Path planning and auxiliary decision-making method for observation task of underwater unmanned vehicle

Abstract

The invention provides a path planning and auxiliary decision-making method for an observation task of an underwater unmanned vehicle, which belongs to the field of underwater path planning and auxiliary decision-making and mainly comprises the following steps of: acquiring sea area information, observing task content, analyzing the task content, selecting an observation interest point access sequence and an intermediate path point, and selecting an observation interest point access sequence; calculating a global path and a local path according to the path point position, changing the observation point when the risk of the observation point position rises, and calculating that the hidden path leaves the current sea area after the observation task is completed. According to the method, compared with single-scale zoning, the dual-scale zoning can achieve the effects of improving the path accuracy and reducing the calculated amount, and compared with a single-path planning system, the dual-scale zoning can automatically make decisions, select proper interest point access paths, observation points, sea area entrances and the like and select proper evacuation opportunities; and the observation efficiency and the course safety of the unmanned underwater vehicle are improved.

Description

Technical Field

[0001] The present invention belongs to the field of underwater path planning and auxiliary decision-making, and specifically is a path planning and auxiliary decision-making method for underwater unmanned vehicle observation tasks. Background Art

[0002] The navigation safety of underwater unmanned vehicles is of vital importance in modern naval operations. Underwater unmanned vehicle operations have become an important means of naval operations. As a concealed underwater combat platform, it performs reconnaissance, attack and defense missions. Its success or failure is directly related to national security and the implementation of military strategies. Underwater unmanned vehicles need to operate in a vast and complex marine environment. Faced with changing marine conditions and potential threats, navigation safety is particularly important. Ensuring the navigation safety of underwater unmanned vehicles involves not only the accuracy of the navigation system, but also real-time monitoring of the marine environment, prediction and avoidance of potential threats, and intelligent decision-making for route planning. Any negligence in navigation safety may lead to mission failure and even serious loss of personnel and equipment. Therefore, the research on technical means and strategies to improve the navigation safety of underwater unmanned vehicles is an important part of ensuring naval combat capabilities and national security.

[0003] The marine risk gridding method is a technology used to improve the navigation safety of underwater unmanned vehicles. It divides the marine environment into multiple small grids, each of which represents a specific geographical area and environmental conditions. Various risk factors are evaluated and annotated in each grid, such as water depth, current, temperature, fish activity, and other potential threats. Through this refined gridding method, the navigation system of underwater unmanned vehicles can more accurately identify and evaluate the risk level of different areas, thereby developing a safer and more efficient navigation path. However, the traditional marine risk gridding zoning method also has some shortcomings. The traditional gridding method usually adopts a fixed grid size and cannot be flexibly adjusted to adapt to the detailed requirements of different areas, which may lead to insufficiently refined risk assessment in some key areas. In addition, the traditional method mainly relies on static data, has insufficient response capabilities to real-time environmental changes, and cannot promptly reflect new risk factors in the dynamic marine environment.

[0004] The underwater unmanned vehicle observation mission is highly complex, including single point of interest observation, multiple point of interest observation and regional observation. Single point of interest observation requires underwater unmanned vehicles to covertly approach and monitor or attack specific targets, which requires precise path planning and concealment strategies; multiple point of interest observation involves multiple targets, and underwater unmanned vehicles need to move between multiple locations, which increases the complexity and requires efficient path selection and dynamic adjustment; regional observation requires underwater unmanned vehicles to be active in a specific area for a long time, and a comprehensive risk assessment and environmental adaptation of the entire area are required to ensure continuous concealment and flexibility in mission execution. These mission forms require underwater unmanned vehicles to have excellent navigation, concealment and real-time decision-making capabilities. This places strict requirements on path planning and decision-making during the observation voyage of underwater unmanned vehicles.

[0005] In recent years, research on underwater unmanned vehicle path planning has made significant progress, especially the introduction of intelligent algorithms. Ant colony algorithms, genetic algorithms, and deep learning technologies have optimized path selection and improved planning accuracy and efficiency. For example, the ant colony algorithm simulates the foraging behavior of ants to find the shortest path and avoid high-risk areas; the genetic algorithm selects the optimal path through natural selection; and the deep learning technology analyzes historical and real-time data to predict environmental changes and dynamically adjust the path. However, these intelligent algorithms also have defects. They have high computational complexity and may not be real-time enough when processing large-scale ocean data. In addition, intelligent algorithms rely on a large amount of training data and computing resources, and may be limited by data acquisition and processing capabilities in applications. The "black box" characteristics of deep learning models reduce the transparency and interpretability of path decisions.

[0006] When performing observation missions, underwater unmanned vehicles face highly complex and dynamic environments, which makes the decision-making process extremely difficult. For example, during the observation process, underwater unmanned vehicles may encounter situations where the risk is abnormally high, such as sudden changes in weather that threaten the safety of the vehicle. At this time, the underwater unmanned vehicle must quickly change the observation point and replan the path to avoid the threat. Such decisions must be made in a very short time, and any hesitation may lead to mission failure or exposure of the underwater unmanned vehicle. The complexity of these decisions shows that underwater unmanned vehicles need a highly intelligent and real-time responsive path planning system in observation missions to ensure mission success while maximizing their own safety. Summary of the invention

[0007] In order to cope with the computational complexity and decision-making complexity of path planning in underwater unmanned vehicle observation tasks, the present invention proposes a path planning and auxiliary decision-making method for underwater unmanned vehicle observation tasks, which divides path planning into two steps to ensure the calculation speed and path accuracy, and at the same time makes systematic decision-making suggestions for underwater unmanned vehicle observation tasks and covert evacuation until the underwater unmanned vehicle completes the observation mission and leaves the dangerous waters.

[0008] The present invention divides the target sea area into two-scale risk zones, uses the small block path planning strategy of node-to-node path planning + dynamic planning to perform fast and accurate path planning, and makes systematic decision suggestions for underwater unmanned vehicle observation tasks. The specific steps include:

[0009] Step 1: Obtain observation mission information and risk zoning;

[0010] The observation mission information includes the current position of the underwater unmanned vehicle, mission content, longitude and latitude range of the target sea area, and the concealment value risk and security risk value information corresponding to the target sea area;

[0011] Perform dual-scale risk zoning within the target sea area, the dual-scale risk zoning includes local zoning and global zoning, the global zoning is composed of several global blocks, and each global block is composed of several local blocks; and add corresponding concealment risk values ​​and security risk values ​​to each local block and each global block;

[0012] Step 2: According to different observation task types, the observation point position of the underwater unmanned vehicle is determined in the local division, and the global path planning from the current underwater unmanned vehicle position to the observation point position is performed in the global division, and the path with the lowest path cost is selected as the global path;

[0013] Step 3: Use inter-node path planning and dynamic planning to plan the path of local blocks. Specifically, plan multiple local paths within each global block in the global path, select corresponding local blocks to connect between global blocks, and use dynamic planning algorithm to select the continuous path with the lowest risk from all local paths as the precise path;

[0014] Step 4: The underwater unmanned vehicle follows a precise path to arrive at the observation point for observation and obtains the surrounding concealment risk value;

[0015] Step 5: Detect the hidden risk value Rc of the current observation point i Is it greater than or equal to the danger identification threshold Rc0? If it is, it means that the surrounding risk is increasing, and go to step 6 to change the observation point. If it is less than, it means that the surrounding risk is normal. After completing the observation task of the current observation point, go to step 7 to continue the observation task;

[0016] Step 6: Reselect the local block with the lowest risk around the current observation point as the new observation point and plan the path, and go to step 4 to continue executing the replanned path;

[0017] Step 7: Get the current observation task status. If the current observation point is not the last observation point, that is, all observation tasks have not been completed, go to the next observation point to continue the current observation task and go to step 4. If the current underwater unmanned vehicle position is the last observation point, that is, all observation tasks have been completed, go to step 8 to covertly evacuate the current sea area.

[0018] Step 8. When performing a covert evacuation, specify the longitude and latitude boundaries of the area to be passed and the evacuation direction, use the path planning algorithm to generate a path from the location of the underwater unmanned vehicle, select the global path with the lowest total path cost, use inter-node path planning and dynamic planning to perform local block path planning to accurately calculate the evacuation path, and execute the evacuation.

[0019] Furthermore, in step 1, the dual-scale risk zoning division method is:

[0020] The size of the local block is determined according to the size of the underwater unmanned vehicle, and risk zoning is performed using large and small grids. The large grid corresponds to the global block, and the small grid corresponds to the local block. The small grid is nested in the large grid, and the hidden risk and security risk information of each local block and global block is obtained to generate a multi-dimensional risk map.

[0021] Furthermore, in step 2, the observation task types include single point of interest observation tasks, multiple point of interest observation tasks and regional observation tasks.

[0022] Furthermore, for the single point of interest observation task, the global path planning method is as follows:

[0023] S101, obtaining the most concealed local block around the point of interest in the local division as the end point of the path, that is, the observation point position;

[0024] S102. Select multiple entrances on the boundary of the area where the current underwater unmanned vehicle is located for path planning, and use the path planning algorithm to calculate the distance from each entrance to the end point of the path and the path risk as the path cost; select the entry point with the lowest path cost as the sea area entrance S1, and obtain the planned path R1{S1, P1, P2, P3...} in the global scale division, where P1, P2, P3... are the middle points on the path.

[0025] Furthermore, for the multi-POI observation task, the global path planning method is as follows:

[0026] S201, obtaining the most concealed local block around each point of interest in the local division as the observation point position of each point of interest;

[0027] S202, planning the paths between all observation points on the global division, and constructing a fully connected graph including all observation points; solving the order of visiting points of interest as a classic TSP problem on the basis of the fully connected graph, and obtaining the node visiting order L{n1,n2,n3...}, where n1,n2,n3,... are the observation points visited in sequence by the underwater unmanned vehicle;

[0028] S203. Select multiple entrances on the boundary of the area where the current underwater unmanned vehicle is located to perform path planning to n1. Through the path planning algorithm, select the entry point with the lowest path cost as the sea area entrance S2, and obtain the path R2{S2, P1, P2, P3...} planned in the global division, where P1, P2, P3... are the middle points on the path.

[0029] Furthermore, for regional observation tasks, the global path planning method is as follows:

[0030] S301, selecting the local block with the best concealment in the global block with the best concealment as the end point of the path, that is, the observation point position;

[0031] S302. Select multiple entrances on the boundary of the area where the current underwater unmanned vehicle is located to plan the path to the end point of the path. Through the path planning algorithm, select the entry point with the lowest path cost as the sea area entrance S3, and obtain the planned path R3 {S3, P1, P2, P3...} on the global division, where P1, P2, P3... are the middle points on the path.

[0032] Furthermore, the specific steps of step three are as follows:

[0033] Step 1), for each path point P in the global path R n Generate multiple possible local paths in the corresponding local block, each local path starts from the current path point P n Go to the next waypoint P n+1 , calculate the risk of each path;

[0034] Step 2), for each global path point P n , select 3 out-edge candidate points and 3 in-edge candidate points to form several possible path combinations, use the dynamic programming algorithm to evaluate these local paths, and select the path with the lowest comprehensive risk among all possible local path combinations as the exact path.

[0035] Compared with traditional methods, the present invention uses dual-scale zoning for path planning, uses a path evaluation algorithm to automatically select path points, and dynamically changes the path according to the current sea area risk, to achieve automatic planning and automatic decision-making of the underwater unmanned vehicle's navigation path. The present invention improves the accuracy of path planning, reduces the amount of calculation, and at the same time enhances the automatic decision-making ability, optimizing the observation efficiency and navigation safety of the underwater unmanned vehicle.

[0036] 1) The dual-scale risk zoning can accurately describe the risk status of the target sea area and improve the calculation efficiency. The small block path planning strategy of inter-node path planning + dynamic planning can refine the navigation path of the underwater unmanned vehicle, and the system provides a large number of decision-making directions for underwater unmanned vehicle observation tasks. Therefore, the present invention can fully adapt to the complexity of current underwater unmanned vehicle observation tasks.

[0037] 2) During the execution of the mission, the present invention can acquire and analyze sea area information in real time, optimize the selection of observation points of interest and intermediate path points, and dynamically adjust the observation points to cope with increased risks, thereby ensuring the efficiency and safety of the mission. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A schematic diagram of a path planning and decision-making assistance method for underwater unmanned vehicle observation tasks according to the present invention;

[0039] Figure 2 It is a schematic diagram of dual-scale grid division;

[0040] Figure 3 is the outer grid (global) risk distribution map;

[0041] Figure 4 This is the inner grid (local) risk distribution map;

[0042] Figure 5 This is a schematic diagram of global path planning;

[0043] Figure 6 A schematic diagram of selecting adjacent points between global blocks to connect corresponding local blocks in local path planning;

[0044] Figure 7 In local path planning, after the adjacent points are determined, the local block is further refined to generate an accurate path diagram. Specific embodiments

[0045] The present invention provides a path planning and auxiliary decision-making method for underwater unmanned vehicle observation tasks. The method of the present invention analyzes the needs of three tasks, performs path planning and auxiliary decision-making, and completes the tasks and leaves the sea area, thereby realizing the full process planning of the observation task, aiming to improve the real-time performance of underwater unmanned vehicles in dealing with risks and responding. The flow chart of the method of the present invention is as follows: Figure 1 shown.

[0046] The present embodiment provides a path planning and decision-making assistance method for underwater unmanned vehicle observation tasks, comprising the following steps:

[0047] Step 1: Obtain observation mission information and risk zoning.

[0048] 1) Obtaining observation mission information, including the current position of the underwater unmanned vehicle, mission content, longitude and latitude range of the target sea area, and the concealment value risk and security risk value information corresponding to the target sea area;

[0049] Assume that the underwater unmanned vehicle is currently located at Q (25.1°N, 121.5°E), and its mission is to monitor and reconnaissance the target sea area between 24.5000°N and 25.5000°N, and 101.0000°E and 122.0000°E. Calculate the concealment risk and security risk of each area in the sea area based on intelligence data. For example, the concealment risk of the enemy radar coverage area is 0.8, and the security risk is 0.9.

[0050] 2) Risk zoning:

[0051] A dual-scale risk zoning is performed within the target sea area, and the dual-scale risk zoning includes a local zoning and a global zoning. The global zoning is composed of several global blocks, and each global block is composed of several local blocks; and corresponding concealment risk values ​​and safety risk values ​​are added to each local block and each global block. The division method of the dual-scale risk zoning is:

[0052] The size of the local block is determined according to the size of the underwater unmanned vehicle, and risk zoning is performed using large and small grids. The large grid corresponds to the global block, and the small grid corresponds to the local block. The small grid is nested in the large grid, and the concealment risk and security risk information of each local block and the global block is obtained to generate a multi-dimensional risk map. For example, if the size of the local block is set to 1.5 times the length of the underwater unmanned vehicle, each global block is set to contain 10*10 local blocks.

[0053] Assuming that the length of the underwater unmanned vehicle is 90 meters, a dual-scale risk zoning is performed based on the length of the underwater unmanned vehicle. The size of the local block is set to 1.5 times the length of the underwater unmanned vehicle, that is, 135 meters × 135 meters. 10 × 10 local blocks form a global block (1350 meters × 1350 meters). And assign a concealment risk value and a safety risk value to each block to support path planning. The schematic diagram of the dual-scale risk zoning is shown in the figure below. Figure 2-Figure 3 As shown, Figure 2-Figure 3 In the figure, each color block represents the cost of the current block, which is calculated by the comprehensive calculation of the concealment risk and security risk. Figure 2 This is a schematic diagram of the dual-scale grid division method. Figure 3 Map the outer grid (global) risk zones, Figure 4 Map the inner grid (local) risk zones.

[0054] Step 2: According to different observation task types, determine the observation point location of the underwater unmanned vehicle in the local division, perform global path planning from the current underwater unmanned vehicle location to the observation point location in the global division, and select the path with the lowest path cost as the global path. The global path planning diagram is shown in the figure below. Figure 5 shown.

[0055] According to the task type, such as single point of interest observation task, multiple point of interest observation task or regional observation task, select the corresponding path planning process.

[0056] 1) For the single point of interest observation task, the global path planning method is as follows:

[0057] S101. In the local division, the local block with the highest concealment in the area with a distance of 1 around the point of interest is obtained as the end point of the path, that is, the observation point location; this process ensures that the underwater unmanned vehicle can select a point with the best concealment as the final observation point location, thereby minimizing the risk of being discovered.

[0058] Assume that the current position Q of the underwater unmanned vehicle is (35.0°N, 120.0°E), and the task requires observation near the point of interest I. Calculate the concealment risk in the local area around the point of interest I. Assume that the local block with the lowest concealment risk is selected as the observation point d1, and the location is (34.8°N, 120.2°E).

[0059] S102: After determining the path end point d1, perform global path planning on the global region.

[0060] On the boundary of the area where the underwater unmanned vehicle is currently located, 10 entry points are randomly selected, such as S1, S2, ..., S 10, and path planning is performed. These entry points will be used to find the optimal path for the underwater unmanned vehicle to reach the path end point d1 from the current position Q. The path planning algorithm is used to calculate the distance from each entry to the path end point and the path risk as the path cost; the entry point with the lowest path cost is selected as the sea area entrance S1, and finally the global path R1{S1,P1,P2,P3,…} from Q to d1 is determined, where P1,P2,P3… are the middle points on the path.

[0061] Use the path planning algorithm to calculate the distance from each entrance to the end point of the path and the path risk as the path cost Path_cost. The specific algorithm process is as follows:

[0062] The path planning algorithm needs to consider both the safety risk Rs and the concealment risk Rc in the sea area, and combine these risk information to construct the cost of a certain block. i,j . Its cost can be expressed as:

[0063] cost i,j =w s ·Rs i,j +w c ·Rc i,j

[0064] Among them, w s and w c is the risk weight, which indicates the importance of hidden risks and security risks. i,j Represents the hidden risk of the block with subscript (i,j), Rs i,j Indicates the security risk of the block with subscript (i, j). When calculating the global path, the subscript i, j represents the index of the global block. When calculating the local path, the subscript i, j represents the index of the local block. The global block and the local block have their own security risk Rs and hidden risk Rc. The specific risk zoning method is as follows: Figure 2-Figure 4 shown.

[0065] Therefore, the overall path cost Path_cost of a path can be expressed as:

[0066]

[0067] 2) For multi-point-of-interest observation tasks, it is necessary to plan paths for multiple observation points and ensure that the underwater unmanned vehicle can effectively move between the observation points to complete the task. The method of global path planning for multi-point-of-interest observation tasks is as follows:

[0068] S201, determine the best observation point position for each point of interest on a local scale. Specifically, in the area with a distance of 1 around each point of interest, calculate the concealment risk of each local block, and select the point with the lowest concealment risk as the observation point. The selection of each observation point ensures that the underwater unmanned vehicle completes the observation task under the safest conditions.

[0069] S202. Next, plan the paths between all observation points on the global division, and construct a fully connected graph containing all observation points. In this graph, nodes represent observation points, and edges represent possible paths between observation points. In order to optimize the order in which the underwater unmanned vehicle visits each observation point, the path planning problem between observation points is converted into the classic traveling salesman problem (TSP). TSP is a problem of finding the shortest path. The goal is to find a path that minimizes the total path cost for the underwater unmanned vehicle to visit all observation points. By solving TSP, the visit order of the observation points L{n1,n2,n3,…} is obtained, where n1,n2,n3,… are the observation points visited in sequence by the underwater unmanned vehicle.

[0070] Assume that the task of the underwater unmanned vehicle is to observe near three points of interest I1, I2 and I3. First, in the area with a distance of 1 around each point of interest, the concealment risk is calculated, and the local block with the lowest concealment risk is selected as the observation points d1, d2 and d3 of the three points of interest. Assume that the coordinates of these three observation points are d1 = (34.8°N, 120.2°E), d2 = (35.2°N, 120.4°E) and d3 = (34.9°N, 120.6°E). Next, a fully connected graph containing d1, d2 and d3 is constructed on a global scale, and the cost cost is calculated for each path between the observation points. The observation point access order problem is converted into a TSP problem, and the optimal order of underwater unmanned vehicles to visit observation points is obtained. L{n1 = d2, n2 = d1, n3 = d3}, that is, the underwater unmanned vehicle first visits d2, then d1, and finally d3.

[0071] The path cost calculation method is the same as the cost calculation method in S102.

[0072] This step ensures that the underwater unmanned vehicle can visit all observation points in the optimal path sequence in the multi-point-of-interest observation mission, thereby minimizing the risk and increasing the success rate of the mission.

[0073] S203, similar to S102, assume that the current position Q of the underwater unmanned vehicle is (35.0°N, 120.0°E), and the task requires the underwater unmanned vehicle to first reach the observation point n1, whose coordinates are (34.8°N, 120.2°E). Five entry points are randomly selected on the boundary of the area where Q is located, namely S1, S2, S3, S4, and S5. For each entry point, a path planning algorithm is used to calculate the path cost from the entry point to n1. For example, the path cost of entry point S1 may be 10, the path cost of S2 may be 8, and so on. Assuming that the calculation shows that the entry point S2 has the lowest cost, S2 is selected as the optimal entry point. Finally, the global scale path from S2 to n1 is determined to be R2{S2, P1, P2, P3, …}, where the path contains several intermediate points P1, P2, P3, … to ensure that the underwater unmanned vehicle can successfully reach n1. Then n2, n3, … are visited in sequence.

[0074] Use the path planning algorithm to calculate the global path with the lowest path cost under the L access order, such as n1→n2, n2→n3.

[0075] The path cost calculation method is the same as the cost calculation method in S102.

[0076] 3) For regional observation tasks, the global path planning method is as follows:

[0077] S301. In the regional observation task, the final observation position of the underwater unmanned vehicle needs to be determined so that the underwater unmanned vehicle can perform covert operations in the designated sea area. To this end, the best concealment area is selected on a global scale, and further refined within the area to select the best concealment local block as the final path end point d3.

[0078] Specifically, firstly, the hidden risk Rc of the entire sea area is analyzed on a global scale. i,j The sea area is divided into multiple global blocks, and the hidden risk Rc of each global block is calculated. i,j , and select the global block with the smallest concealment risk (i.e. the strongest concealment). Then, within the selected global block, further concealment risk analysis is performed on the local blocks. The system calculates the concealment risk Rc of each local block i,j , and select the local block with the lowest concealment risk as the final observed path end point d3.

[0079] Assume that the mission requires the underwater unmanned vehicle to perform a regional observation mission in a certain sea area from (34.5°N, 119.5°E) to (35.5°N, 120.5°E). The system first divides the entire sea area into multiple global blocks and calculates the concealment risk of each global block. Assume that after analysis, the system finds that the global block with the lowest concealment risk in the sea area is G3, which has the lowest concealment risk. Then, G3 is further divided into multiple local blocks, and the concealment risks of these local blocks are calculated. Assume that within G3, the concealment risk of local block d3 is the lowest, so d3 is selected as the final observation position. Finally, the position of d3 is determined as the path end point d3 = (34.8°N, 120.3°E). The underwater unmanned vehicle will perform the covert observation mission at this position to ensure that the most concealed position is selected in the regional observation mission, thereby effectively reducing the risk of being detected by the enemy.

[0080] S302: Similar to the path planning for the single point of interest observation task in S102, the most suitable sea area entrance S3 is selected through the path planning algorithm, and the path R3 {S3, P1, P2, P3...} planned on a global scale is obtained, where P1, P2, P3... are the intermediate points on the path.

[0081] Step 3: Use inter-node path planning and dynamic planning to perform local block path planning, such as Figure 6 and Figure 7 As shown, Figure 6 Select a schematic diagram for the neighbor point, Figure 7 Generate a schematic diagram for an exact path.

[0082] Multiple local paths {r1, r2, r3...} are planned within each global block in the global path. Corresponding local blocks are selected to connect the global blocks. The dynamic programming algorithm is used to select the path with the lowest continuous path cost from all local paths as the precise path.

[0083] In the observation mission, the global path needs to be further refined to generate an accurate path to ensure that the underwater unmanned vehicle can safely and covertly pass through each global block area. The core of this step is to select a continuous path with the lowest comprehensive risk of the entire path through local path planning.

[0084] The specific steps of step three are as follows:

[0085] Step 1), first, at each path point P in the global path R n Multiple possible local paths are generated in the corresponding local block. Each local path starts from the current path point P n Go to the next waypoint P n+1, and the risk of each path is determined by the calculation method of the path cost in S102.

[0086] Step 2), for each global path point P n , select 3 candidate outgoing points and 3 candidate incoming points to form several possible path combinations. Then, a dynamic programming algorithm is used to evaluate these local paths, with the goal of selecting a continuous path with the lowest comprehensive risk. In other words, it is not only the risk of a single local path that is considered, but the cumulative risk of all connected local paths on the entire path chain. Finally, a path with the lowest comprehensive risk is selected from all possible local path combinations as the precise path. This precise path r will start from the initial entry point, pass through all planned path points, and finally reach the observation end point d3 of the underwater unmanned vehicle.

[0087] For example, the global path R of the underwater unmanned vehicle contains the following path points: P1 = (34.9°N, 120.1°E), P2 = (34.8°N, 120.2°E), P3 = (34.7°N, 120.3°E). In the local grid between path points P1 and P2, the system generates three local paths r1, r2, and r3. Similarly, three local paths are generated between path points P2 and P3. All possible path combinations from P1 to P3 are evaluated through the dynamic programming algorithm. For example, the following possible path combinations are evaluated:

[0088] 1.

[0089] 2.

[0090] 3.

[0091] Select the combined path with the lowest comprehensive risk. Assuming that the combination r2+r5 is selected, this path will be used as the final precise path for the underwater unmanned vehicle.

[0092] Step 4: Execute the planned precise path, conduct observations after arriving at the observation point, use the passive equipment of the underwater unmanned vehicle to obtain surrounding information and calculate the concealment risk value Rc of the local block where the underwater unmanned vehicle is currently located i .

[0093] Step 5: Detect the hidden risk value Rc of the current observation point i Is it greater than or equal to the danger identification threshold Rc0? If so, it means that the surrounding risk is rising abnormally, and go to step 6 to change the observation point; if less than, it means that the surrounding risk is normal, and after completing the observation task of the current observation point, go to step 7 to continue the observation task.

[0094] Step 6: Reselect the local area with the lowest risk around the current observation point as the new observation point and plan the path, then go to step 4 to continue executing the replanned path.

[0095] Step 7. Get the current observation task status. If the current underwater unmanned vehicle position is not the last observation point, that is, all observation tasks have not been completed, go to step 4 and go to the next observation point to continue the current observation task. If the current underwater unmanned vehicle position is the last observation point, that is, all observation tasks have been completed, go to step 8 to covertly evacuate the current sea area.

[0096] Step 8. Execute covert evacuation. Specify the longitude and latitude boundaries of the area to be passed and the evacuation direction. From the current position of the underwater unmanned vehicle, use the path planning algorithm to generate the global block path R4 with the lowest overall path cost. Use the small block path planning strategy of inter-node path planning + dynamic planning to accurately calculate the evacuation path and execute the evacuation.

[0097] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes in form and details may be made without departing from the spirit and scope of the present invention.

Claims

1. A path planning and decision-making assistance method for underwater unmanned vehicle observation tasks, characterized in that: The specific steps are as follows: Step 1: Obtain observation mission information and risk zoning; The observation mission information includes the current position of the underwater unmanned vehicle, mission content, longitude and latitude range of the target sea area, and the concealment value risk and security risk value information corresponding to the target sea area; Perform dual-scale risk zoning within the target sea area, the dual-scale risk zoning includes local zoning and global zoning, the global zoning is composed of several global blocks, and each global block is composed of several local blocks; and add corresponding concealment risk values ​​and security risk values ​​to each local block and each global block; Step 2: According to different observation task types, the observation point position of the underwater unmanned vehicle is determined in the local division, and the global path planning from the current underwater unmanned vehicle position to the observation point position is performed in the global division, and the path with the lowest path cost is selected as the global path; Step 3: Use inter-node path planning and dynamic planning to plan the path of local blocks. Specifically, plan multiple local paths within each global block in the global path, select corresponding local blocks to connect between global blocks, and use dynamic planning algorithm to select the continuous path with the lowest risk from all local paths as the precise path; Step 4: The underwater unmanned vehicle follows a precise path to arrive at the observation point for observation and obtains the surrounding concealment risk value; Step 5: Detect the hidden risk value Rc of the current observation point i Is it greater than or equal to the danger identification threshold Rc0? If it is, it means that the surrounding risk is increasing, and go to step 6 to change the observation point. If it is less than, it means that the surrounding risk is normal. After completing the observation task of the current observation point, go to step 7 to continue the observation task; Step 6: Reselect the local block with the lowest risk around the current observation point as the new observation point and plan the path, and go to step 4 to continue executing the replanned path; Step 7: Get the current observation task status. If the current observation point is not the last observation point, that is, all observation tasks have not been completed, go to the next observation point to continue the current observation task and go to step 4. If the current underwater unmanned vehicle position is the last observation point, that is, all observation tasks have been completed, go to step 8 to covertly evacuate the current sea area. Step 8. When performing a covert evacuation, specify the longitude and latitude boundaries of the area to be passed and the evacuation direction, use the path planning algorithm to generate a path from the location of the underwater unmanned vehicle, select the global path with the lowest total path cost, use inter-node path planning and dynamic planning to perform local block path planning to accurately calculate the evacuation path, and execute the evacuation.

2. The method according to claim 1, characterized in that In step 1, the dual-scale risk zoning is divided as follows: The size of the local block is determined according to the size of the underwater unmanned vehicle, and risk zoning is performed using large and small grids. The large grid corresponds to the global block, and the small grid corresponds to the local block. The small grid is nested in the large grid, and the hidden risk and security risk information of each local block and global block is obtained to generate a multi-dimensional risk map.

3. The method according to claim 1, characterized in that In step 2, the observation task types include single point of interest observation tasks, multiple point of interest observation tasks and regional observation tasks.

4. The method according to claim 3, characterized in that For the single point of interest observation task, the global path planning method is as follows: S101, obtaining the most concealed local block around the point of interest in the local division as the end point of the path, that is, the observation point position; S102, selecting multiple entrances on the boundary of the area where the current underwater unmanned vehicle is located for path planning, and using a path planning algorithm to calculate the distance from each entrance to the end point of the path and the path risk as the path cost; The entry point with the lowest path cost is selected as the sea area entrance S1, and the path R1{S1, P1, P2, P3...} planned in the global scale zoning is obtained, where P1, P2, P3... are the middle points on the path.

5. The method according to claim 3, characterized in that: For multi-POI observation tasks, the global path planning method is as follows: S201, obtaining the most concealed local block around each point of interest in the local division as the observation point position of each point of interest; S202, planning the paths between all observation points on the global division, and constructing a fully connected graph including all observation points; solving the order of visiting points of interest as a classic TSP problem on the basis of the fully connected graph, and obtaining the node visiting order L{n1,n2,n3...}, where n1,n2,n3,... are the observation points visited in sequence by the underwater unmanned vehicle; S203. Select multiple entrances on the boundary of the area where the current underwater unmanned vehicle is located to perform path planning to n1. Through the path planning algorithm, select the entry point with the lowest path cost as the sea area entrance S2, and obtain the path R2{S2, P1, P2, P3...} planned in the global division, where P1, P2, P3... are the middle points on the path.

6. The method according to claim 3, characterized in that: For regional observation tasks, the global path planning method is as follows: S301, selecting the local block with the best concealment in the global block with the best concealment as the end point of the path, that is, the observation point position; S302. Select multiple entrances on the boundary of the area where the current underwater unmanned vehicle is located to plan the path to the end point of the path. Through the path planning algorithm, select the entry point with the lowest path cost as the sea area entrance S3, and obtain the planned path R3 {S3, P1, P2, P3...} on the global division, where P1, P2, P3... are the middle points on the path.

7. The method according to claim 1, characterized in that The specific steps of step three are as follows: Step 1), for each path point P in the global path R n Generate multiple possible local paths in the corresponding local block, each local path starts from the current path point P n Go to the next waypoint P n+1 , calculate the risk of each path; Step 2), for each global path point P n , select 3 out-edge candidate points and 3 in-edge candidate points to form several possible path combinations, use the dynamic programming algorithm to evaluate these local paths, and select the path with the lowest comprehensive risk among all possible local path combinations as the exact path.

Images