algorithm of the present invention. DETAILED DESCRIPTION

[0069] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be further described below.

[0070] This embodiment proposes a dynamic route planning method suitable for polar ships, such as Figure 1 As shown, a dynamic map is generated by environmental data, and then the optimal route in the ice area is obtained by running the improved D*Lite algorithm. The method includes the following steps:

[0071] S1: Read environmental data; construct a grid environmental map, calculate the ship navigation risk value of each grid, and set the risk threshold , determine the non-navigable area. The specific implementation method of step S1 is:

[0072] S1.1: Read data

[0073] Read the required ice data, i.e. the average thickness of sea ice, sea ice density and the corresponding geographic longitude and latitude data, and convert them into rasterized form; read the highest continuous ice breaking thickness value of the ship and the risk threshold of this voyage ;

[0074] S1.2: Data trimming

[0075] The data is trimmed after the area to which the waterway belongs is determined based on the geographical locations of the starting port and the destination port. Specifically, it can be divided into the Northeast Passage and the Northwest Passage, and only the environmental data of the area to which the waterway belongs is retained.

[0076] S1.3: Calculate the navigation risk value for all grids

[0077] The navigation risk value is calculated based on environmental data including average sea ice thickness and sea ice density and the maximum continuous ice breaking thickness parameter for ships. , the calculation formula is:

[0078] ,

[0079] in, is the average sea ice thickness, is the sea ice density, It is the maximum continuous ice breaking thickness of the ship.

[0080] S1.4: Determine the set of unnavigable areas

[0081] Identify raster collections marked as land and islands in environmental data And include it in the unnavigable area grid set; according to the highest continuous ice breaking thickness of the ship, the grid set with average sea ice thickness greater than this value Include the unnavigable area grid set; the calculated result is greater than the risk threshold A collection of grids Incorporate the unnavigable area grid set to obtain the unnavigable area grid set of the grid environment map for:

[0082] ,

[0083] in Indicates land and islands, Indicates sea areas where the average sea ice thickness exceeds the maximum continuous icebreaking thickness of ships. Indicates sea areas where the navigation risk is higher than the risk threshold.

[0084] S2: Determine the starting and ending points, minimize the estimated cost values ​​of all nodes according to the modified edge cost function and heuristic function, and obtain the initial route through the modified path extraction method. The specific method of this step is:

[0085] S2.1: Determine the start and end positions and initialize the cost value

[0086] Set the grid where the current ship position is located as the starting point, and the grid where the destination port is located as the end point; set the estimated cost value of all grid points And the actual minimum cost Initialize to infinity; set the estimated cost value of the end point Set to 0 and insert into the node update queue to complete initialization.

[0087] S2.2: Minimize the value

[0088] According to the modified edge cost function and the heuristic function Minimize the estimated cost of all nodes Value, modified edge cost function and the heuristic function as follows:

[0089] ,

[0090] ,

[0091] in: , are the weight values ​​of risk factor and voyage respectively; and Node and risk value of s; and Represents the horizontal and vertical coordinates of point s; and Representation Node The horizontal and vertical coordinates of and Indicates the horizontal and vertical coordinates of the starting point.

[0092] S2.3: Extract initial route

[0093] The initial route is obtained according to the modified path extraction algorithm. The modified path extraction logic is: from the unexplored nodes around the current path node, select the node with the lowest sum of the actual minimum cost value and the cost value to the current path node as the next path point. The specific formula is as follows:

[0094] ,

[0095] in, Indicates the successor node of the current node. Represents the current node, Represents unexplored nodes around the current node.

[0096] like Figure 3As shown, in this embodiment, the implementation method of the modified path extraction algorithm is: (1) the starting point is included in the path list and used as the current exploration point; (2) the actual minimum cost value of the unexplored nodes around the current exploration point and the sum of the cost to reach the current exploration point are calculated; (3) the node with the smallest sum is included in the path list and used as the new exploration point; (4) the old exploration point and its surrounding nodes are marked as explored; (5) it is determined whether the current exploration point is the end point. If not, return to step (2); if so, continue to step (6); (6) output the path list.

[0097] Compare the modified path extraction algorithm in the method of this embodiment with the original path extraction algorithm results. Figure 2 As shown in the figure, sub-graphs (a) and (c) mark the cost values ​​of different regions, and sub-graphs (b) and (d) mark the actual minimum cost values ​​of reaching the end point in different regions under the corresponding environment. The orange area represents the starting point, and the blue area represents the end point. The cost values ​​of each node in the figure are different. The red line in sub-graph b represents the optimal path extracted by the original path extraction algorithm under the initial environment. After the environmental change triggers the update of the local node cost value, as shown in the green area in the figure, the cost value of the predecessor node after the update may be higher than that of the successor node at the junction of the updated area and the unupdated area. The original path extraction algorithm will fall into an infinite loop in the path extraction, as shown in the red line in sub-graph d. The improved algorithm can successfully extract the optimal path after the environment is updated, as shown in the blue line in sub-graph d.

[0098] S3: Optimize the initial route through a path optimization algorithm to obtain an optimal route. The specific method of this step is:

[0099] S3.1: Calculate the cost between two points on the initial path

[0100] Calculate the cumulative cost of the subsequent target node under the initial path from the initial point of the current path point list , the calculation formula is:

[0101] ,

[0102] in, Indicate point With point The cost between is calculated as:

[0103] ,

[0104] in, Represents the cost value per unit distance of a single node; Represents the cost value per unit distance of the i-th path node; Represents the cost value per unit distance of the i+1th path node.

[0105] S3.2: Calculate the direct cost of two points

[0106] Calculate the cumulative cost of direct connection to subsequent target nodes starting from the initial point of the current path point list , the calculation formula is:

[0107] ,

[0108] in, Indicates that the line is at a grid point The length of the inner

[0109] S3.3: Determine whether there is room for optimization

[0110] if When the line between the two points does not pass through an obstacle, it means that the cost value of the direct connection between the two nodes is lower than the cost value under the initial path, and there is room for optimization. Directly connect the two path nodes and use the next node of the subsequent target node as the new current point. Repeat steps S3-1 to S3-3, loop through and find nodes with optimization space, and optimize their direct connection until there are no nodes that can be optimized. The optimal path after optimization is obtained, and the route optimization is completed;

[0111] The above optimization initial path cost calculation and its optimization results are as follows: Figure 4 and Figure 5 As shown, from Figure 4 It can be seen that the navigation cost value of different grid unit distances is ~ ,The black line represents the original path, the distance between two adjacent points is unit distance, and the path cost is , the blue line represents the optimized path, and the distances through different grids are different, respectively. ~ , and its path cost is . Figure 5 The cost value of each grid unit distance is marked in the figure. The grid with a cost value of unit distance greater than or equal to 0.7 is an obstacle point grid (see Figure 5 The light gray area in the figure), the dark gray in the upper left corner indicates the starting point, the dark gray grid in the lower right corner indicates the end point, the solid line is the original optimal path, and the dotted line is the optimized optimal path.

[0112] S4: Update the unnavigable area grid set based on the latest ice data , and the non-navigable area grid set of the previous round Compare; update the cost values ​​of the nodes in the corresponding area according to the modified node cost value update rules;

[0113] S4.1: Determine the trigger update node

[0114] Determine the latest unnavigable area raster set based on the latest ice data And compare it with the unnavigable area grid set in the last round of route update By comparison, a set of all navigable change areas is obtained;

[0115] S4.2: Local Node Update

[0116] The cost value of the corresponding node is updated according to the modified node cost value update rule. The modified update rule is: when the node that triggers the local update is a newly added obstacle point, the node and the nodes affected by it use the old environment data to update the cost value; if the node that triggers the local update is a newly added navigable point, the cost value of the node is updated using the new environment data, and the nodes affected by it use the old environment data to update the cost value. The expression is as follows:

[0117] ,

[0118] in, Indicates the surrounding nodes affected by the change of the triggering update node, Indicates the latest ice data. Represents the initial historical data after partial modification.

[0119] S5: Replan the route and perform optimization processing to obtain the latest optimal route, and repeat step S4 until the ship arrives at the grid where the destination port is located.

[0120] Based on the current position of the ship and the last updated position Value (historical movement cost), perform local updates until the exit condition of the loop update is met again, that is, the key value of the node with the highest priority in the current list to be updated Greater than or equal to the starting point key value and the starting point The value is equal to value, and then use the modified path extraction algorithm to search for a new initial route and optimize it to obtain the latest optimized route. Repeat step S4 and obtain the latest optimal route until the ship arrives at the grid where the destination port is located.

[0121] Using the method of this embodiment, the cost of each node in the original map and the two changed maps is simulated and verified. The results are as follows: Figure 6 As shown, from Figure 6 As can be seen from the figure, all grids are marked with the unit distance cost value of the grid, and the grids greater than or equal to 0.7 are obstacle point grids, that is, the light gray area in the figure. The dark gray grid in the upper left corner is the starting point, and the dark gray grid in the lower right corner is the end point. Sub-image (a) is compared with Figure 5 , the unit distance cost of all grids is reduced by 0.05, representing the ablation period of the Arctic Ocean. Sub-graph (b) is compared with Figure 5 , the unit distance proxy values ​​of all grids are increased by 0.05, representing the freezing period of the Arctic Ocean.

[0122] The method of this embodiment is used to simulate the path planning during the ablation period and compared with the method before modification. The comparison results are as follows: Figure 7 As shown, from Figure 7 It can be seen that when the starting point and the map environment are improved as a whole, the number of grids (light gray grids) whose cost values ​​need to be updated during the update process of the improved D*Lite algorithm represented by subgraph (a) is significantly lower than the number of grids (light gray grids) whose cost values ​​need to be updated during the update process of the original D*lite algorithm represented by subgraph (b), indicating that the improved D*Lite algorithm can operate effectively in the environmental changes during the ablation period.

[0123] The method of this embodiment is used to simulate the freezing period path planning, and compared with the method before modification. The comparison results are as follows: Figure 8 As shown, from Figure 8 It can be seen that when the starting point and the map environment deteriorate as a whole, the number of grids (light gray grids) whose cost values ​​need to be updated during the update process of the improved D*Lite algorithm represented by sub-graph a is significantly lower than the number of grids (light gray grids) whose cost values ​​need to be updated during the update process of the original D*lite algorithm represented by sub-graph b, indicating that the improved D*Lite algorithm can operate effectively in the environmental changes during the freezing period.

[0124] After experiencing the above-mentioned two changes of the ablation period and the freezing period, the updated number of path update nodes after the method of this embodiment and the original method are compared. The comparison results are as follows: Fig. 9 As shown in the figure, in the simulated map, the number of nodes that need to be updated when updating the path of the improved D*Lite algorithm is significantly reduced compared with the original D*Lite algorithm, whether in the ablation period or the freezing period, which is reduced by 75% and 70% respectively. This shows that the improved D*Lite algorithm can effectively cope with different ice conditions in the Arctic Ocean.

[0125] The above is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any technician in the relevant technical field, without departing from the scope of the technical solution of the present invention, makes any form of equivalent replacement or modification to the technical solution and technical content disclosed in the present invention, which does not depart from the content of the technical solution of the present invention and still falls within the protection scope of the present invention.

Claims

1. A dynamic route planning method suitable for polar ships, characterized in that: The original D*Lite algorithm is improved, the cost value update rule and path extraction rule of the node are modified, the node diffusion update constraint is limited to the local area, the real successor node of the current node is correctly identified, and the correct path extraction is completed; the dynamic route planning method includes the following steps: S1: Read environmental data, build a grid environment map, calculate the ship navigation risk value of each grid, and obtain the risk threshold r max , determine the non-navigable area; S2: Determine the starting and ending positions, minimize the estimated cost values ​​of all nodes according to the modified edge cost function and heuristic function, and obtain the initial route through the modified path extraction method; Step S2 is specifically as follows: S2.1: Set the grid where the current ship position is located as the starting point, and the grid where the destination port is located as the end point; initialize the estimated cost value rhs(s) and the actual minimum cost value g(s) of all grid points to infinity; set the estimated cost value rhs(s) of the end point to 0, and insert it into the node update queue to complete the initialization; S2.2: Minimize the estimated cost rhs(s) of all nodes based on the modified edge cost function c(s',s) and the heuristic function h(s',s); In step S2.2, the modified edge cost function c(s',s) is: Among them, ω1 and ω2 are the weight values ​​of risk coefficient and voyage respectively; r s' and r s are the risk values ​​of nodes s' and s respectively; x s and s represents the horizontal and vertical coordinates of point s; x s' and s' represents the horizontal and vertical coordinates of the node s'; The modified heuristic function h(s',s) is: Among them, x s and s represents the horizontal and vertical coordinates of point s, x start and start Indicates the horizontal and vertical coordinates of the starting point; S2.3: Obtain the initial route according to the modified path extraction algorithm. The logic of the modified path extraction algorithm is to select the node with the lowest sum of the actual minimum cost value and the cost value to the current path node from the unexplored nodes around the current path node as the next path point. The specific formula is as follows: Among them, s new Indicates the successor node of the current node, s last Indicates the current node, s opt Indicates the unexplored nodes around the current node; S3: Optimizing the initial route through a path optimization algorithm to obtain an optimal route; S4: Update the unnavigable area grid set new_obs according to the latest ice data, and compare it with the unnavigable area grid set old_obs of the previous round; update the cost value of the corresponding area node according to the modified node cost value update rule; S5: Replan the route and perform optimization processing to obtain the latest optimal route, and repeat step S4 until the ship arrives at the grid where the destination port is located.

2. The dynamic route planning method for polar ships according to claim 1, characterized in that: Step S1 is specifically as follows: S1.1: Read environmental data, including average sea ice thickness, sea ice density and corresponding geographic longitude and latitude data, and convert them into rasterized form; read the highest continuous ice breaking thickness of the current ship and the risk threshold of the current voyage r max ; S1.2: According to the geographical location of the starting port and the destination port, the area to which the waterway belongs is determined and the data is clipped, and only the environmental data of the area to which the waterway belongs is retained; S1.3: Calculate all grid navigation risk values: Calculate the navigation risk value r according to the environmental data and the highest continuous ice breaking thickness; S1.4: Determine the set of unnavigable areas: Identify the grid set obs marked as land and islands in the environmental data land , and include it in the unnavigable area grid set; according to the highest continuous icebreaking thickness of the ship, the grid set obs with an average sea ice thickness greater than this value ice_thick Include the unnavigable area grid set; the calculated result is greater than the risk threshold r max The grid collection obs risk The non-navigable area grid set Obs in the grid environment map is obtained by incorporating the non-navigable area grid set Obs: Note=note land ∪obs ice_thick ∪obs risk Among them, obs land Indicates land and islands, obs ice_thick Indicates the sea area where the average sea ice thickness exceeds the highest continuous icebreaking thickness of the ship, obs risk Indicates sea areas where the navigation risk is higher than the risk threshold.

3. The dynamic route planning method for polar ships according to claim 2 is characterized in that ,In step S1.3, the navigation risk value r is calculated as follows: Among them, h is the average sea ice thickness, C is the sea ice density, Ice max It is the maximum continuous ice breaking thickness of the ship.

4. The dynamic route planning method for polar ships according to claim 1, characterized in that: In step S3, the specific method of optimizing the initial route by using the optimization algorithm is: S3.1: Calculate the cumulative cost value cost1 of the initial point and subsequent target node in the current path point list under the initial path. The calculation formula is: Among them, cost(S i ,S i+1 ) represents the i-th node S i and the i+1th node S i+1 The cost between S3.2: Calculate the cumulative cost value cost2 of the direct connection between the initial point of the current path point list and the subsequent target node. The calculation formula is: Among them, Cost single S i Indicates point S i Cost per unit distance, d i Indicates that the line is at the grid point S i The length of the inner S3.3: By comparing cost1 and cost2, determine whether there is room for optimization of the route. The judgment method is: 1) If cost2≤cost1 and the line between the two points does not pass through the obstacle point, it means that the cost value of the direct connection between the two nodes is lower than the cost value under the initial path, and there is room for optimization; 2) directly connect the two path nodes and use the next node of the subsequent target node as the new current point; Repeat steps S3.1-S3.3, loop through to find nodes with optimization space, and directly connect them for optimization until there are no nodes that can be optimized. The optimized optimal path can be obtained and the route optimization is completed.

5. The dynamic route planning method for polar ships according to claim 4, characterized in that: The i-th node S i and the i+1th point S i+1 The cost between i ,S i+1 ) is calculated as: Among them, Cost single Represents the cost value per unit distance of a single node.

6. The dynamic route planning method for polar ships according to claim 1, characterized in that: Step S4 is specifically as follows: S4.1: Determine the latest unnavigable area grid set new_obs based on the latest ice data, and compare it with the unnavigable area grid set old_obs during the last round of route update to obtain the set of all navigable change areas; S4.2: Update the cost value of the corresponding node according to the modified node cost value update rule, and the modified node cost value update rule is: when the node that triggers the local update is a newly added obstacle point, the node and the nodes affected by it use the old environmental data to update the cost value; if the node that triggers the local update is a newly added navigable point, the cost value of the node is updated using the new environmental data, and the nodes affected by it use the old environmental data to update the cost value.

7. The dynamic route planning method for polar ships according to claim 1, characterized in that: Step S5 is specifically as follows: S5.1: Calculate the historical movement cost k based on the current position of the ship and the last updated position m The value is locally updated until the exit condition of the cyclic update is met again. The exit condition of the cyclic update is: the key value k of the highest priority node in the current list to be updated is greater than or equal to the starting point key value and the estimated cost value rhs(s) of the starting point is equal to its actual minimum cost value g(s); S5.2: Use the modified path extraction algorithm to search for a new initial route and optimize it to obtain the latest optimal route, and repeat step S4 until the ship arrives at the grid where the destination port is located.

Images