Submersible navigation method and device based on spatial positioning

Abstract

The invention discloses a submersible navigation method and device based on space positioning, relates to the technical field of underwater navigation, and realizes self-adaptive path updating, improvement of navigation frequency at a high speed to enhance safety and reduction of frequency at a low speed to save energy by combining the real-time traveling speed and dynamic navigation frequency of a submersible. Meanwhile, obstacles are periodically detected and avoided, the navigation path can be optimized in real time, and the navigation safety, reliability and intelligent level of the submersible in a complex unknown water area are remarkably improved.

Description

Technical Field

[0001] This application relates to the field of underwater navigation technology, specifically to a submersible navigation method and device based on spatial positioning. Background Technology

[0002] Submersibles, especially autonomous underwater vehicles (AUVs), play a vital role in marine scientific research, resource exploration, pipeline inspection, and military reconnaissance. One of the core capabilities of submersibles lies in their autonomous navigation and path planning, enabling them to safely and efficiently navigate from a starting point to a destination.

[0003] Existing submersible navigation technologies mainly suffer from the following problems: Limitations of path planning algorithms: Traditional path planning algorithms, such as While algorithms like Dijkstra's algorithm can find an optimal path, they can suffer from problems such as high computational cost, susceptibility to local optima, and poor adaptability to dynamic environments in complex, high-dimensional underwater environments. Some intelligent optimization algorithms, such as ant colony optimization, while possessing strong global search capabilities and robustness, still have room for improvement in convergence speed and path quality in standard applications, especially when dealing with large maps, where efficiently constructing environmental models and guiding the search remains a challenge.

[0004] Fixed navigation frequency: Most submersible systems use a fixed navigation update cycle. This fixed frequency strategy has obvious drawbacks: when the submersible is traveling at high speed, a long update cycle may cause it to be unable to respond to newly appearing obstacles in time, thus raising the risk of collision; when the submersible is traveling at low speed or hovering, an excessively short update cycle will cause unnecessary consumption of computing resources and energy, shortening the submersible's endurance.

[0005] Insufficient adaptability to dynamic environments: The underwater environment is complex and ever-changing, containing dynamic obstacles such as moving marine life, floating fishing nets, other vehicles, and sudden changes in terrain. Traditional navigation methods are usually based on static prior maps for planning, making it difficult to perceive and avoid these dynamic obstacles in real time. This can cause the planned path to become infeasible or unsafe in actual navigation, reducing the reliability and safety of navigation. Summary of the Invention

[0006] This application provides a submersible navigation method and apparatus based on spatial positioning to solve the problems of low path planning efficiency, inflexible navigation frequency, and poor adaptability to dynamic obstacles in the prior art.

[0007] The first aspect of this application provides a submersible navigation method based on spatial positioning, comprising: The starting point spatial positioning information and the target point positioning information of the submersible are obtained, and the navigation path of the first submersible is planned based on the starting point spatial positioning information and the target point positioning information. The real-time speed of the submersible is obtained, and the correlation table between the speed and the real-time navigation frequency is queried based on the real-time speed to obtain the target real-time navigation cycle corresponding to the real-time speed. After each target real-time navigation cycle, obstacle information around the submersible is collected, and the navigation path of the first submersible is updated based on the obstacle information around the submersible to obtain the navigation path of the second submersible. Based on the second submersible navigation path, the submersible is navigated to complete submersible navigation based on spatial positioning.

[0008] In one possible design approach, acquiring the starting point spatial positioning information and the target point positioning information of the submersible, and planning the navigation path of the first submersible based on the starting point spatial positioning information and the target point positioning information, includes: Acquire the spatial positioning information of the submersible's starting point and the positioning information of the target point; Based on the spatial positioning information of the starting point and the positioning information of the target point, the environmental model is constructed using the grid method to obtain the environmental model. The pheromone concentration between adjacent grid cells in the ant colony algorithm and the environment model are zero during initialization. For any ant, the grid in the environmental model where the spatial location information of the starting point is located is taken as the ant's starting grid, and the grid in the environmental model where the location information of the target point is located is taken as the ant's ending grid. Ants are restricted to moving only to adjacent grids during their movement. Based on the environmental model and the pheromone concentration between adjacent grids, the ant is controlled to move to the next grid, and the grid where the ant is located is added to the taboo list. Determine whether the endpoint grid is located in the taboo list. If so, determine that the corresponding ant has completed the search and obtain the target path between the starting point spatial positioning information and the target point positioning information of the ant. Otherwise, control the ant to move repeatedly until the endpoint grid is located in the taboo list and determine that the corresponding ant has completed the search and obtain the target path between the starting point spatial positioning information and the target point positioning information of the ant. Based on the target path, the pheromone concentration in the environmental model is updated to obtain the updated pheromone concentration; If the number of times the path length of the optimal target path has not decreased has reached a preset threshold, then the optimal target path is determined as the navigation path of the first submersible; otherwise, the ants are controlled to perform the next search based on the updated pheromone concentration.

[0009] In one possible design approach, based on the spatial positioning information of the starting point and the positioning information of the target point, an environmental model is constructed using the grid method, resulting in an environmental model including: Using the spatial positioning information of the starting point and the positioning information of the target point as the coordinates of the diagonal corners, a two-dimensional planar map containing the water area is constructed, and the two-dimensional planar map is evenly divided into M×N squares to obtain M×N grids; The environmental model is obtained by setting the attribute value corresponding to the grid that intersects with the land to 1; wherein, setting the attribute value corresponding to the grid that intersects with the land to 1 means that the grid cannot be walked on by a submersible.

[0010] In one possible design approach, based on the environmental model and pheromones between adjacent grids, the movement of ants to the next grid is controlled, including: Based on the environmental model, add all grids with an attribute value of 1 to the taboo list; Based on the pheromone between adjacent grids, determine the transition probability between other grids that are not in the taboo list and are adjacent to the grid where the current ant is located and the grid where the current ant is located; Based on the transition probability, the next grid corresponding to the ant is selected, and the ant is controlled to move to the next grid.

[0011] In one possible design approach, based on the pheromones between adjacent grids, the transition probability between the current ant's grid and other grids not in the taboo list and adjacent to the current ant's grid is determined as follows:

[0012] in, Indicates the first t During this planning process, the current grid where the ant is located... i Other grilles adjacent to it j The transition probability between them Indicates the grid where the ant is currently located. i Adjacent sets of other grids that are not on the taboo list, Indicates the current grid position of the ant. i Other grilles adjacent to it j The pheromone concentration along the path between them This represents the heuristic function value, and is set to 1. Represents the information heuristic factor, This represents the expected heuristic factor.

[0013] In one possible design approach, selecting the next grid corresponding to the ant based on the transition probability includes: selecting the next grid corresponding to the ant using a roulette wheel strategy based on the transition probability.

[0014] In one possible design approach, the pheromone concentration in the environmental model is updated according to the target path to obtain the updated pheromone concentration, including: Determine the optimal target path among all target paths corresponding to ants, assign a first fixed pheromone to the sub-paths between each adjacent grid in the target paths corresponding to all ants, and additionally assign a second fixed pheromone to the sub-paths between each adjacent grid in the optimal target path, to obtain the updated pheromone concentration.

[0015] In one possible design approach, before updating the pheromone concentration in the environmental model, the method further includes: volatilizing the pheromone concentration in the environmental model.

[0016] In one possible design approach, after each target real-time navigation cycle, obstacle information around the submersible is collected, and the first submersible navigation path is updated based on the obstacle information to obtain a second submersible navigation path, including: After each target real-time navigation cycle, the obstacles around the submersible are collected, and obstacles that are not at the same height as the submersible are removed to obtain obstacle information; Based on the obstacle information around the submersible, the attribute value of the grid containing obstacles in the environment model is set to 1 to obtain the updated environment module; Based on the updated environment module, the navigation path of the first submersible is corrected to obtain the navigation path of the second submersible.

[0017] A second aspect of this application provides a space-based positioning-based submersible navigation device, comprising: The initial path acquisition module is used to acquire the starting point spatial positioning information and the target point positioning information of the submersible, and to plan the navigation path of the first submersible based on the starting point spatial positioning information and the target point positioning information. The navigation cycle determination module is used to obtain the real-time speed of the submersible and query the correlation table between the speed and the real-time navigation frequency based on the real-time speed to obtain the target real-time navigation cycle corresponding to the real-time speed. The navigation path update module is used to collect obstacle information around the submersible after each target real-time navigation cycle, and update the first submersible navigation path according to the obstacle information around the submersible to obtain the second submersible navigation path. The submersible navigation module is used to navigate the submersible based on the second submersible navigation path, thereby completing submersible navigation based on spatial positioning.

[0018] This application provides a submersible navigation method and apparatus based on spatial positioning. By combining the submersible's real-time travel speed with dynamic navigation frequency, adaptive path updates are achieved. The navigation frequency is increased at high speeds to enhance safety, and decreased at low speeds to save energy. Simultaneously, periodic obstacle detection and avoidance optimize the navigation path in real time, significantly improving the submersible's navigation safety, reliability, and intelligence in complex and unknown waters. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0020] Figure 1 A flowchart illustrating a submersible navigation method based on spatial positioning, provided as an embodiment of this application.

[0021] Figure 2 This is a schematic diagram of a submersible navigation device based on spatial positioning, provided as an embodiment of this application.

[0022] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0023] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0024] The embodiments of this application are described in detail below with reference to the accompanying drawings.

[0025] like Figure 1 As shown, this application provides a submersible navigation method based on spatial positioning, including: S101. Obtain the starting point spatial positioning information and the target point positioning information of the submersible, and plan the navigation path of the first submersible based on the starting point spatial positioning information and the target point positioning information.

[0026] For example, spatial positioning information of the starting and target points can be obtained, relying on high-precision global navigation satellite systems (such as GPS, when on the surface) or pre-calibrated inertial navigation systems. This information mainly includes two-dimensional coordinates such as latitude and longitude, providing the submersible with precise coordinates in the three-dimensional underwater space. Subsequently, based on these coordinate points, combined with pre-set electronic charts or known underwater topographic databases, advanced path planning algorithms (such as...) can be applied. The algorithm (Dijkstra's algorithm) is used to plan an initial navigation path for the submersible from the starting point to the target point. This initial path takes into account various factors such as travel distance, energy consumption, known terrain obstacles, preset safety corridors, and hydrological environment (such as ocean currents), aiming to generate a globally optimal or suboptimal, safe and feasible baseline route, providing a reliable initial plan for subsequent dynamic adjustments.

[0027] Optionally, the starting point spatial positioning information and the target point positioning information of the submersible can both be set as latitude and longitude coordinates.

[0028] S102. Obtain the real-time travel speed of the submersible, and query the correlation table between the travel speed and the real-time navigation frequency based on the real-time travel speed to obtain the target real-time navigation cycle corresponding to the real-time travel speed.

[0029] During its voyage, the submersible obtains its own velocity v in real time using its onboard Doppler velocimeter or other speed sensors. The submersible's control system stores a pre-set table relating velocities to real-time navigation frequencies. This table, established using extensive experimental and simulation data, reflects the minimum navigation update frequency required to ensure safety at different speeds.

[0030] This application's embodiments abandon the traditional fixed navigation cycle mode and innovatively introduce a speed-frequency correlation mechanism. By monitoring the submersible's speed in real time and querying a preset correlation table, the system dynamically adjusts the target's real-time navigation cycle. When the submersible is traveling at high speed, the system automatically shortens the navigation cycle, increases the frequency of environmental perception and path updates, and ensures timely detection and avoidance of rapidly approaching obstacles, greatly improving navigation safety at high speeds. Conversely, when the submersible is traveling at low speed or performing delicate operations, the system automatically extends the navigation cycle, reducing unnecessary sensor data acquisition, path calculation, and energy consumption, effectively saving computing resources and precious electrical energy, and significantly extending the submersible's endurance. This intelligent adaptive mechanism achieves a perfect balance between safety and economy.

[0031] S103. After each target real-time navigation cycle, obstacle information around the submersible is collected, and the navigation path of the first submersible is updated according to the obstacle information around the submersible to obtain the navigation path of the second submersible.

[0032] A timer can be started, triggering obstacle information collection every time a target real-time navigation cycle T is completed. The submersible uses its onboard sensors, such as forward-looking sonar, mechanical scanning sonar, and underwater cameras, to detect the environment within a certain range. After acquiring the raw obstacle data, it compares the depth of all obstacles with its current depth information. If the depth of an obstacle (also known as the submersible's altitude in the water) differs from the submersible's depth by more than a safety threshold (e.g., ±5 meters), the obstacle is considered to be outside the same navigation plane and does not pose a direct threat to the submersible, thus being removed from the obstacle list. The final result is information on obstacles at the same altitude as the submersible that require close monitoring.

[0033] The selected obstacle information is mapped into the environment model. Specifically, based on the spatial coordinates of the obstacles, the corresponding grid cells are determined, and the attribute values ​​of these grid cells are changed from 0 to 1. This allows the environment model to be updated in real time, reflecting the latest obstacle distribution, resulting in an updated environment model.

[0034] Based on the updated environment model, the currently executing navigation path of the first submersible is corrected online. The correction algorithm can employ local path planning algorithms, such as the dynamic window method, the artificial potential field method, or a simplified method. Algorithm. For example, if a submersible detects a new obstacle on its path, it can use its current position as the starting point and a point ahead on the original path as the temporary endpoint. Within the updated environment model, it can quickly plan a local path to bypass the new obstacle and replace the blocked portion of the original path with this local path. Finally, the path integrating the corrected portion becomes the second submersible navigation path.

[0035] S104. Based on the second submersible navigation path, navigate the submersible to complete the submersible navigation based on spatial positioning.

[0036] The embodiments of this application do not require determining the depth of the submersible, but can still achieve real-time dynamic navigation of the submersible. Compared with the prior art, the technical principle is simpler and can achieve the same or similar technical effects.

[0037] In one possible design approach, acquiring the starting point spatial positioning information and the target point positioning information of the submersible, and planning the navigation path of the first submersible based on the starting point spatial positioning information and the target point positioning information, includes: Acquire the spatial positioning information of the submersible's starting point and the positioning information of the target point; Based on the spatial positioning information of the starting point and the positioning information of the target point, the environmental model is constructed using the grid method to obtain the environmental model. The pheromone concentration between adjacent grid cells in the ant colony algorithm and the environment model are zero during initialization. For any ant, the grid in the environmental model where the spatial location information of the starting point is located is taken as the ant's starting grid, and the grid in the environmental model where the location information of the target point is located is taken as the ant's ending grid. Ants are restricted to moving only to adjacent grids during their movement. Based on the environmental model and the pheromone concentration between adjacent grids, the ant is controlled to move to the next grid, and the grid where the ant is located is added to the taboo list. Determine whether the endpoint grid is located in the taboo list. If so, determine that the corresponding ant has completed the search and obtain the target path between the starting point spatial positioning information and the target point positioning information of the ant. Otherwise, control the ant to move repeatedly until the endpoint grid is located in the taboo list and determine that the corresponding ant has completed the search and obtain the target path between the starting point spatial positioning information and the target point positioning information of the ant. Based on the target path, the pheromone concentration in the environmental model is updated to obtain the updated pheromone concentration; If the number of times the path length of the optimal target path has not decreased has reached a preset threshold, then the optimal target path is determined as the navigation path of the first submersible; otherwise, the ants are controlled to perform the next search based on the updated pheromone concentration.

[0038] This application employs an improved ant colony algorithm for initial path planning. First, the complex continuous aquatic environment is discretized into a grid map with clearly defined attributes using a grid method, greatly simplifying the complexity of the environmental model and providing a clear and efficient computational foundation for the algorithm. Second, during the iterative process of the ant colony algorithm, pheromone updates are performed on all paths traversed by ants, and additional pheromone rewards are given to the optimal path in the current iteration. This elitist strategy accelerates the convergence of pheromones to high-quality paths, significantly improving the algorithm's convergence speed and avoiding the premature convergence and search stagnation problems common in standard ant colony algorithms. Finally, by setting the number of times the optimal path length does not decrease as a termination condition, the algorithm is ensured to stop promptly after finding a satisfactory solution, balancing solution quality and time cost. Therefore, the planned first submersible navigation path is a relatively optimal path found globally, comprehensively considering path length and obstacle avoidance, laying a solid foundation for the safe and efficient navigation of the submersible.

[0039] Optionally, to ensure accuracy during navigation and the maneuverability of the submersible, when using the grid method to grid the water area (i.e., temporarily disregarding the height direction), the side length of the grid should be at least three times the maximum length of the submersible.

[0040] In one possible design approach, based on the spatial positioning information of the starting point and the positioning information of the target point, an environmental model is constructed using the grid method, resulting in an environmental model including: Using the spatial positioning information of the starting point and the positioning information of the target point as the coordinates of the diagonal corners, a two-dimensional planar map containing the water area is constructed, and the two-dimensional planar map is evenly divided into M×N squares to obtain M×N grids; Here, M represents the number of rows in the gridding process, and N represents the number of columns. For example, using starting point A (i.e., the grid center corresponding to the starting point's spatial positioning information) and target point B (i.e., the grid center corresponding to the target point's positioning information) as diagonal points, a two-dimensional rectangular map that can completely cover the predetermined navigation area is constructed. This rectangular map is evenly divided into M rows and N columns of square grids, with each grid representing a small area of ​​water. By loading electronic nautical charts or known topographic data, land or permanent obstacle areas in the map are identified. The attribute value of any grid intersecting with land or permanent obstacles is set to 1, indicating that the grid is an obstacle grid and the submersible cannot pass through. The attribute value of grids completely within navigable waters is set to 0, indicating free grids. Thus, an environmental model with defined attribute values, consisting of M×N grids, is obtained.

[0041] The environmental model is obtained by setting the attribute value corresponding to the grid that intersects with the land to 1; wherein, setting the attribute value corresponding to the grid that intersects with the land to 1 means that the grid cannot be walked on by a submersible.

[0042] In one possible design approach, based on the environmental model and pheromones between adjacent grids, the movement of ants to the next grid is controlled, including: Based on the environment model, add all grids with an attribute value of 1 to the taboo list; Based on the pheromone between adjacent grids, determine the transition probability between other grids that are not in the taboo list and are adjacent to the grid where the current ant is located and the grid where the current ant is located; Based on the transition probability, the next grid corresponding to the ant is selected, and the ant is controlled to move to the next grid.

[0043] In one possible design approach, based on the pheromones between adjacent grids, the transition probability between the current ant's grid and other grids not in the taboo list and adjacent to the current ant's grid is determined as follows:

[0044] in, Indicates the first t During this planning process, the current grid where the ant is located... i Other grilles adjacent to it j The transition probability between them Indicates the grid where the ant is currently located. i Adjacent sets of other grids that are not on the taboo list, Indicates the current grid position of the ant. i Other grilles adjacent to it j The pheromone concentration along the path between them This represents the heuristic function value, and is set to 1. Represents the information heuristic factor, This represents the expected heuristic factor.

[0045] It is worth noting that, to avoid collisions with obstacles during the submersible's movement, the submersible can only move to adjacent grids that share a common edge with the grid where the current ant is located. Therefore, the distance the ant moves each time is fixed, and the heuristic function value in the ant colony algorithm can be set to 1. The information heuristic factor and the expectation heuristic factor are extensively described in existing technologies; this application does not limit these two parameters, and reference to existing technologies is sufficient.

[0046] In one possible design approach, selecting the next grid corresponding to the ant based on the transition probability includes: selecting the next grid corresponding to the ant using a roulette wheel strategy based on the transition probability.

[0047] In one possible design approach, the pheromone concentration in the environmental model is updated according to the target path to obtain the updated pheromone concentration, including: Determine the optimal target path among all target paths corresponding to ants, assign a first fixed pheromone to the sub-paths between each adjacent grid in the target paths corresponding to all ants, and additionally assign a second fixed pheromone to the sub-paths between each adjacent grid in the optimal target path, to obtain the updated pheromone concentration.

[0048] In one possible design approach, before updating the pheromone concentration in the environmental model, the method further includes: volatilizing the pheromone concentration in the environmental model.

[0049] For example, a volatile factor between (0,1) can be set, and the volatile factor can be multiplied by the pheromone concentration in the environmental model to obtain the pheromone concentration after volatile treatment.

[0050] In one possible design approach, after each target real-time navigation cycle, obstacle information around the submersible is collected, and the first submersible navigation path is updated based on the obstacle information to obtain a second submersible navigation path, including: After each target real-time navigation cycle, the obstacles around the submersible are collected, and obstacles that are not at the same height as the submersible are removed to obtain obstacle information; Based on the obstacle information around the submersible, the attribute value of the grid containing obstacles in the environment model is set to 1 to obtain the updated environment module; Based on the updated environment module, the navigation path of the first submersible is corrected to obtain the navigation path of the second submersible.

[0051] Optionally, to improve navigation accuracy, this application provides an obstacle recognition method, specifically including: Step 1: Train a highly accurate obstacle recognition model; An obstacle recognition model is constructed using the YOLO model, and the hyperparameters of the obstacle recognition model are initialized to obtain a hyperparameter vector; for example, the hyperparameter vector is obtained by randomly generating and encoding the hyperparameters between the upper and lower bounds. Divergent neighborhood search of the hyperparameter vector is as follows:

[0052]

[0053] in, Indicates the first t During the training process, the first i The th hyperparameter vector of the th ... d Dimensional parameters, i =1,2,…,NP, where NP represents the total number of hyperparameter vectors. d =1,2,…,D, where D represents the total dimension of the parameters in the hyperparameter vector. Indicates the first i The corresponding hyperparameter vector of the th hyperparameter vector p The first divergent neighborhood search value d Dimensional parameters; Represents the sine function. Represents pi (π). =1,2,…, h , h Indicates the total number of search directions; This represents the preset step size factor. This represents the current training step size in the t-th training process. This indicates the preset maximum number of training iterations. This represents the step size adjustment coefficient. This indicates the current number of training iterations.

[0054] In the early stages of optimization, the large step size and multi-directional sinusoidal perturbations allow the algorithm to jump out of the initial point and efficiently explore a vast hyperparameter space, greatly reducing the risk of getting trapped in local optima. This is crucial for complex non-convex optimization problems like YOLO. The dynamic decay mechanism of the step size allows the search process to smoothly transition from the global exploration phase to the local exploration phase. Early on, promising regions are quickly identified, and later, a refined search is conducted within those regions, balancing efficiency and accuracy. Compared to completely random searches, this directional divergence based on the current optimal solution can find better-performing hyperparameter combinations more quickly, accelerating the entire optimization process.

[0055] The divergent neighborhood search value with the highest fitness corresponding to the hyperparameter vector is used as the initial hyperparameter vector after training, and information fusion is performed on the initial trained hyperparameter vector as follows:

[0056]

[0057]

[0058] in, Indicates the first j The hyperparameter vector after the initial training and the first training step k The Euclidean distance between the initial trained hyperparameter vectors Indicates the first d The upper limit of the dimension parameter, Indicates the first d The lower bound of the dimension parameter, The first hyperparameter vector represents the optimal hyperparameter vector. d Dimensional parameters, Indicates the first j The first hyperparameter vector after information fusion d Dimensional parameters, Indicates information fusion parameters, This indicates the upper limit of the information fusion parameters, and is set to 1; This represents the lower limit of the information fusion parameters, and is set to 0.00001; Indicates the first j The hyperparameter vector after the initial training has relation to the th... k The influence parameter of the hyperparameter vector after initial training. j =1,2,…,NP This represents the information collection intensity parameter, and is set to 0.5; This represents the information collection scale parameter, and is set to 1.5; Represents the natural constant. Indicates the first t During the training process, the first j The first hyperparameter vector after initial training. d Dimensional parameters, Indicates the first t During the training process, the first k The first hyperparameter vector after initial training. d Dimensional parameters, k =1,2,…,NP, and j Not equal to k ;| | represents absolute value.

[0059] The information fusion mechanism makes the algorithm less susceptible to being swayed by individual, abruptly changed, bad solutions. Even if a vector accidentally jumps to a poor region, it will be pulled back to a more promising region by the majority of other vectors. This collective consensus mechanism enhances the stability and resilience of the optimization process. When two vectors are far apart, their influence is weak, encouraging individuals to maintain diversity and continue exploring; when they are close, their influence is strong, encouraging individuals to learn from their superior peers and engage in local development. This achieves a dynamic balance between exploration and development.

[0060] The chaotic mapping search for the hyperparameter vector after information fusion is as follows:

[0061]

[0062] in, Indicates the first t During the training process, the first n The first hyperparameter vector after information fusion d Dimensional parameters, express The corresponding chaotic decision factor, Represents the arctangent function; This represents the first random number between (0,1). Indicates the first n The hyperparameter vector after exploring the chaotic map. d Dimensional parameters. Simulated annealing can be used to probabilistically accept inferior solutions in chaotic map search, ensuring training speed.

[0063] Chaotic sequences can generate highly ergodic points within a tiny region, essentially performing a thorough, detailed scan of the neighborhood of the current optimal solution. This helps identify superior, peak solutions overlooked by conventional gradient descent or fixed-step-size search. Later in the optimization process, when all individuals cluster near a local optimum, conventional algorithms may prematurely converge, stopping the search. The inherent randomness of chaotic maps provides a perturbation, helping the algorithm escape this stagnation and continue searching for better solutions, ensuring the quality of the final solution.

[0064] Repeatedly perform divergent neighborhood search, information fusion, and chaotic mapping search until the preset maximum number of training iterations is reached. Then, re-obtain the optimal hyperparameter vector and use the hyperparameters of the re-obtained optimal hyperparameter vector as the final hyperparameters of the obstacle recognition model to obtain the trained obstacle recognition model.

[0065] The obstacle recognition model trained in the above manner can identify obstacles more accurately, ensuring the accuracy of path planning for submersibles in water conservancy.

[0066] Step 2: Perform binocular obstacle recognition using the trained obstacle recognition model, and locate the same obstacle based on the obstacle recognition results of the two cameras. Determine the relative position information between the submersible and the obstacle, and remove obstacles that are not at the same height as the submersible based on the relative position relationship.

[0067] Step 3: Based on the relative position information between the submersible and the obstacle, and the grid position of the submersible, determine the grid where the obstacle is located, so that the attribute value of the grid containing the obstacle in the environment model can be set to 1.

[0068] This application's embodiments construct a closed-loop dynamic navigation strategy encompassing perception, decision-making, and execution. During each real-time navigation cycle, obstacle information around the submersible can be actively collected. Specifically, by eliminating obstacles at different altitudes from the submersible, interference information that poses no threat to the current navigation level is filtered out, making obstacle perception more accurate and efficient. Subsequently, newly detected obstacles can be updated in real-time in the environmental model (by setting the corresponding grid attribute to 1), and based on this dynamically updated environmental model, the original first submersible navigation path is corrected online. This rolling time-domain path update strategy enables the submersible to readily cope with dynamic threats such as moving marine life and sudden underwater structures, always navigating along the optimal feasible path in the current environment. This significantly improves the submersible's autonomous survivability and mission completion rate in unknown, unstructured dynamic environments.

[0069] like Figure 2 As shown, this application provides a space-based positioning-based submersible navigation device, including: The initial path acquisition module 201 is used to acquire the starting point spatial positioning information and the target point positioning information of the submersible, and plan the navigation path of the first submersible based on the starting point spatial positioning information and the target point positioning information. The navigation cycle determination module 202 is used to obtain the real-time speed of the submersible, and query the correlation table between the speed and the real-time navigation frequency based on the real-time speed to obtain the target real-time navigation cycle corresponding to the real-time speed. The navigation path update module 203 is used to collect obstacle information around the submersible after each target real-time navigation cycle, and update the first submersible navigation path according to the obstacle information around the submersible to obtain the second submersible navigation path. The submersible navigation module 204 is used to navigate the submersible based on the second submersible navigation path, thereby completing submersible navigation based on spatial positioning.

[0070] The submersible navigation device based on spatial positioning provided in this application embodiment can execute the above-described method and technical solution. Its technical principle and beneficial effects are similar, and will not be repeated here.

[0071] This application integrates three core functions—global path planning, adaptive frequency control, and local dynamic obstacle avoidance—within a unified technical framework. The logic between modules is clear, and data flow is smooth. The initial path planning module 201 provides an initial, globally optimal guiding direction; the navigation cycle determination module 202 ensures navigation cycle adjustment under different operating conditions; and the navigation path update module 203 handles real-time environmental changes. This layered, progressive, and collaborative architecture allows the entire navigation system to possess both a global perspective and local flexibility. Furthermore, the algorithms employed—such as the grid method, ant colony algorithm, and lookup table method—are easily implemented on computers, facilitating software development and deployment in the embedded control system of submersibles, thus possessing strong engineering practical value and broad market application prospects.

[0072] This application also provides a computer-readable storage medium storing a computer program / instructions thereon, which, when executed by a processor, implements the steps in the method disclosed in this application.

[0073] This application also provides a computer program product that, when run on an electronic device, causes a processor to execute the steps of the method disclosed in this application.

[0074] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0075] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, apparatuses, electronic devices, and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0076] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0077] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0078] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0079] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0080] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A submersible navigation method based on spatial positioning, characterized in that, include: The starting point spatial positioning information and the target point positioning information of the submersible are obtained, and the navigation path of the first submersible is planned based on the starting point spatial positioning information and the target point positioning information. The real-time speed of the submersible is obtained, and the correlation table between the speed and the real-time navigation frequency is queried based on the real-time speed to obtain the target real-time navigation cycle corresponding to the real-time speed. After each target real-time navigation cycle, obstacle information around the submersible is collected, and the navigation path of the first submersible is updated based on the obstacle information around the submersible to obtain the navigation path of the second submersible. Based on the second submersible navigation path, the submersible is navigated to complete submersible navigation based on spatial positioning.

2. The submersible navigation method based on spatial positioning according to claim 1, characterized in that, Acquiring the starting point spatial positioning information and target point positioning information of the submersible, and planning the navigation path of the first submersible based on the starting point spatial positioning information and target point positioning information, including: Acquire the spatial positioning information of the submersible's starting point and the positioning information of the target point; Based on the spatial positioning information of the starting point and the positioning information of the target point, the environmental model is constructed using the grid method to obtain the environmental model. The pheromone concentration between adjacent grid cells in the ant colony algorithm and the environment model are zero during initialization. For any ant, the grid in the environmental model where the spatial location information of the starting point is located is taken as the ant's starting grid, and the grid in the environmental model where the location information of the target point is located is taken as the ant's ending grid. Ants are restricted to moving only to adjacent grids during their movement. Based on the environmental model and the pheromone concentration between adjacent grids, the ant is controlled to move to the next grid, and the grid where the ant is located is added to the taboo list. Determine whether the endpoint grid is located in the taboo list. If so, determine that the corresponding ant has completed the search and obtain the target path between the starting point spatial positioning information and the target point positioning information of the ant. Otherwise, control the ant to move repeatedly until the endpoint grid is located in the taboo list and determine that the corresponding ant has completed the search and obtain the target path between the starting point spatial positioning information and the target point positioning information of the ant. Based on the target path, the pheromone concentration in the environmental model is updated to obtain the updated pheromone concentration; If the number of times the path length of the optimal target path has not decreased has reached a preset threshold, then the optimal target path is determined as the navigation path of the first submersible; otherwise, the ants are controlled to perform the next search based on the updated pheromone concentration.

3. The submersible navigation method based on spatial positioning according to claim 2, characterized in that, Based on the spatial positioning information of the starting point and the positioning information of the target point, the environmental model is constructed using the grid method, resulting in an environmental model including: Using the spatial positioning information of the starting point and the positioning information of the target point as the coordinates of the diagonal corners, a two-dimensional planar map containing the water area is constructed, and the two-dimensional planar map is evenly divided into M×N squares to obtain M×N grids; The environmental model is obtained by setting the attribute value corresponding to the grid that intersects with the land to 1; wherein, setting the attribute value corresponding to the grid that intersects with the land to 1 means that the grid cannot be walked on by a submersible.

4. The submersible navigation method based on spatial positioning according to claim 2, characterized in that, Based on the environmental model and the pheromones between adjacent grids, the ants are controlled to move to the next grid, including: Based on the environmental model, add all grids with an attribute value of 1 to the taboo list; Based on the pheromones between adjacent grids, determine the transition probability between other grids that are not in the taboo list and are adjacent to the grid where the current ant is located and the grid where the current ant is located; Based on the transition probability, the next grid corresponding to the ant is selected, and the ant is controlled to move to the next grid.

5. The submersible navigation method based on spatial positioning according to claim 4, characterized in that, Based on the pheromones between adjacent grids, the transition probability between the current ant's grid and other grids that are not in the taboo list and are adjacent to the current ant's grid is determined as follows: in, Indicates the first t During this planning process, the current grid where the ant is located... i Other grilles adjacent to it j The transition probability between them Indicates the grid where the ant is currently located. i Adjacent sets of other grids that are not on the forbidden list, Indicates the current grid position of the ant. i Other grilles adjacent to it j The pheromone concentration along the path between them This represents the heuristic function value, and is set to 1. Represents the information heuristic factor, This represents the expected heuristic factor.

6. The submersible navigation method based on spatial positioning according to claim 4, characterized in that, Selecting the next grid corresponding to the ant based on the transition probability includes: selecting the next grid corresponding to the ant using a roulette wheel strategy based on the transition probability.

7. The submersible navigation method based on spatial positioning according to claim 6, characterized in that, Based on the target path, the pheromone concentration in the environmental model is updated to obtain the updated pheromone concentration, including: Determine the optimal target path among all target paths corresponding to ants, assign a first fixed pheromone to the sub-paths between each adjacent grid in the target paths corresponding to all ants, and additionally assign a second fixed pheromone to the sub-paths between each adjacent grid in the optimal target path, to obtain the updated pheromone concentration.

8. The submersible navigation method based on spatial positioning according to claim 7, characterized in that, Before updating the pheromone concentration in the environmental model, the method further includes: performing a volatilization treatment on the pheromone concentration in the environmental model.

9. The submersible navigation method based on spatial positioning according to claim 8, characterized in that, After each target real-time navigation cycle, obstacle information around the submersible is collected, and the first submersible navigation path is updated based on the obstacle information to obtain a second submersible navigation path, including: After each target real-time navigation cycle, the obstacles around the submersible are collected, and obstacles that are not at the same height as the submersible are removed to obtain obstacle information; Based on the obstacle information around the submersible, the attribute value of the grid containing obstacles in the environment model is set to 1 to obtain the updated environment module; Based on the updated environment module, the navigation path of the first submersible is corrected to obtain the navigation path of the second submersible.

10. A submersible navigation device based on spatial positioning, characterized in that, include: The initial path acquisition module is used to acquire the starting point spatial positioning information and the target point positioning information of the submersible, and to plan the navigation path of the first submersible based on the starting point spatial positioning information and the target point positioning information. The navigation cycle determination module is used to obtain the real-time speed of the submersible and query the correlation table between the speed and the real-time navigation frequency based on the real-time speed to obtain the target real-time navigation cycle corresponding to the real-time speed. The navigation path update module is used to collect obstacle information around the submersible after each target real-time navigation cycle, and update the first submersible navigation path according to the obstacle information around the submersible to obtain the second submersible navigation path. The submersible navigation module is used to navigate the submersible based on the second submersible navigation path, thereby completing submersible navigation based on spatial positioning.

Images