Maritime unmanned cluster target searching method based on prior graph

Abstract

The invention relates to the technical field of offshore unmanned clusters, in particular to an offshore unmanned cluster target search method based on a prior graph, which comprises the following steps: comprehensively analyzing prior information to obtain a plurality of target points of a current target search task, and forming a prior set corresponding to the current target search task by all the target points; performing Gaussian modeling according to the specific position and the distribution range of each target point to obtain a plurality of Gaussian functions; superposing the Gaussian functions to generate a target search probability graph; expanding the prior set based on the reliability of each target point, and optimizing the target search probability graph based on the sum of Gaussian functions of expanded target points obtained through expansion to obtain an optimized target search probability graph; extracting a plurality of valuable search areas from the optimized target search probability graph; and indicating the offshore unmanned cluster to perform target search according to the priority sequence of each valuable search area. According to the method, the target searching efficiency and accuracy are remarkably improved.

Description

Technical Field

[0001] Embodiments of the present application relate to the technical field of maritime unmanned cluster technology, and particularly to a method for searching for maritime unmanned cluster targets based on a priori maps. Background Art

[0002] Maritime rescue refers to a comprehensive operation to provide emergency assistance to targets such as distressed personnel, ships, and aircraft in the sea or coastal areas. Its core objective is to ensure life safety, reduce property losses, and prevent and control marine environmental pollution. The core tasks of maritime rescue include life rescue, vehicle assistance, emergency disposal of environmental pollution, and ensuring channel safety.

[0003] Target search is a basic means for performing maritime rescue tasks. Most traditional target search methods rely on a single unmanned system, and their search efficiency and coverage are often limited. In the complex and changing marine environment, traditional target search methods are difficult to cope with various challenges, such as poor underwater communication, environmental interference, and dynamic changes of targets.

[0004] In recent years, the technology of maritime unmanned clusters based on cluster cooperation has gradually emerged. A maritime unmanned cluster refers to a system composed of a certain number of drones, unmanned boats, unmanned submersibles, etc., interconnected through air, surface, and underwater communication networks, and realizing collaborative tasks in a group manner. Through the collaborative operation of multiple unmanned systems, this technology can significantly improve the flexibility and efficiency of target search.

[0005] However, when facing maritime rescue tasks, the target search strategy of maritime unmanned clusters is very important. How to optimize the target search strategy to improve the efficiency and accuracy of target search is an urgent problem to be solved. Summary of the Invention

[0006] In view of this, embodiments of the present application propose a method for searching for maritime unmanned cluster targets based on a priori maps, which can significantly improve the efficiency and accuracy of target search, provide a more intelligent solution for target recognition in complex marine environments, and open up a new path for maritime target search.

[0007] To achieve the above object, an embodiment of the present application proposes a method for searching for targets of a maritime unmanned cluster based on a prior map. The method includes: obtaining prior information, comprehensively analyzing the obtained prior information to obtain a number of target points corresponding to the current target search task; wherein the prior information includes historical search data, environmental monitoring records, and expert analysis opinions, and the number of target points constitutes a prior set corresponding to the current target search task; performing Gaussian modeling according to the specific positions and distribution ranges of the target points to obtain a number of Gaussian functions; using the Parzen window theory and non-parametric data-driven estimation to superimpose the Gaussian functions to generate a target search probability map corresponding to the current target search task; expanding the prior set based on the reliability of each target point, and optimizing the target search probability map based on the sum of the Gaussian functions of the expanded target points obtained by the expansion to obtain an optimized target search probability map; performing non-parametric data-driven estimation in the optimized target search probability map, and then analyzing each cluster to quantitatively extract a number of valuable search regions; instructing the maritime unmanned cluster to perform target search according to the priority order of each valuable search region; wherein, the greater the target probability of the valuable search region, the higher its corresponding priority.

[0008] To achieve the above object, an embodiment of the present application also proposes a system for searching for targets of a maritime unmanned cluster based on a prior map. The system includes: a prior analysis module, a Gaussian modeling module, a target search probability map generation module, a target search probability map optimization module, a valuable search region extraction module, and a search execution module; the prior analysis module is used to obtain prior information, comprehensively analyze the obtained prior information to obtain a number of target points corresponding to the current target search task, wherein the prior information includes historical search data, environmental monitoring records, and expert analysis opinions, and the number of target points constitutes a prior set corresponding to the current target search task; the Gaussian modeling module is used to perform Gaussian modeling according to the specific positions and distribution ranges of the target points to obtain a number of Gaussian functions; the target search probability map generation module is used to use the Parzen window theory and non-parametric data-driven estimation to superimpose the Gaussian functions to generate a target search probability map corresponding to the current target search task; the target search probability map optimization module is used to expand the prior set based on the reliability of each target point, and optimize the target search probability map based on the sum of the Gaussian functions of the expanded target points obtained by the expansion to obtain an optimized target search probability map; the valuable search region extraction module is used to perform non-parametric data-driven estimation in the optimized target search probability map, and then analyze each cluster to quantitatively extract a number of valuable search regions; the search execution module is used to instruct the maritime unmanned cluster to perform target search according to the priority order of each valuable search region, wherein, the greater the target probability of the valuable search region, the higher its corresponding priority.

[0009] To achieve the above object, an embodiment of the present application further provides an electronic device, where the electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a method for searching for targets of an unmanned marine cluster based on a prior map as described above.

[0010] To achieve the above object, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which when executed by a processor, can implement a method for searching for targets of an unmanned marine cluster based on a prior map as described above.

[0011] A method for searching for targets of an unmanned marine cluster based on a prior map proposed in the present application first comprehensively analyzes the obtained prior information to obtain several target points corresponding to the current target search task. Subsequently, Gaussian modeling is performed according to the specific positions and distribution ranges of the target points, and the Parzen window theory and non-parametric data-driven estimation are used to superimpose the Gaussian functions obtained by modeling to generate a target search probability map corresponding to the current target search task. Then, the prior set is expanded based on the reliability of each target point, and the target search probability map is optimized based on the sum of the Gaussian functions of the expanded target points obtained by the expansion. Non-parametric data-driven estimation is performed on the optimized target search probability map, and each cluster is analyzed to quantitatively extract several valuable search areas. Finally, the unmanned marine cluster is instructed to perform target search in the order of the priorities of the valuable search areas. Such a search process can not only significantly improve the efficiency and accuracy of target search, but also provide a more intelligent solution for target recognition in complex marine environments. Through the full utilization of prior information, the scientific construction of the target probability map, the effective application of clustering extraction, and the collaborative work of the unmanned system, the present application opens up a new path for the field of marine target search. In the future, with the continuous progress of technology, especially the development of artificial intelligence and big data analysis technologies, this method is expected to further improve its intelligence level and achieve more accurate and efficient target search. This will not only provide strong technical support for the development and protection of marine resources, but also provide an important data basis and tool for marine scientific research. The application prospect of the present application is very broad and can play an important role in multiple fields, including but not limited to marine resource exploration, environmental monitoring, marine safety and rescue, etc., and has extremely high social value and economic benefits. Through continuous adjustment and optimization in practical applications, a method for searching for targets of an unmanned marine cluster based on a prior map proposed in the present application will become an important technical support in the future field of marine exploration, providing a strong guarantee for realizing more intelligent marine management and development.

[0012] In some alternative embodiments, the prior information further includes relevant literature and unverified indication reports, and there is a chance that the unverified indication reports contain preliminary clues of the existence of the target; Historical search data is used to provide the frequency and location of the target appearance, providing a basis for evaluating whether there are potential targets in the search area. Environmental monitoring records are used to provide the impact of changes in marine conditions on the target distribution. The changes in marine conditions include but are not limited to tides, flow rates, and temperatures. Expert analysis opinions are used to provide a professional perspective for the interpretation of relevant literature and unverified indication reports.

[0013] In some alternative embodiments, after obtaining the prior information, comprehensively analyzing the obtained prior information, and obtaining several target points corresponding to the current target search task, the method further includes: Defining three key attributes for each target point, namely specific location, reliability, and distribution range. Denote the specific location of the th target point as , the reliability as , and the distribution range as ; The specific location is used to represent the central position of the expected target, which is evaluated based on historical search data; The reliability is used to represent the credibility of the specific location of the target point, which is the probability of the expected target appearing at the specific location obtained through quantitative evaluation and is obtained by analyzing historical search data; The distribution range is used to represent the possible range of the area where the expected target exists, which is set based on the characteristics of the expected target and the environmental conditions in the environmental monitoring records to ensure a comprehensive understanding of the potential distribution of the expected target.

[0014] In some alternative embodiments, Gaussian modeling is performed according to the specific locations and distribution ranges of each target point to obtain several Gaussian functions, which is achieved through the following formula: ; ; where represents the Gaussian function corresponding to the th target point, represents the specific location of the th target point, represents the distribution range of the th target point, is the total number of target points, is the independent variable, representing any position within the search area corresponding to the target search task.

[0015] In some alternative embodiments, the prior set is expanded based on the reliability of each target point, and the target search probability map is optimized based on the sum of the Gaussian functions of the expanded target points obtained by the expansion, so as to obtain the optimized target search probability map, including: Based on the preset maximum expansion number and the reliability of each target point, each target point is expanded respectively to obtain a number of expanded target points. The number of expanded target points obtained by expanding different target points is different. The number of expanded target points obtained by expanding the th target point is expressed by the formula: ; ; wherein, represents the reliability of the th target point, is the preset maximum expansion number at , represents the number of expanded target points obtained by expanding based on the th target point; According to the specific positions and distribution ranges of the respective expanded target points, Gaussian modeling is performed to obtain the Gaussian functions corresponding to the respective expanded target points, and the sum of the Gaussian functions of the respective expanded target points is calculated; Through the following formula, the target search probability map is optimized based on the sum of the Gaussian functions of the respective expanded target points to obtain the optimized target search probability map; ; ; wherein, represents the total number of expanded target points, represents the Gaussian function corresponding to the th expanded target point, represents the specific position of the th target point, represents the distribution range of the th target point, represents the unoptimized target search probability map, the value of is the target search probability at represents the optimized target search probability map.

[0016] In some alternative embodiments, non-parametric data-driven estimation is performed on the optimized target search probability map, and then each cluster is analyzed, so as to quantitatively extract a number of valuable search regions, including: Through the following formula, non-parametric data-driven estimation is performed on the optimized target search probability map to obtain the non-parametric data-driven estimation result: ; Among them, represents the number of environmental changes, is the preset probability decrease value caused by environmental changes, represents the non-parametric data-driven estimation result; Based on the non-parametric data-driven estimation result, each cluster is analyzed to ensure that each cluster contains adjacent points with positive values and all clusters are isolated from each other, so as to quantitatively extract several valuable search areas; Among them, as increases, gradually decreases, and the size of the valuable search area shrinks accordingly.

[0017] In some alternative embodiments, after instructing the unmanned maritime cluster to perform target search in the priority order of each valuable search area, the method further includes: Based on the estimated return, sailing time, and coverage time of each valuable search area, a comprehensive evaluation is performed to obtain the comprehensive evaluation value of each valuable search area respectively. Denote the comprehensive evaluation value of the th valuable search area as , The calculation process of is expressed by the formula as: ; Among them, represents the estimated return of the th valuable search area, which is calculated based on the probability integral of each point in the th valuable search area, represents the sailing time of the th valuable search area, represents the coverage time of the th valuable search area. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the related art, the following will briefly introduce the drawings required for the description of the embodiments of the present application or the related technology. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0019] Figure 1 is a flowchart of a method for target search of an unmanned maritime cluster based on a priori graph provided in an embodiment of the present application; Figure 2It is the target search probability graph provided in an embodiment of the present application; Figure 3 It is the optimized target search probability graph provided in an embodiment of the present application; Figure 4 It is the schematic diagram of cluster analysis provided in an embodiment of the present application; Figure 5 It is the schematic diagram of valuable search area extraction provided in an embodiment of the present application; Figure 6 It is the schematic diagram of the change of search accuracy over time provided in an embodiment of the present application; Figure 7 It is the schematic diagram of the cumulative reward curve provided in an embodiment of the present application; Figure 8 It is the structural schematic diagram of a maritime unmanned cluster target search system based on a prior graph provided in another embodiment of the present application; Figure 9 It is the structural schematic diagram of an electronic device provided in another embodiment of the present application. Detailed implementation manners

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the embodiments of the present application will be elaborated in detail below with reference to the accompanying drawings. However, those of ordinary skill in the art can understand that in the embodiments of the present application, many technical details are proposed for the readers to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical solutions required to be protected by the present application can be implemented. The following division of each embodiment is for convenience of description and should not constitute any limitation on the specific implementation manner of the present application. Each embodiment can be combined and cross-referenced with each other on the premise of not being contradictory.

[0021] To solve the problem that the efficiency and accuracy of traditional target search strategies are too low, an embodiment of the present application proposes a maritime unmanned cluster target search method based on a prior graph, which is applied to an electronic device. The electronic device can be a terminal or a server. In this embodiment and the following embodiments, the server is taken as an example for illustration. The implementation details of a maritime unmanned cluster target search method based on a prior graph proposed in this embodiment are specifically described below. The following content is only implementation details provided for convenient understanding and is not necessary for implementing this solution.

[0022] The specific process of a maritime unmanned cluster target search method based on a prior graph proposed in this embodiment can be as Figure 1 shown and includes: Step 101: Obtain prior information, comprehensively analyze the obtained prior information, and obtain several target points corresponding to the current target search task. All the target points form the prior set corresponding to the current target search task.

[0023] In specific implementation, the server first needs to obtain prior information, and then comprehensively analyze the obtained prior information to obtain several target points corresponding to the current target search task. All the target points form the prior set corresponding to the current target search task. The obtained prior information includes but is not limited to historical search data, environmental monitoring records, and expert analysis opinions.

[0024] In an example, when implementing the method for searching for unmanned cluster targets at sea based on a prior map, the primary task is to systematically collect prior information related to target search. This prior information includes historical search data, environmental monitoring records, expert analysis opinions, relevant literature, and unvalidated indication reports, etc. The collection of prior information is crucial because high-quality prior information can significantly improve the accuracy and effectiveness of subsequent modeling.

[0025] In an example, historical search data is used to provide the frequency and location of target appearances, providing a basis for evaluating whether there are potential targets in the search area. Environmental monitoring records are used to provide the impact of changes in marine conditions on target distribution. Changes in marine conditions include but are not limited to tides, flow rates, and temperatures. There is a chance that relevant literature and unvalidated indication reports contain preliminary clues about the existence of targets. Expert analysis opinions can provide a professional perspective for the interpretation of relevant literature and unvalidated indication reports.

[0026] In an example, after the server obtains several target points corresponding to the current target search task, it also needs to define three key attributes for each target point, namely specific location, reliability, and distribution range. The server records the specific location of the th target point as , the reliability as , and the distribution range as . The specific location is used to represent the central location of the expected target, which is evaluated based on historical search data. Reliability is used to represent the credibility of the specific location of the target point, which is a quantified probability of the expected target appearing at the specific location and can be obtained by analyzing historical search data. The distribution range is used to represent the possible range of the area where the expected target exists, which is set based on the characteristics of the expected target and the environmental conditions in the environmental monitoring records to ensure a comprehensive understanding of the potential distribution of the expected target.

[0027] Step 102: Perform Gaussian modeling based on the specific locations and distribution ranges of each target point to obtain several Gaussian functions.

[0028] In a specific implementation, to more effectively express the uncertainty of the target position, the server needs to perform Gaussian modeling based on the specific positions and distribution ranges of each target point, thereby obtaining several Gaussian functions. The mean of each Gaussian function is the specific position of the target point, and the standard deviation is the distribution range of the target point. This modeling method can effectively generate the probability distribution of the target points, providing a reliable basis for subsequent search strategies. The choice of the Gaussian distribution is reasonable because it can effectively describe the distribution characteristics of many natural phenomena. By applying the Gaussian function to each target point, the existence probability of the target point in space can be reflected, making the target search more scientific and controllable.

[0029] In one example, the server performs Gaussian modeling based on the specific positions and distribution ranges of each target point to obtain several Gaussian functions, which can be achieved through the following formula: ; ; where represents the Gaussian function corresponding to the th target point, represents the specific position of the th target point, represents the distribution range of the th target point, is the total number of target points, is the independent variable, representing any position within the search area corresponding to the target search task.

[0030] Step 103, use the Parzen window theory and non-parametric data-driven estimation to superimpose each Gaussian function to generate a target search probability map corresponding to the current target search task.

[0031] In a specific implementation, the server uses the Parzen window theory to superimpose multiple Gaussian functions to construct a comprehensive target search probability map. The target search probability map intuitively shows the target existence probability of each area, providing a clear direction for subsequent target search. The construction of the target search probability map not only improves the search efficiency but also enables the unmanned system to make more accurate decisions in complex environments. The information reflected by the target search probability map is extremely rich, not only showing the high-probability areas of potential targets but also marking the low-probability areas, enabling the rational allocation of search resources. By analyzing the target search probability map, the unmanned system can focus its efforts on searching the high-probability areas first, greatly improving the overall search efficiency.

[0032] In one example, the target search probability map can be as Figure 2 shown. Figure 2The X-axis in it represents the abscissa of the search area corresponding to the target search task, the Y-axis represents the ordinate of the search area corresponding to the target search task, and the Z-axis represents any position. The corresponding target search probability.

[0033] Step 104: Expand the prior set based on the reliability of each target point, and optimize the target search probability map based on the sum of the Gaussian functions of the expanded target points obtained by the expansion, so as to obtain an optimized target search probability map.

[0034] In specific implementation, there is only one target point in the prior set, which results in the target search probability map still being not precise enough. Therefore, the server needs to expand the prior set based on the reliability of each target point, and optimize the target search probability map based on the sum of the Gaussian functions of the expanded target points obtained by the expansion, so as to obtain an optimized target search probability map.

[0035] In an example, the server first needs to expand each target point respectively based on the preset maximum expansion number and the reliability of each target point to obtain several expanded target points, and the number of expanded target points obtained by expanding different target points is different. The number of expanded target points obtained by expanding the th target point is expressed by the formula: ; ; where, represents the reliability of the th target point, is the preset maximum expansion number, represents the number of expanded target points obtained by expanding based on the th target point.

[0036] Subsequently, the server performs Gaussian modeling according to the specific positions and distribution ranges of each expanded target point to obtain the Gaussian function corresponding to each expanded target point, and calculates the sum of the Gaussian functions of each expanded target point.

[0037] Finally, the server optimizes the target search probability map based on the sum of the Gaussian functions of each expanded target point through the following formula to obtain an optimized target search probability map; ; ; where, represents the total number of expanded target points, represents the Gaussian function corresponding to the th expanded target point, represents the The specific position of a target point Indicates the distribution range of the nth target point, where the value of is the target search probability at represents the optimized target search probability map.

[0038] In one example, the optimized target search probability map can be as shown in Figure 3 , and compared with the target search probability map ( Figure 2 ), the information contained in the optimized target search probability map is more accurate. Correspondingly, Figure 3 the X-axis in represents the abscissa of the search area corresponding to the target search task, the Y-axis represents the ordinate of the search area corresponding to the target search task, and the Z-axis represents the optimized target search probability corresponding to any position.

[0039] Step 105: Perform non-parametric data-driven estimation on the optimized target search probability map, and then analyze each cluster to quantitatively extract several valuable search areas.

[0040] In a specific implementation, after optimizing the target search probability map, the server needs to perform non-parametric data-driven estimation on the optimized target search probability map and then analyze each cluster to quantitatively extract several valuable search areas. Cluster extraction can identify the cluster areas of potential targets, thereby concentrating search resources and avoiding ineffective searches. In a marine environment, targets are usually non-uniformly distributed, so cluster extraction can help the unmanned system focus on several key areas to maximize search efficiency. By analyzing the clusters, the system can determine the high-probability areas (valuable search areas) and prioritize access. This process not only reduces the overall search time but also improves the detection success rate of the target. In addition, cluster extraction also provides a basis for subsequent strategy adjustment to ensure that the unmanned system can respond flexibly in a dynamic environment.

[0041] In one example, the server first performs non-parametric data-driven estimation on the optimized target search probability map through the following formula to obtain the non-parametric data-driven estimation result: ; where represents the number of environmental changes, is the preset probability decrease value due to environmental changes, represents the non-parametric data-driven estimation result.

[0042] Subsequently, based on the non-parametric data-driven estimation results, the server analyzes each cluster to ensure that each cluster contains adjacent points with positive values and that all clusters are isolated from each other, thereby quantitatively extracting several valuable search regions. Among them, as increases, gradually decreases, and the size of the valuable search region also shrinks accordingly.

[0043] In one example, the process of cluster extraction can be as shown in Figure 4 , and the valuable search regions extracted by the server can be as shown in Figure 5 . In Figure 4 and Figure 5 , the X-axis represents the abscissa of the search region corresponding to the target search task, and the Y-axis represents the ordinate of the search region corresponding to the target search task. In Figure 4 and Figure 5 , two connected domains, namely connected domain 1 and connected domain 2, are clearly divided.

[0044] Step 106: Instruct the maritime unmanned cluster to perform target search according to the priority order of each valuable search region. Among them, the greater the target probability of the valuable search region, the higher its corresponding priority.

[0045] In specific implementation, after the server extracts each valuable search region, it can instruct the maritime unmanned cluster to perform target search according to the priority order of each valuable search region. Among them, the greater the target probability of the valuable search region, the higher its corresponding priority.

[0046] In one example, to further improve the search efficiency, multiple mobile marine unmanned systems can work collaboratively. These unmanned systems dynamically adjust their respective search paths and task assignments by sharing real-time data and the optimized target search probability map. This way of cluster collaboration enables the unmanned systems to cooperate efficiently in the vast sea area and ensures flexible response in complex environments. In actual operation, communication and collaboration among the unmanned systems are crucial. By establishing a reliable communication network, the system can achieve real-time sharing of data, ensuring that each unmanned system can obtain the latest environmental information and target probability map. This information sharing not only improves the decision-making ability of individual unmanned systems but also enhances the collaborative effect of the entire cluster.

[0047] During the search process, the sensors built into the unmanned system can monitor the changes in the marine environment in real time, including water flow, temperature, acoustic signals, etc. These real-time data will be fed back to the control center (server), enabling the unmanned system to adjust the search strategy in a timely manner according to the environmental changes. For example, if the environmental conditions in a certain area change drastically, the unmanned system can immediately evaluate the possibility of the existence of the target in that area, and thus decide whether to adjust the search path or strategy. Through such real-time monitoring, the unmanned system can continuously optimize the target probability map and search strategy, making the search process more intelligent and efficient. This flexibility ensures that the unmanned system can maintain high-efficiency target search capabilities in a complex and ever-changing marine environment.

[0048] In one example, after completing the target search task, the server can also analyze and evaluate the results of the entire search process, and conduct a comprehensive evaluation based on the estimated rewards, sailing time, and coverage time of each valuable search area, respectively obtaining the comprehensive evaluation values of each valuable search area, so as to continuously optimize future search strategies. This continuous learning and optimization mechanism not only improves the success rate of the search task, but also provides reliable support for subsequent marine exploration and resource development. By analyzing historical data, the server can identify the key factors affecting the search efficiency and make adjustments.

[0049] For example, if it is found that a certain search strategy is ineffective under specific conditions, the server will be able to record this experience and improve it in future tasks. This self-optimization ability enables the unmanned system to maintain high-efficiency search capabilities in a constantly changing environment.

[0050] In one example, as Figure 6 shown, the search accuracy of the unmanned system continuously improves over time, as Figure 7 shown, the cumulative reward value of the search task also continuously improves over time.

[0051] In one example, the server records the comprehensive evaluation value of the th valuable search area as , The calculation process of which can be expressed by the formula: ; where represents the estimated reward of the th valuable search area, is calculated based on the probability integral of each point in the th valuable search area, represents the sailing time of the th valuable search area, represents the coverage time of the th valuable search area.

[0052] In this embodiment, the prior information obtained is first comprehensively analyzed to obtain several target points corresponding to the current target search task. Subsequently, Gaussian modeling is performed according to the specific positions and distribution ranges of the target points. Using the Parzen window theory and non-parametric data-driven estimation, the Gaussian functions obtained by modeling are superimposed to generate a target search probability map corresponding to the current target search task. Then, based on the reliability of each target point, the prior set is expanded, and the target search probability map is optimized based on the sum of the Gaussian functions of the expanded target points obtained by the expansion. Non-parametric data-driven estimation is performed on the optimized target search probability map, and then each cluster is analyzed to quantitatively extract several valuable search regions. Finally, the unmanned marine cluster is instructed to perform target search according to the priority order of each valuable search region. Such a search process can not only significantly improve the efficiency and accuracy of target search, but also provide a more intelligent solution for target recognition in complex marine environments. Through the full utilization of prior information, the scientific construction of the target probability map, the effective application of clustering extraction, and the collaborative work of unmanned systems, this embodiment has opened up a new path for the field of marine target search. In the future, with the continuous progress of technology, especially the development of artificial intelligence and big data analysis technologies, this method is expected to further improve its intelligent level and achieve more accurate and efficient target search. This will not only provide strong technical support for the development and protection of marine resources, but also provide important data bases and tools for marine scientific research. This embodiment has broad application prospects and can play an important role in multiple fields, including but not limited to marine resource exploration, environmental monitoring, marine safety and rescue, etc., and has extremely high social value and economic benefits. Through continuous adjustment and optimization in practical applications, a method for target search of an unmanned marine cluster based on a prior map proposed in this embodiment will become an important technical support in the future field of marine exploration and provide a strong guarantee for realizing more intelligent marine management and development.

[0053] The step division of the above various methods is only for clear description. When implemented, they can be combined into one step or some steps can be split into multiple steps. As long as the same logical relationship is included, it is within the protection scope of this application; adding insignificant modifications to the algorithm or process or introducing insignificant designs, but not changing the core design of its algorithm and process are all within the protection scope of this application.

[0054] Another embodiment of this application proposes a system for target search of an unmanned marine cluster based on a prior map. The details of the system for target search of an unmanned marine cluster based on a prior map proposed in this embodiment are specifically described below. The following content is only implementation details provided for easy understanding and is not necessary for implementing this example. Figure 8It is a schematic diagram of a maritime unmanned cluster target search system based on a prior map proposed in this embodiment, including: a prior analysis module 201, a Gaussian modeling module 202, a target search probability map generation module 203, a target search probability map optimization module 204, a valuable search area extraction module 205, and a search execution module 206.

[0055] The prior analysis module 201 is used to obtain prior information, comprehensively analyze the obtained prior information, and obtain several target points corresponding to the current target search task. Among them, the prior information includes historical search data, environmental monitoring records, and expert analysis opinions, and the several target points form a prior set corresponding to the current target search task.

[0056] The Gaussian modeling module 202 is used to perform Gaussian modeling according to the specific positions and distribution ranges of each target point to obtain several Gaussian functions.

[0057] The target search probability map generation module 203 is used to superimpose each Gaussian function by using the Parzen window theory and non-parametric data-driven estimation to generate a target search probability map corresponding to the current target search task.

[0058] The target search probability map optimization module 204 is used to expand the prior set based on the reliability of each target point, and optimize the target search probability map based on the sum of the Gaussian functions of the expanded target points obtained by the expansion to obtain an optimized target search probability map.

[0059] The valuable search area extraction module 205 is used to perform non-parametric data-driven estimation in the optimized target search probability map, and then analyze each cluster to quantitatively extract several valuable search areas.

[0060] The search execution module 206 is used to instruct the maritime unmanned cluster to perform target search according to the priority order of each valuable search area. Among them, the greater the target probability of the valuable search area, the higher its corresponding priority.

[0061] It is worth mentioning that each module involved in this embodiment is a logical module. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, in order to highlight the innovative part of this application, units that are not closely related to solving the technical problems proposed in this application are not introduced in this embodiment, but this does not mean that there are no other units in this embodiment.

[0062] It is not difficult to find that this embodiment is a system embodiment corresponding to the above method embodiment. This embodiment can be implemented in cooperation with the above method embodiment. The relevant technical details and technical effects mentioned in the above method embodiment are still valid in this embodiment. To avoid repetition, they will not be elaborated here. Correspondingly, the relevant technical details mentioned in this embodiment can also be applied to the above method embodiment.

[0063] Another embodiment of the present application proposes an electronic device, and its specific structure can be as Figure 9 shown, including: at least one processor 301; and a memory 302 communicatively connected to the at least one processor 301; wherein, the memory 302 stores instructions executable by the at least one processor 301, and the instructions are executed by the at least one processor 301 to enable the at least one processor 301 to execute a method for searching for targets of an unmanned marine cluster based on a priori graph as described in the above method embodiment.

[0064] Among them, the memory and the processor are connected by a bus. The bus can include any number of interconnected buses and bridges, and the bus connects various circuits of one or more processors and memories together. The bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art, so they will not be further described herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be an element or multiple elements, such as multiple receivers and transmitters, and provides a unit for communicating with various other devices on the transmission medium. The data processed by the processor is transmitted over the wireless medium through the antenna. Further, the antenna also receives data and transmits the data to the processor.

[0065] The processor is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interface, voltage regulation, power management, and other control functions. The memory can be used to store the data used by the processor when executing operations.

[0066] Another embodiment of the present application proposes a computer-readable storage medium storing a computer program, which when executed by a processor, can implement a method for searching for targets of an unmanned marine cluster based on a priori graph as described in the above method embodiment.

[0067] That is, those skilled in the art can understand that all or part of the steps in the methods of the above embodiments can be completed by instructing relevant hardware through a program. The program is stored in a storage medium, including several instructions to enable a device (which can be a single-chip microcomputer, a chip, etc.) or a processor to execute all or part of the steps of the methods described in various embodiments of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM (Read-Only Memory), RAM (Random Access Memory), magnetic disks, or optical discs.

[0068] Those of ordinary skill in the art can understand that the above embodiments are specific embodiments for implementing the present application. In actual applications, various changes can be made to them in form and details without departing from the spirit and scope of the present application.

Claims

1. A method for searching unmanned cluster targets at sea based on a priori graph, characterized in that: include: Acquire prior information, conduct a comprehensive analysis on the acquired prior information, and obtain a number of target points corresponding to this target search task; wherein the prior information includes historical search data, environmental monitoring records and expert analysis opinions, and a number of target points constitute a prior set corresponding to this target search task; Gaussian modeling is performed according to the specific location and distribution range of each target point to obtain several Gaussian functions; Using Parzen window theory and non-parametric data-driven estimation, each Gaussian function is superimposed to generate a target search probability map corresponding to this target search task; The prior set is expanded based on the reliability of each target point, and the target search probability map is optimized based on the sum of the Gaussian functions of the expanded target points to obtain an optimized target search probability map; Non-parametric data-driven estimation is performed in the optimized target search probability map, and then each cluster is analyzed to quantitatively extract several valuable search areas; Instruct the unmanned marine swarm to search for targets in the order of priority of each valuable search area; the greater the probability of a target in a valuable search area, the higher its corresponding priority.

2. The method for searching unmanned marine cluster targets based on a priori graph according to claim 1 is characterized in that: Prior information also includes relevant literature and unverified indication reports, which are likely to contain preliminary clues to the existence of the target; Historical search data is used to provide the frequency and location of target appearances, providing a basis for assessing whether there are potential targets in the search area. Environmental monitoring records are used to provide the impact of changing ocean conditions on target distribution. Changes in ocean conditions include but are not limited to tides, current speeds and temperatures. Expert analysis opinions are used to provide a professional perspective for the interpretation of relevant literature and unverified indication reports.

3. The method for searching unmanned marine cluster targets based on a priori graph according to claim 1 is characterized in that: After obtaining the prior information and comprehensively analyzing the obtained prior information to obtain a number of target points corresponding to the target search task, the method further includes: Define three key attributes for each target point, namely, specific location, reliability and distribution range. The specific location of the target point is , the reliability is , the distribution range is ; The specific location is used to characterize the central location of the expected target, which is evaluated based on historical search data; Reliability is used to characterize the credibility of the specific location of the target point. It is a quantitative assessment of the probability of the expected target appearing at a specific location, which is obtained by analyzing historical search data; The distribution range is used to characterize the possible range of the area where the expected target exists, which is set based on the characteristics of the expected target and the environmental conditions in the environmental monitoring records to ensure a comprehensive understanding of the potential distribution of the expected target.

4. The method for searching unmanned marine cluster targets based on a priori graph according to claim 3 is characterized in that: Gaussian modeling is performed according to the specific location and distribution range of each target point to obtain several Gaussian functions, which are implemented by the following formula: ； ； in, Indicates The Gaussian function corresponding to the target point is Indicates The specific location of the target point, Indicates The distribution range of target points, is the total number of target points, is the independent variable, representing any position in the search area corresponding to the target search task.

5. The method for searching unmanned marine cluster targets based on a priori graph according to claim 4 is characterized in that: The prior set is expanded based on the reliability of each target point, and the target search probability map is optimized based on the sum of the Gaussian functions of the expanded target points to obtain an optimized target search probability map, including: Based on the preset maximum expansion number and the reliability of each target point, each target point is expanded to obtain a number of expanded target points. The number of expanded target points obtained by expanding different target points is different. The number of extended target points obtained by expanding target points is expressed by the formula: ； ； in, Indicates The reliability of the target point, For preset The maximum number of expansions when Indicates that based on The number of extended target points obtained by expanding the target points; Gaussian modeling is performed according to the specific location and distribution range of each extended target point to obtain the Gaussian function corresponding to each extended target point, and the sum of the Gaussian functions of each extended target point is calculated; The target search probability map is optimized based on the sum of the Gaussian functions of each extended target point through the following formula to obtain an optimized target search probability map; ； ； in, Indicates the total number of expansion target points, Indicates Gaussian function corresponding to the extended target point, Indicates The specific location of the target point, Indicates The distribution range of target points, represents the unoptimized target search probability map, The value of The target search probability at Represents the optimized target search probability map.

6. The method for searching unmanned marine cluster targets based on a priori graph according to claim 1 is characterized in that: Non-parametric data-driven estimation is performed in the optimized target search probability map, and each cluster is analyzed to quantitatively extract several valuable search areas, including: The non-parametric data-driven estimation is performed in the optimized target search probability map through the following formula to obtain the non-parametric data-driven estimation result: ； in, Indicates the number of times the environment changes. is the preset probability decrease value caused by environmental changes, Represents the nonparametric data-driven estimation results; Based on the nonparametric data-driven estimation results, each cluster is analyzed to ensure that each cluster contains incidental positive values. All clusters are isolated from each other, thus quantitatively extracting several valuable search areas; Among them, with increase, As it decreases, the size of the valuable search area decreases.

7. A method for searching unmanned marine cluster targets based on a priori graph according to any one of claims 1 to 6, characterized in that: After instructing the unmanned marine cluster to search for targets according to the priority order of each valuable search area, the method further includes: Based on the estimated return, navigation time and coverage time of each valuable search area, a comprehensive evaluation is performed to obtain the comprehensive evaluation value of each valuable search area. The comprehensive evaluation value of the valuable search area is , The calculation process is expressed by the formula: ； in, Indicates The estimated return of valuable search areas, Based on The probability integral of each point in the valuable search area is calculated. Indicates The time of navigation for each valuable search area, Indicates The coverage time of a valuable search area.

8. A marine unmanned cluster target search system based on a priori graph, characterized in that: include: The prior analysis module is used to obtain prior information, conduct a comprehensive analysis on the obtained prior information, and obtain several target points corresponding to this target search task. The prior information includes historical search data, environmental monitoring records and expert analysis opinions. Several target points constitute the prior set corresponding to this target search task. The Gaussian modeling module is used to perform Gaussian modeling according to the specific location and distribution range of each target point to obtain several Gaussian functions; The target search probability map generation module is used to use the Parzen window theory and non-parametric data-driven estimation to superimpose various Gaussian functions to generate the target search probability map corresponding to this target search task; A target search probability map optimization module is used to expand the prior set based on the reliability of each target point, and optimize the target search probability map based on the sum of Gaussian functions of the expanded target points to obtain an optimized target search probability map; The valuable search region extraction module is used to perform non-parametric data-driven estimation in the optimized target search probability map, and then analyze each cluster to quantitatively extract a number of valuable search regions; The search execution module is used to instruct the unmanned marine cluster to search for targets according to the priority order of each valuable search area, wherein the greater the target probability of the valuable search area, the higher its corresponding priority.

9. An electronic device, characterized in that: include: at least one processor; and, a memory communicatively coupled to the at least one processor; In which, the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a method for searching unmanned marine cluster targets based on a priori graph as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it can implement a method for searching unmanned marine cluster targets based on a priori graph as described in any one of claims 1 to 7.

Images