Unmanned swarm search and rescue mission planning method, system, equipment and storage medium

Abstract

The present application discloses a method, system, device and storage medium for planning an unmanned cluster search and rescue mission, which relates to the field of search and rescue technology. The method includes: analyzing multi-source distress signals in the search and rescue area to determine the accident point information; dynamically updating the target discovery probability of each grid sub-area in the search and rescue area based on the environmental feature information and target movement feature information of the search and rescue area; determining the current task planning strategy based on the number of types of unmanned equipment in the unmanned cluster and the accident point information; determining the regional search and rescue priority list of each unmanned equipment based on the current task planning strategy and the target discovery probability of the grid sub-area, and performing task allocation for the unmanned cluster. Through the above method, combined with environmental features and target movement features, unified task planning and path allocation are performed for multiple types of unmanned equipment, which quickly covers the accident point, improves the search and rescue efficiency, and increases the success rate of the search and rescue mission.

Description

Technical Field

[0001] The present application relates to the field of maritime search and rescue technology, and in particular to unmanned swarm search and rescue mission planning methods, systems, equipment, and storage media. Background Art

[0002] In recent years, maritime search and rescue technology has garnered widespread attention for its role in ensuring the safety of people and property. Unmanned aerial vehicles, with their rapid and flexible maneuverability and relatively low operating costs, demonstrate potential for patrol and rescue missions across large maritime areas. However, current maritime search and rescue efforts primarily rely on manual search or single-device assistance. This approach suffers from low efficiency, delayed response times, and limited coverage, making it difficult to meet the search and rescue needs of today's complex maritime environments.

[0003] Currently, unmanned swarm-based maritime search technology has yet to be implemented. While individual devices like drones and unmanned boats have been applied in certain areas, they can only provide limited support and lack a comprehensive system for the coordinated operation of multiple unmanned devices. Furthermore, algorithms for search and rescue mission planning and allocation are virtually nonexistent, and no method has been developed specifically for optimizing the tasks of heterogeneous unmanned swarms in dynamic maritime environments.

[0004] Maritime search operations are highly complex, requiring comprehensive consideration of multiple dynamic and environmental factors. For one thing, targets in distress constantly drift under the influence of wind, waves, and tides, making their positions unpredictable over time. Furthermore, unique terrain features such as shallow islands and areas with complex currents further complicate search and rescue missions. These characteristics require that search and rescue mission planning not only efficiently handle multiple incident points and the coordination of multiple equipment, but also be able to respond to dynamic environments in real time.

[0005] The above content is only used to assist in understanding the technical solution of the present invention and does not constitute an admission that the above content is prior art. Summary of the Invention

[0006] The main purpose of this application is to provide an unmanned swarm search and rescue mission planning method, system, equipment and storage medium, aiming to solve the technical problem in the existing technology that the mission planning scheme of unmanned swarms in complex search and rescue scenarios has poor search and rescue efficiency, affecting the success rate of search and rescue missions.

[0007] To achieve the above objectives, this application provides a method for planning an unmanned swarm search and rescue mission, which includes:

[0008] Analyze multiple distress signals in the search and rescue area to determine the accident point information;

[0009] Dynamically updating the target discovery probability of each grid sub-area in the search and rescue area based on the environmental feature information and target movement feature information of the search and rescue area;

[0010] Determining a current mission planning strategy based on the types and number of unmanned equipment in the unmanned swarm and the accident point information;

[0011] Determining a regional search and rescue priority list for each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area;

[0012] Tasks are allocated based on the regional search and rescue priority list, scenario adaptability, and execution cost of each unmanned equipment to determine the target search and rescue area, execution sequence, and search and rescue path of each unmanned equipment.

[0013] In one embodiment, the accident point information includes at least the number of accident points and the scene type of the accident points. The step of analyzing the multi-source distress signals in the search and rescue area to determine the accident point information includes:

[0014] Extracting spatiotemporal features from multi-source distress signals in the area to be searched and rescued;

[0015] Inputting the spatiotemporal characteristics into an accident point quantity recognition model for recognition, and determining the number of accident points;

[0016] extracting scene features from multi-source distress signals in the area to be searched and rescued;

[0017] The scene features are input into the accident point scene classification model for classification to determine the accident point scene type.

[0018] In one embodiment, the step of determining the current mission planning strategy based on the number and types of unmanned equipment in the unmanned cluster and the accident point information includes:

[0019] When the number of types of the unmanned equipment is greater than or equal to a preset type number threshold, determining that the current mission planning strategy is a mission planning strategy based on a competition mechanism;

[0020] When the number of types of the unmanned equipment is less than a preset type number threshold and the accident point information does not conform to a multiple accident point scenario, determining that the current task planning strategy is a task planning strategy based on probability sorting;

[0021] When the number of types of the unmanned equipment is less than a preset type number threshold and the accident point information meets the multi-accident point scenario, the current task planning strategy is determined to be a task planning strategy based on probability and speed sorting.

[0022] In one embodiment, the step of determining the regional search and rescue priority list for each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area includes:

[0023] When the current mission planning strategy is a mission planning strategy based on probability and speed sorting, obtaining a correspondence between target discovery probability, area distance, speed of the unmanned equipment, and search and rescue priority;

[0024] Determining the search and rescue priority of the unassigned area corresponding to each unassigned area based on the target discovery probability of the unassigned area in the grid sub-area, the speed of the unmanned equipment, the area distance between the unmanned equipment and the unassigned area, and the corresponding relationship;

[0025] The search and rescue priorities of the unassigned areas corresponding to each unmanned equipment are sorted in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

[0026] In one embodiment, the step of determining the regional search and rescue priority list for each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area includes:

[0027] When the current mission planning strategy is a mission planning strategy based on a competition mechanism, obtaining a corresponding relationship between the target discovery probability, the area distance, the speed of the unmanned equipment, and the search and rescue priority;

[0028] Determining the search and rescue priority of the grid sub-area corresponding to each unmanned device based on the target discovery probability of the grid sub-area, the speed of the unmanned device, the regional distance between the unmanned device and the grid sub-area, and the corresponding relationship;

[0029] The search and rescue priorities of the grid sub-areas corresponding to each unmanned equipment are sorted in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

[0030] In one embodiment, the step of determining the regional search and rescue priority list for each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area includes:

[0031] When the current mission planning strategy is a mission planning strategy based on probability sorting, the target discovery probability of the unassigned area in the grid sub-area is used as the search and rescue priority of the corresponding unassigned area;

[0032] The search and rescue priorities of the unassigned areas corresponding to each unmanned equipment are sorted in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

[0033] In one embodiment, the step of allocating tasks based on the regional search and rescue priority list, scenario adaptability, and execution cost of each unmanned equipment, and determining the target search and rescue area, execution sequence, and search and rescue path of each unmanned equipment includes:

[0034] Based on the regional search and rescue priority list, mapping the grid sub-region into a task set and determining the weight of each task in the task set;

[0035] Constructing a task allocation variable, and constructing a comprehensive benefit function of the unmanned swarm based on the task allocation variable, task weight, scenario adaptability, execution cost, and balance coefficient;

[0036] Based on the comprehensive benefit function and the corresponding constraints, the optimal allocation strategy is solved to obtain the target search and rescue area, execution sequence and search and rescue path of each unmanned equipment.

[0037] In addition, to achieve the above objectives, this application also proposes an unmanned swarm search and rescue mission planning system, which includes:

[0038] The scene recognition module is used to analyze the multi-source distress signals in the search and rescue area and determine the accident point information;

[0039] A probability calculation module, configured to dynamically update the target discovery probability of each grid sub-area in the search and rescue area based on the environmental characteristic information and target movement characteristic information of the search and rescue area;

[0040] A strategy selection module, configured to determine a current mission planning strategy based on the types and number of unmanned equipment in the unmanned swarm and the accident point information;

[0041] A mission planning module, configured to determine a regional search and rescue priority list for each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area;

[0042] The task planning module is also used to allocate tasks based on the regional search and rescue priority list, scene adaptability and execution cost of each unmanned equipment, and determine the target search and rescue area, execution sequence and search and rescue path of each unmanned equipment.

[0043] In addition, to achieve the above-mentioned purpose, the present application also proposes an unmanned cluster search and rescue mission planning device, which includes: a memory, a processor, and a computer program stored in the memory and runnable on the processor. The computer program is configured to implement the steps of the unmanned cluster search and rescue mission planning method as described above.

[0044] In addition, to achieve the above-mentioned purpose, the present invention also proposes a storage medium, which is a computer-readable storage medium. A computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the unmanned cluster search and rescue mission planning method as described above are implemented.

[0045] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the unmanned cluster search and rescue mission planning method as described above.

[0046] The present application provides an unmanned cluster search and rescue mission planning method, which analyzes multi-source distress signals in the search and rescue area to determine the accident point information; dynamically updates the target discovery probability of each grid sub-area in the search and rescue area based on the environmental feature information and target movement feature information of the search and rescue area; determines the current mission planning strategy based on the type and number of unmanned equipment in the unmanned cluster and the accident point information; determines the regional search and rescue priority list of each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area; allocates tasks based on the regional search and rescue priority list of each unmanned equipment, the scene adaptability and the execution cost, and determines the target search and rescue area, execution sequence and search and rescue path of each unmanned equipment. The environmental characteristics and target movement characteristics are updated in real time, and combined with planning strategies in different scenarios, task planning and path allocation are unified for multiple types of unmanned equipment. The speed, remaining energy and environmental adaptability of each equipment can be comprehensively considered, and multiple accident points, different scenarios, and dynamic changes in the environment and distressed targets can be considered to calculate the optimal or near-optimal task planning scheme in real time, thereby achieving rapid coverage of multiple accident points and timely response to emergencies, improving the search and rescue efficiency and accuracy in complex search and rescue scenarios, and improving the success rate of search and rescue missions. It solves the technical problem that the task planning scheme of unmanned clusters in complex search and rescue scenarios has poor search and rescue efficiency, affecting the success rate of search and rescue missions. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0048] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0049] Figure 1 This is a flowchart of the first embodiment of the unmanned swarm search and rescue mission planning method of this application;

[0050] Figure 2 A schematic diagram of target discovery probability calculation for the unmanned swarm search and rescue mission planning method provided in Example 1 of the present application;

[0051] Figure 3A schematic diagram of the convolution kernel coefficient calculation for the unmanned swarm search and rescue mission planning method provided in Example 1 of the present application;

[0052] Figure 4 This is a flow chart of Example 2 of the unmanned swarm search and rescue mission planning method of this application;

[0053] Figure 5 A schematic diagram of a simplified process flow of the unmanned swarm search and rescue mission planning method provided in Example 2 of this application;

[0054] Figure 6 This is a schematic diagram of the module structure of the unmanned swarm search and rescue mission planning system according to an embodiment of the present application;

[0055] Figure 7 This is a schematic diagram of the device structure of the hardware operating environment involved in the unmanned cluster search and rescue mission planning method in the embodiment of the present application.

[0056] The realization of the objectives, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0057] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0058] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0059] The main solutions of the embodiments of the present application are: analyzing multi-source distress signals in the search and rescue area to determine the accident point information; dynamically updating the target discovery probability of each grid sub-area in the search and rescue area based on the environmental feature information and target movement feature information of the search and rescue area; determining the current task planning strategy based on the type and number of unmanned equipment in the unmanned cluster and the accident point information; determining the regional search and rescue priority list of each unmanned equipment based on the current task planning strategy and the target discovery probability of the grid sub-area; allocating tasks based on the regional search and rescue priority list, scene adaptability and execution cost of each unmanned equipment, and determining the target search and rescue area, execution sequence and search and rescue path of each unmanned equipment.

[0060] The present application provides a solution that updates environmental characteristics and target movement characteristics in real time, and combines planning strategies in different scenarios to uniformly plan tasks and allocate paths for multiple types of unmanned equipment. It can comprehensively consider the speed, remaining energy and environmental adaptability of each equipment, and consider multiple accident points, different scenarios, and the dynamic changes of the environment and distressed targets, and calculate the optimal or near-optimal task planning scheme in real time, thereby achieving rapid coverage of multiple accident points and timely response to emergencies, improving the search and rescue efficiency and accuracy in complex search and rescue scenarios, and improving the success rate of search and rescue missions. It solves the technical problem that the task planning scheme of unmanned clusters in complex search and rescue scenarios has poor search and rescue efficiency, which affects the success rate of search and rescue missions.

[0061] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication, and program execution capabilities, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the aforementioned functions, such as an unmanned swarm search and rescue mission planning device, and this embodiment does not specifically limit this. The following uses the unmanned swarm search and rescue mission planning device as an example to illustrate this embodiment and the following embodiments.

[0062] In addition, the unmanned swarm is a collaborative operation system composed of various types of unmanned equipment, which can be applied to various search and rescue scenarios, such as deserts, seas, ruins, etc. This embodiment uses the sea search and rescue scenario as an example for explanation.

[0063] The present invention provides a method for planning an unmanned swarm search and rescue mission. Figure 1 , Figure 1 This is a flow chart of the first embodiment of the unmanned swarm search and rescue mission planning method of the present application.

[0064] In this embodiment, the unmanned swarm search and rescue mission planning method includes steps S10 to S50:

[0065] Step S10, analyzing the multi-source distress signals in the search and rescue area to determine the accident point information;

[0066] It should be noted that the search and rescue area, i.e., the area where the unmanned swarm is needed for search and rescue, is the area where the target in distress may be located. This area can be a single area at sea or multiple areas at sea, with no specific restrictions. The target in distress, i.e., the target in distress that is in need of rescue due to an emergency, can be a vessel or a person in distress, with no specific restrictions.

[0067] In addition, it should be noted that multi-source distress signals are distress signals received in the area to be searched and rescued. Distress signals generally include two categories, namely active distress signals and passive discovery signals. Active distress signals are data sent by the target in distress through a distress device (such as an SOS beacon, radio station, etc.), and passive discovery signals are data generated when a suspicious target or abnormal situation is automatically identified by a patrol or monitoring device (such as a drone patrol video, sonar detection, shore camera, radar, etc.). Since there are many ways to generate distress signals, and different methods generate different data, it can be considered that the distress signals are multi-source. Generally speaking, the area to be searched and rescued can be determined based on the received distress signal, and a suitable range can be divided as the area to be searched and rescued according to the location where the distress signal is sent.

[0068] It is understandable that the accident point is the location where the distress situation occurs, which can be considered as the location / position from which the distress signal was sent. Accident point information is relevant information about the accident point, including at least the number of accident points and the accident point scene type. The number of accident points is the number of accident points. If the number of accident points is 1, it is generally considered a single accident point. If the number of accident points is greater than 1, it is generally considered to be multiple accident points. It can be seen that the number of accident points can be divided into two situations: single accident point and multiple accident points. The accident point scene type is the category of the accident point scene, such as: sea surface, shallow underwater, island, etc., and there is no specific limitation on this.

[0069] It should be understood that this embodiment analyzes the received multi-source distress signals to determine the number of accident points and the types of accident scene, thereby determining the search and rescue environment, and providing accurate basis for environmental perception and accident point distribution for subsequent path planning and unmanned equipment scheduling.

[0070] Step S20, dynamically updating the target discovery probability of each grid sub-area in the search and rescue area based on the environmental characteristic information and target movement characteristic information of the search and rescue area;

[0071] It should be noted that the members of an unmanned swarm (MoUS) can include various types of unmanned equipment, such as unmanned aerial vehicles (UAVs), unmanned ships (USVs), robot dogs (DOGs), and unmanned underwater vehicles (UUVs). All members can perform search and rescue missions individually or in any combination. In this embodiment, the unmanned equipment includes at least one of a UAV, an unmanned ship, a robot dog, and an unmanned underwater vehicle. The unmanned swarm can be composed of a single type of unmanned equipment, such as a swarm of UAVs, or multiple (two or more) types of unmanned equipment, such as a swarm of UAVs and an unmanned ship. The specific unmanned equipment used will be determined based on actual circumstances and is not limited to this.

[0072] Understandably, unmanned boats can serve as relay charging platforms for drones and as deployment and recovery centers for unmanned submersibles (UUVs are activated only when they detect shallow underwater scenes). They can also independently perform surface search and rescue missions. Drones can conduct large-scale patrols and reconnaissance in cooperative mode. Robot dogs and UUVs can search island areas and shallow underwater areas, respectively, forming a three-dimensional collaborative network with drones and unmanned boats. The robot dog is deployed and recovered by the drone (it is activated only when it detects an island scene). In cooperative mode, the two can conduct large-scale patrols and searches. As can be seen, drones, unmanned boats, UUVs, and robot dogs each have different operating modes: drones are suitable for aerial reconnaissance of the sea surface / islands, with fast patrol speeds but limited endurance; unmanned boats have slightly slower maneuverability and can serve as deployment and charging platforms; UUVs have underwater sonar search capabilities and can detect distressed targets in shallow waters, requiring deployment from unmanned boats; robot dogs have strong search capabilities on the surface of islands and require drone deployment.

[0073] Compared to using only a single type of unmanned equipment, the use of unmanned swarms can more effectively leverage the unique characteristics of different types of unmanned equipment, avoiding the limitations of a single type of unmanned equipment in search and rescue operations. For example, unmanned boats are relatively slow, but by collaborating with drones, the drone's speed can be fully utilized, expanding the boat's search and rescue range and improving search efficiency. Drones have limited flight range, but unmanned boats can serve as charging platforms for drones. Before a drone runs out of power, it will fly to the nearest unmanned boat to recharge, thus compensating for the drone's limited range and enabling extended searches in a designated sea area.

[0074] In the specific implementation, the feasibility and dependency relationship between unmanned equipment are determined according to the number of accident points and the type of accident scene, so as to determine the type and quantity of unmanned equipment required, and then decide on the unmanned cluster used for search and rescue.

[0075] It should be noted that the environmental characteristic information, i.e., the environmental characteristics of the area to be searched and rescued, includes at least terrain information, wind and wave information, tidal information, and ocean current information. The terrain information may include information such as the sea surface range, island terrain, and shallow underwater terrain. In addition to wind and wave information, tidal information, and ocean current information, other sea condition information may also be provided, which is not specifically limited in this embodiment. Since the target in distress may be attached to floating objects or lifeboats, the search and rescue target is usually mobile, and its motion trajectory will be similar to that of floating objects. The target movement characteristic information, i.e., the movement characteristics of the target in distress within the area to be searched and rescued, includes at least the speed of the target in distress itself and the convolution kernel parameters after being affected by ocean currents, tides, and island blocking effects. These data can be used as the initial possible diffusion direction and range of the target in distress.

[0076] In addition, it should be noted that the area to be searched and rescued can be divided into a plurality of grid units, each grid unit being a sub-area of ​​the area to be searched and rescued, that is, a grid sub-area. The size of the grid sub-area is set according to the actual situation and is not limited to this. The probability of target discovery (P-DSRT) is the probability of the distress target appearing, that is, the probability of discovering the distress target, which is recorded as .

[0077] It is understandable that after receiving the initial distress signal, there may be a probability of finding the target in distress within multiple grid sub-areas. However, over time, the target discovery probability in the grid sub-areas will continue to spread around the initial location. Initially, the target discovery probability is highest at the location where the initial distress signal was sent. The longer the search and rescue time, the more likely the target in distress will appear farther away from the initial distress signal location. At this time, the trend of the target discovery probability shows an outward diffusion characteristic. When the unmanned swarm completes the search of a grid sub-area, the target discovery probability in that area will decrease. However, since the target in distress may move, the target discovery probability in the surrounding areas will shift to the area that has already been searched. In other words, the target discovery probability in the surrounding areas will decrease, and the target discovery probability in the area that has already been searched will increase.

[0078] It should be understood that the location of the target in distress may change over time. Considering the search area as a discrete space and using a grid dynamics model to simulate the diffusion of the probability of the target in distress being found in the discrete space is consistent with the movement logic of the target in distress. The target discovery probability of a certain grid sub-area will evolve based on the target discovery probability of the area at the previous moment and the target discovery probability of the surrounding areas, thereby calculating the target discovery probability of any grid sub-area at any time, and the target discovery probability of the area will also be affected by the passing of unmanned clusters. In discrete space, the target discovery probability will change over time according to the direction in which the target in distress may move, which is similar to a convolution calculation. Reference Figure 2 , combining the environmental feature information and the target movement feature information, through the iterative calculation of the convolution kernel and the grid dynamics model, the target appearance probability of the grid sub-region at the next moment is obtained. The calculation formula for the target discovery probability of the grid sub-area at a certain moment is as follows:

[0079]

[0080] Where, express The time coordinate is The target discovery probability of the grid sub-area is Represents the convolution kernel coefficient, reflecting the distress target from the adjacent grid sub-region Drift to the current grid sub-region The probability ratio is the probability that the target in distress moves to the periphery. In addition, on the boundary The relationship between the target discovery probability of the grid sub-area at the moment is: .

[0081] It can be understood that, assuming the side length of each grid sub-region is , the external environmental factors make the distress target fixed and move vector , the vector of random movement of the distress target is ,consider At a radius of The target discovery probability of the grid sub-area is uniformly distributed in the circle. . refer to Figure 3 , the probability that the target in distress moves to the surrounding area is a vector The area of ​​overlap between the circle and the corresponding grid sub-area (rectangle) is the convolution kernel coefficient. The target detection probability of the central area at the next moment is affected by the target detection probability of the nearby area. The probability of the distressed target moving to the periphery is calculated using the following relationship:

[0082]

[0083] Where, Indicates the probability that the target in distress moves to the periphery, that is, the convolution kernel coefficient, 、 are the horizontal and vertical coordinates of the area where the target in distress moves to at the next moment relative to the area at the current moment. It should be noted that if the randomly moving vector When the length is larger, a larger convolution kernel is needed to represent it; if the randomly moving vector When the length is small, the side length of the discrete space needs to be reduced accordingly ; If the vector It is not evenly distributed in the circle. The volume of the area corresponding to the convolution kernel is calculated by integrating at different positions as the probability of the distress target moving to the periphery. When the convolution kernel size is When , indicating that the target in distress will not move.

[0084] When an unmanned vehicle passes through a certain grid sub-area, it will detect the target with a certain probability, thereby reducing the residual probability of the grid sub-area. At this time, the calculation relationship of the target detection probability can be simplified to:

[0085]

[0086] Where, is the probability that the unmanned equipment can detect the target, The coordinates for the next moment are The target discovery probability of the grid sub-area is for The time coordinate is The probability of finding a target in a grid sub-area. When , it means that the target in distress can be found as long as the unmanned equipment passes by. At this time, it is only necessary to make the paths of all unmanned equipment cover the search and rescue area.

[0087] In practical applications, environmental and target movement characteristics are updated in real time, continuously revising the target detection probability of each grid sub-area for more accurate mission planning. Furthermore, grid sub-areas can be colored, with high target detection probabilities being darker and low target detection probabilities being lighter.

[0088] Step S30: determining a current mission planning strategy based on the number and types of unmanned equipment in the unmanned cluster and the accident point information;

[0089] It should be noted that the task planning strategy is the task planning algorithm, and the current task planning strategy is the task planning algorithm currently used. In this embodiment, a variety of task planning strategies are set, including at least a task planning strategy based on probability sorting (Algorithm 1), a task planning strategy based on probability and speed sorting (Algorithm 2), and a task planning strategy based on a competition mechanism (Algorithm 3).

[0090] Understandably, different scenarios require different task planning strategies. Algorithm 1 is suitable for single incident sites and homogeneous unmanned swarms, characterized by fast computation and intended for emergency response. Algorithm 2 is suitable for multiple incident sites and homogeneous unmanned swarms, optimizing for equipment speed and arrival time. When two or more types of unmanned equipment are activated, Algorithm 3 is automatically assigned for task allocation and path planning. Algorithm 3 combines cooperative and competitive mechanisms to achieve efficient coverage of multiple incident sites within the complex coupling of unmanned vessels, drones, robot dogs, and unmanned underwater vehicles.

[0091] In a feasible embodiment, step S30 may include: when the number of types of the unmanned equipment is greater than or equal to a preset type number threshold, determining that the current task planning strategy is a task planning strategy based on a competition mechanism; when the number of types of the unmanned equipment is less than the preset type number threshold and the accident point information does not conform to a multi-accident point scenario, determining that the current task planning strategy is a task planning strategy based on probability sorting; when the number of types of the unmanned equipment is less than the preset type number threshold and the accident point information conforms to a multi-accident point scenario, determining that the current task planning strategy is a task planning strategy based on probability and speed sorting.

[0092] It should be noted that the number of unmanned equipment types refers to the number of categories of unmanned equipment in the unmanned swarm. For example, if the unmanned swarm is a drone swarm, that is, the unmanned equipment in the unmanned swarm is drones, then the number of types is 1. If the unmanned swarm is a drone swarm and an unmanned ship, that is, the unmanned equipment in the unmanned swarm is drones and unmanned ships, then the number of types is 2. The preset type number threshold is a set threshold for the number of types, and the value is usually 2. If the number of unmanned equipment types is greater than or equal to 2, heterogeneous unmanned equipment is being used, that is, multiple types of unmanned equipment are being used, and a task planning strategy based on a competition mechanism is adopted. If the number of unmanned equipment types is less than 2, homogeneous unmanned equipment is being used, that is, a single type of unmanned equipment is being used, and further judgment is made based on the accident point information.

[0093] It's understandable that if the number of story points indicates a single story point, it doesn't fit the multiple-accident-point scenario. If the number of story points indicates multiple story points, it fits the multiple-accident-point scenario. If the number of unmanned equipment types is less than two and the scenario is a single story point, a task planning strategy based on probability sorting is used. If the number of unmanned equipment types is less than two and the scenario is multiple story points, a task planning strategy based on probability and speed sorting is used.

[0094] Step S40: determining a regional search and rescue priority list for each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area;

[0095] In a feasible implementation, step S40 may include steps A11 to A12:

[0096] Step A11, when the current mission planning strategy is a mission planning strategy based on probability sorting, using the target discovery probability of the unassigned area in the grid sub-area as the search and rescue priority of the corresponding unassigned area;

[0097] It should be noted that the search and rescue priority is the calculated priority of each grid sub-area. When using a mission planning strategy based on probability sorting, the search and rescue priority of a grid sub-area is the target discovery probability of the grid sub-area.

[0098] It can be understood that the core idea of ​​Algorithm 1 is to allocate areas that have not yet been searched and rescued and have the maximum target detection probability to the unmanned equipment in the unmanned cluster, so as to detect the distressed target in time and avoid the probability diffusion caused by the target movement.

[0099] Step A12: sort the search and rescue priorities of the unassigned areas corresponding to the unmanned equipment in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

[0100] It should be noted that an unassigned area is an area that has not yet been assigned to an unmanned rescue system. The regional search and rescue priority list is a list of different areas sorted by search and rescue priority. In this embodiment, the unassigned areas are sorted in descending order of search and rescue priority to form the regional search and rescue priority list for each unmanned system.

[0101] In the specific implementation, first according to the target discovery probability All grid sub-areas are sorted, and then the unassigned areas with the maximum target detection probability are allocated to the idle unmanned equipment in the unmanned cluster.

[0102] The advantage of Algorithm 1 is that it can determine the next target search and rescue area for the unmanned swarm in real time, thereby forming a coherent search and rescue path. Traditional intelligent algorithms require multiple iterations to determine the optimal path. Since only the target detection probability between grid sub-areas is compared, the calculation speed is relatively fast. However, since the search and rescue priority is determined solely by the target detection probability, the distance between the assigned unmanned equipment and the area is not considered. This may cause unmanned equipment to waste travel time when reaching the target search and rescue area, and may lead to resource waste such as overlapping planned paths within the unmanned swarm. The target search and rescue area is the area currently assigned to each unmanned device for search and rescue.

[0103] In another feasible implementation, step S40 may include steps B11 to B13:

[0104] Step B11, when the current mission planning strategy is a mission planning strategy based on probability and speed sorting, obtaining a correspondence between target discovery probability, area distance, speed of the unmanned equipment, and search and rescue priority;

[0105] It is understandable that the core idea of ​​Algorithm 2 is to comprehensively consider two factors: the time required to reach the target search and rescue area and the probability of target detection. For each unmanned vehicle waiting for a task, it can be assigned to an area with the shortest arrival time and the highest probability of target detection.

[0106] It should be noted that the corresponding relationship between the target detection probability, area distance, speed of the unmanned equipment and the search and rescue priority is the calculation formula of the search and rescue priority, as shown below:

[0107]

[0108] Where, Indicates the Unmanned equipment at all times The corresponding coordinates are The search and rescue priority of the grid sub-area, Indicates at time The coordinates are The target discovery probability of the grid sub-area is Indicates the Unmanned equipment at all times Go to the coordinates The distance between the grid sub-areas, that is, at the time No. Unmanned equipment and coordinates are The area distance between the grid sub-areas, Indicates the The speed of an unmanned device, Indicates the Unmanned equipment at all times Go to the coordinates The time required for a grid sub-region.

[0109] Step B12: determining the search and rescue priority of the unassigned area corresponding to each unassigned equipment based on the target discovery probability of the unassigned area in the grid sub-area, the speed of the unmanned equipment, the area distance between the unmanned equipment and the unassigned area, and the corresponding relationship;

[0110] It can be understood that the target discovery probability of the unallocated area in the grid sub-area, the speed of the unmanned equipment, and the area distance between the unmanned equipment and the unallocated area are substituted into the above corresponding relationship to calculate the search and rescue priority of the unallocated area.

[0111] Step B13: sort the search and rescue priorities of the unassigned areas corresponding to the unmanned equipment in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

[0112] In this embodiment, the search and rescue priorities of the unassigned areas are sorted in descending order, thereby forming a regional search and rescue priority list for each unmanned equipment.

[0113] It should be understood that according to the calculated search and rescue priority , dispatch no. The unmanned equipment is heading to the coordinates By searching the unassigned areas, the drone can reach the area with the highest probability of finding the target in the shortest time, thereby finding the target in distress earlier.

[0114] Algorithm 2 comprehensively considers the UAV's speed, the distance between the UAV and the target search and rescue area, and the target detection probability within the target search and rescue area. Therefore, in a multi-accident rescue mission, even if the target detection probability at a certain location decreases, the UAV currently searching at the current location will not be immediately dispatched to another accident site. This prevents UAVs from moving back and forth between multiple, distant accident sites, reduces the overlap of paths between UAVs within the swarm, and thus saves travel time.

[0115] In another feasible implementation, step S40 may include steps C11 to C13:

[0116] Step C11, when the current mission planning strategy is a mission planning strategy based on a competition mechanism, obtaining a corresponding relationship between the target discovery probability, the area distance, the speed of the unmanned equipment, and the search and rescue priority;

[0117] It should be noted that Algorithm 3, based on Algorithm 2, introduces a competition mechanism between different UAVs. Specifically, multiple UAVs plan their search behavior in real time based on their own speed and target detection probability in the grid subregion. The core idea is to leverage the advantages of the UAV swarm's high speed and search efficiency, reducing the time the search and rescue area is occupied by slower UAVs with longer arrival times, while also preventing other nearby UAVs from being unable to assist in the search.

[0118] It can be understood that when using a mission planning strategy based on a competition mechanism, the calculation method of the search and rescue priority is the same as that of the mission planning strategy based on probability and speed sorting.

[0119] Step C12: determining the search and rescue priority of the grid sub-area corresponding to each unmanned device based on the target discovery probability of the grid sub-area, the speed of the unmanned device, the distance between the unmanned device and the grid sub-area, and the corresponding relationship;

[0120] It can be understood that the target discovery probability of the grid sub-area, the speed of the unmanned equipment, and the regional distance between the unmanned equipment and the grid sub-area are substituted into the above corresponding relationship to calculate the search and rescue priority of the grid sub-area.

[0121] Step C13: sorting the search and rescue priorities of the grid sub-areas corresponding to the unmanned equipment in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

[0122] In this embodiment, the search and rescue priorities of all grid sub-areas are sorted in descending order, thereby forming a regional search and rescue priority list for each unmanned equipment.

[0123] It's understandable that when assigning tasks, Algorithm 3 no longer considers only idle UAVs; UAVs currently performing tasks can also participate in task allocation (including those en route to the target search and rescue area and those searching the target area). Algorithm 3 still uses the task planning framework of Algorithm 2, but when an UAV, while en route to its original target search and rescue area, discovers that the search and rescue priority of a grid sub-area is higher than that of the original target search and rescue area, the UAV will change its target search and rescue area. This allows for real-time task allocation to UAVs, rather than the moment they complete their search and rescue mission in a specific area. This helps prevent the search and rescue area from being occupied by slower UAVs. This algorithm facilitates efficient resource utilization, especially when a swarm of UAVs has varying speeds.

[0124] It should be understood that after the unmanned equipment completes the previous round of search tasks, all algorithms will assign it the highest priority area. Among them, Algorithm 1 and Algorithm 2 will search for the highest priority area from the unassigned area, but Algorithm 3 will search in all grid sub-areas. Algorithm 2 and Algorithm 3 have different sorting rules for areas than Algorithm 1. Algorithm 1 will directly compare the target discovery probability of each grid sub-area. , Algorithm 2 and Algorithm 3 will compare the search and rescue priorities ,After sorting by the algorithm, the area with the highest value will be searched first.

[0125] When the competition mechanism is triggered, if the drone cooperates with the unmanned boat for search and rescue, when the unmanned boat is closer to the target and has sufficient resources, the corresponding search task can be directly assigned to the unmanned boat for execution; if the robot dog cooperates with the drone for search and rescue, when the robot dog is closer to the target and has sufficient resources, the corresponding search task can be directly assigned to the robot dog for execution.

[0126] In this embodiment, the optimal planning algorithm is flexibly selected for the search and rescue needs of a single or multiple types of unmanned equipment. When two or more types of unmanned equipment are detected participating in the search and rescue, a task planning algorithm based on a competitive mechanism (Algorithm 3) is automatically assigned. This fully utilizes the advantages of unmanned vessels as control centers and maritime relay platforms, as well as the complementary advantages of drones, robot dogs, and unmanned underwater vehicles, to achieve three-dimensional, all-round, and efficient search and rescue in sea surface, island, reef, and shallow water environments.

[0127] Step S50 , performing task allocation based on the regional search and rescue priority list, scene adaptability, and execution cost of each unmanned equipment, and determining the target search and rescue area, execution sequence, and search and rescue path of each unmanned equipment.

[0128] In a feasible implementation, step S50 may include: based on the regional search and rescue priority list, mapping the grid sub-area into a task set, and determining the weight of each task in the task set; constructing a task allocation variable, and determining the comprehensive benefit function of the unmanned cluster based on the task allocation variable, the weight of the task, the scene adaptability, the execution cost and the balance coefficient; based on the comprehensive benefit function and the corresponding constraints, solving the optimal allocation strategy to obtain the target search and rescue area, execution sequence and search and rescue path of each unmanned equipment.

[0129] It should be noted that according to the regional search and rescue priority list, the grid sub-regions are mapped to task sets, and the search and rescue priorities calculated by the three task planning strategies for the grid sub-regions are converted into task weights. , used to indicate unmanned equipment Whether to execute the assigned task Combining the scene adaptability of the unmanned equipment, the weight of the task, the execution cost of the unmanned equipment, and the balance coefficient, a profit function is constructed as follows:

[0130]

[0131] Where, is the profit function, and is the balance coefficient, is the scene adaptability, is the execution cost (such as distance, time, energy consumption, etc.), is the weight of the task. Thus, the overall benefit function, that is, the comprehensive benefit function, can be obtained as follows:

[0132]

[0133] Where, is the profit function, Assign variables to tasks, is the comprehensive income function.

[0134] It is understandable that, when constraints, dependencies, and resource limitations are met (e.g., equipment life does not exceed a threshold), ), the comprehensive benefit function is maximized and the optimal allocation strategy (optimal task allocation plan) is solved to determine the target search and rescue area and execution sequence of each unmanned equipment and plan the search and rescue path.

[0135] Among them, the constraints for unmanned equipment to perform tasks are as follows: DOG performing island tasks requires UAV to perform them simultaneously, ; UUV needs USV to deploy, ; UAV needs USV resupply in sea scenarios, ; For infeasible scenarios (such as DOG cannot be executed underwater), you can directly Or make the profit the minimum.

[0136] It should be understood that the basic parameters of the drone, unmanned ship, robot dog and unmanned underwater vehicle can be obtained in advance, such as speed, endurance, load conditions, etc.

[0137] Furthermore, a graphical user interface (GUI) can be used to display the current position, mission status, and path planning results of each unmanned vehicle, and target detection probability in grid subareas can be plotted as a heat map. When two or more types of unmanned vehicles are detected, the GUI automatically displays "Algorithm 3" as the selected algorithm. Through the GUI, users can view the switching logic between the unmanned vessel acting as a relay center and performing search and rescue missions. Furthermore, users can intervene in the GUI to manually change the scheduling of unmanned vehicles or force the activation of specific algorithms.

[0138] This embodiment provides an unmanned swarm search and rescue mission planning method. The method analyzes multiple distress signals in a search and rescue area to determine accident point information; dynamically updates the target detection probability of each grid sub-area in the search and rescue area based on the environmental characteristics and target movement characteristics of the search and rescue area; determines the current mission planning strategy based on the number of unmanned equipment types in the unmanned swarm and the accident point information; determines the regional search and rescue priority list for each unmanned equipment based on the current mission planning strategy and the target detection probability of the grid sub-area; and allocates tasks based on the regional search and rescue priority list, scenario adaptability, and execution cost of each unmanned equipment to determine the target search and rescue area, execution sequence, and search and rescue path for each unmanned equipment. The method updates the environmental characteristics and target movement characteristics in real time and combines planning strategies for different scenarios to uniformly plan and allocate tasks for multiple types of unmanned equipment. The method can comprehensively consider the speed, remaining energy, and environmental adaptability of each equipment, multiple accident points, different scenarios, and the dynamic changes of the environment and the target in distress, and calculate the optimal or near-optimal mission planning scheme in real time. This method achieves rapid coverage of multiple accident points and timely response to emergencies, improves search and rescue efficiency and accuracy in complex search and rescue scenarios, and increases the success rate of search and rescue missions.

[0139] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above embodiment 1 can be referred to the above introduction and will not be described in detail later. Figure 4 , step S10 may include steps S101 to S104:

[0140] Step S101, extracting spatiotemporal features from multi-source distress signals in the area to be searched and rescued;

[0141] It is understandable that spatiotemporal features are extracted simultaneously from multi-source distress signals (radar, radio, sonar, images, etc.).

[0142] Step S102: inputting the spatiotemporal characteristics into an accident point quantity recognition model for recognition to determine the number of accident points;

[0143] It should be noted that the accident point quantity recognition model is a model that can identify the number of accident points. This embodiment uses a trained deep learning model as the accident point quantity recognition model. The model is based on a convolutional neural network (CNN), and CNN can fuse these spatiotemporal features.

[0144] It is understandable that the accident point number recognition model includes convolutional layer, pooling layer and fully connected layer. Assume that the input data set is , after the convolution layer and the pooling layer, the feature vector is obtained , and then the classification prediction of the number of accident points is performed through the fully connected layer, and the output is a single accident point or multiple accident points, as shown below:

[0145]

[0146] Where, is the identification result of the number of accident points, is the activation function, and Respectively for category The weights and biases of is the index of the accident point quantity category, If it is determined to be a multiple accident point, all scattered coordinate information will be marked and the location of each accident point will be processed later.

[0147] Step S103, extracting scene features from the multi-source distress signals in the search and rescue area;

[0148] It should be noted that after determining the number of accident points, scene-related features, namely scene features, are obtained by extracting features from multi-source distress signals such as radar echoes, underwater cameras, and satellite images.

[0149] Step S104: input the scene features into the accident point scene classification model for classification to determine the accident point scene type.

[0150] It should be noted that the accident point scene classification model is a model that can classify and identify the accident point scene type. This embodiment uses a trained deep learning model as the accident point scene classification model, and the model combines a convolutional neural network with an attention mechanism.

[0151] It is understandable that the accident scene classification model includes multiple layers of convolutional layers, pooling layers, attention layers, and fully connected layers. Assume that the initial feature vector obtained after the operation of multiple layers of convolutional layers and pooling layers is , and then generate key feature vectors through the attention layer , and then classify it through the fully connected layer to output which of the three types of scenes the accident point belongs to: sea surface, shallow sea surface, and island, as shown below:

[0152]

[0153] Where, is the classification result of the accident scene type, is the activation function, and Respectively for category The weights and biases of is the index of the accident scene type, is the initial eigenvector, is the key feature vector. After completing the preliminary determination of the scene type, the accident point coordinates and scene type labels are combined to use a clustering algorithm such as DBSCAN or K-means to identify accident point clusters within the same area and calibrate their attributes. For example, if the center of a cluster falls within the boundary of an island, it is labeled IslandCluster; if underwater characteristic signals are identified, it is labeled ShallowUnderwaterCluster.

[0154] This embodiment provides an unmanned swarm search and rescue mission planning method. It extracts spatiotemporal features from multi-source distress signals in the search and rescue area; inputs these spatiotemporal features into an accident point number recognition model for identification and determination of the number of accident points; extracts scene features from the multi-source distress signals in the search and rescue area; and inputs these scene features into an accident point scene classification model for classification and determination of the scene type of the accident points. A deep learning model is used to automatically identify the number of accident points and classify the scene types of the accident points from the input data. This replaces manual, item-by-item input with an automatic determination mode, automatically generating a scene configuration. This provides accurate environmental perception and accident point distribution data for subsequent path planning and unmanned equipment scheduling, improving search and rescue efficiency and accuracy.

[0155] For example, in order to help understand the implementation process of the unmanned swarm search and rescue mission planning method obtained by combining this embodiment with the above-mentioned embodiment 2, please refer to Figure 5 , Figure 5 A brief flowchart of an unmanned swarm search and rescue mission planning method is provided. Specifically:

[0156] The first step is data input. The basic parameters of the drone, unmanned vessel, unmanned underwater vehicle, and robot dog (such as speed, endurance, and payload) are input, along with the grid size, convolution kernel, and target motion feature vector within the search and rescue area. The initial probability distribution of the accident point is pre-assigned. If there are multiple accident points, they can be mapped to multiple grid cells.

[0157] The second step is grid dynamics calculation. The convolution kernel coefficients are updated based on factors such as ocean currents, wind direction, and random drift. If the equipment search is successful, the probability of the corresponding grid decreases, and the process is iterated in a time-step or event-driven mode.

[0158] In the third step, multiple algorithms are used to prioritize the grids. Based on different scenarios (single / multiple accident points, homogeneous / heterogeneous equipment), algorithms 1, 2, or 3 are automatically selected to obtain a grid priority list or value index.

[0159] The fourth step is to generate dispatch tasks and allocate them to multiple devices. Based on the priority list, grids or grid clusters are mapped to task sets and task weights are assigned. Task allocation variables are solved or iteratively assigned, outputting the specific execution sequence and path for each device. If Algorithm 3 allows for competitive order grabbing, dynamic scheduling between devices can also be triggered during execution.

[0160] Step 5: Simulation and visualization. In multi-agent simulation modeling (e.g., using the AnyLogic platform), the probabilistic evolution of grid dynamics and the behavioral state diagrams of drones, unmanned vessels, underwater vehicles, and robotic dogs are implemented. The GUI displays real-time grid probability distribution, equipment positions and remaining battery life, algorithm allocation results, and competition processes, facilitating user interaction and solution evaluation.

[0161] It should be noted that the above examples are only used to understand this application and do not constitute a limitation on the unmanned cluster search and rescue mission planning method of this application. More simple transformations based on this technical concept are all within the scope of protection of this application.

[0162] This application also provides an unmanned swarm search and rescue mission planning system, please refer to Figure 6 , the unmanned swarm search and rescue mission planning system includes:

[0163] A scene recognition module 10 is used to analyze the multi-source distress signals in the search and rescue area and determine the accident point information;

[0164] A probability calculation module 20 is configured to dynamically update the target discovery probability of each grid sub-area in the search and rescue area based on the environmental characteristic information and target movement characteristic information of the search and rescue area;

[0165] A strategy selection module 30 is configured to determine a current mission planning strategy based on the number and types of unmanned equipment in the unmanned cluster and the accident point information;

[0166] A mission planning module 40 is configured to determine a regional search and rescue priority list for each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area;

[0167] The task planning module 40 is further used to allocate tasks based on the regional search and rescue priority list, scene adaptability and execution cost of each unmanned equipment, and determine the target search and rescue area, execution sequence and search and rescue path of each unmanned equipment.

[0168] In a feasible implementation manner, the scene recognition module 10 is further configured to extract spatiotemporal features from multi-source distress signals in the search and rescue area;

[0169] Inputting the spatiotemporal characteristics into an accident point quantity recognition model for recognition, and determining the number of accident points;

[0170] extracting scene features from multi-source distress signals in the area to be searched and rescued;

[0171] The scene features are input into the accident point scene classification model for classification to determine the accident point scene type.

[0172] In a feasible implementation manner, the strategy selection module 30 is further configured to determine that the current mission planning strategy is a mission planning strategy based on a competition mechanism when the number of types of the unmanned equipment is greater than or equal to a preset type number threshold;

[0173] When the number of types of the unmanned equipment is less than a preset type number threshold and the accident point information does not conform to a multiple accident point scenario, determining that the current task planning strategy is a task planning strategy based on probability sorting;

[0174] When the number of types of the unmanned equipment is less than a preset type number threshold and the accident point information meets the multi-accident point scenario, the current task planning strategy is determined to be a task planning strategy based on probability and speed sorting.

[0175] In a feasible embodiment, the mission planning module 40 is further configured to obtain a correspondence between the target discovery probability, the area distance, the speed of the unmanned equipment, and the search and rescue priority when the current mission planning strategy is a mission planning strategy based on probability and speed sorting;

[0176] Determining the search and rescue priority of the unassigned area corresponding to each unassigned area based on the target discovery probability of the unassigned area in the grid sub-area, the speed of the unmanned equipment, the area distance between the unmanned equipment and the unassigned area, and the corresponding relationship;

[0177] The search and rescue priorities of the unassigned areas corresponding to each unmanned equipment are sorted in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

[0178] In a feasible implementation manner, the mission planning module 40 is further configured to obtain a correspondence between the target discovery probability, the area distance, the speed of the unmanned equipment, and the search and rescue priority when the current mission planning strategy is a mission planning strategy based on a competition mechanism;

[0179] Determining the search and rescue priority of the grid sub-area corresponding to each unmanned device based on the target discovery probability of the grid sub-area, the speed of the unmanned device, the regional distance between the unmanned device and the grid sub-area, and the corresponding relationship;

[0180] The search and rescue priorities of the grid sub-areas corresponding to each unmanned equipment are sorted in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

[0181] In a feasible embodiment, the mission planning module 40 is further configured to, when the current mission planning strategy is a probability-based mission planning strategy, use the target discovery probability of the unassigned area in the grid sub-area as the search and rescue priority of the corresponding unassigned area;

[0182] The search and rescue priorities of the unassigned areas corresponding to each unmanned equipment are sorted in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

[0183] In a feasible embodiment, the task planning module 40 is further configured to map the grid sub-regions into task sets based on the regional search and rescue priority list, and determine the weight of each task in the task set;

[0184] Constructing a task allocation variable, and determining a comprehensive benefit function of the unmanned swarm based on the task allocation variable, task weights, scene adaptability, execution costs, and a balance coefficient;

[0185] Based on the comprehensive benefit function and the corresponding constraints, the optimal allocation strategy is solved to obtain the target search and rescue area, execution sequence and search and rescue path of each unmanned equipment.

[0186] The unmanned cluster search and rescue mission planning system provided by this application adopts the unmanned cluster search and rescue mission planning method in the above-mentioned embodiment, which can solve the technical problem that the search and rescue efficiency of the unmanned cluster mission planning scheme in complex search and rescue scenarios is poor, affecting the success rate of the search and rescue mission. Compared with the existing technology, the beneficial effects of the unmanned cluster search and rescue mission planning system provided by this application are the same as the beneficial effects of the unmanned cluster search and rescue mission planning method provided by the above-mentioned embodiment, and the other technical features of the unmanned cluster search and rescue mission planning system are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here.

[0187] The present application provides an unmanned cluster search and rescue mission planning device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein 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 the unmanned cluster search and rescue mission planning method in the above-mentioned embodiment one.

[0188] Reference below Figure 7 , which shows a schematic diagram of the structure of an unmanned swarm search and rescue mission planning device suitable for implementing an embodiment of the present application. The unmanned swarm search and rescue mission planning device in the embodiment of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 7 The unmanned cluster search and rescue mission planning device shown is merely an example and should not impose any limitations on the functions and scope of use of the embodiments of the present application.

[0189] like Figure 7As shown, the unmanned swarm search and rescue mission planning device may include a processing system 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in ROM (Read Only Memory) 1002 or programs loaded from storage system 1003 into RAM (Random Access Memory) 1004. RAM 1004 also stores various programs and data required for the operation of the unmanned swarm search and rescue mission planning device. Processing system 1001, ROM 1002, and RAM 1004 are interconnected via bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems may be connected to I / O interface 1006: an input system 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; an output system 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; a storage system 1003 including, for example, a magnetic tape, hard disk, etc.; and a communication system 1009. Communication system 1009 may allow the unmanned swarm search and rescue mission planning device to communicate wirelessly or wired with other devices to exchange data. While the figure illustrates an unmanned swarm search and rescue mission planning device with various systems, it should be understood that implementation or presence of all illustrated systems is not required. More or fewer systems may alternatively be implemented or present.

[0190] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication system, or installed from a storage system 1003, or installed from a ROM 1002. When the computer program is executed by the processing system 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0191] The unmanned cluster search and rescue mission planning device provided by this application adopts the unmanned cluster search and rescue mission planning method in the above-mentioned embodiment, which can solve the technical problem that the search and rescue efficiency of the unmanned cluster mission planning scheme in complex search and rescue scenarios is poor, affecting the success rate of the search and rescue mission. Compared with the existing technology, the beneficial effects of the unmanned cluster search and rescue mission planning device provided by this application are the same as the beneficial effects of the unmanned cluster search and rescue mission planning method provided by the above-mentioned embodiment, and the other technical features of the unmanned cluster search and rescue mission planning device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0192] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0193] The above are only specific embodiments of the present application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0194] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, a computer program) stored thereon, and the computer-readable program instructions are used to execute the unmanned cluster search and rescue mission planning method in the above-mentioned embodiment.

[0195] The computer-readable storage medium provided herein may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems, or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including, but not limited to, wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0196] The above-mentioned computer-readable storage medium may be included in the unmanned swarm search and rescue mission planning device; or it may exist independently without being assembled into the unmanned swarm search and rescue mission planning device.

[0197] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the unmanned cluster search and rescue mission planning device, the unmanned cluster search and rescue mission planning device: analyzes the multi-source distress signals in the search and rescue area to determine the accident point information; dynamically updates the target discovery probability of each grid sub-area in the search and rescue area based on the environmental characteristic information and target movement characteristic information of the search and rescue area; determines the current task planning strategy based on the type and number of unmanned equipment in the unmanned cluster and the accident point information; determines the regional search and rescue priority list of each unmanned equipment based on the current task planning strategy and the target discovery probability of the grid sub-area; performs task allocation based on the regional search and rescue priority list, scene adaptability and execution cost of each unmanned equipment, and determines the target search and rescue area, execution sequence and search and rescue path of each unmanned equipment.

[0198] Computer program code for performing the operations of the present application may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0199] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0200] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0201] The computer-readable storage medium provided in this application stores computer-readable program instructions (i.e., a computer program) for executing the unmanned swarm search and rescue mission planning method described above. This computer-readable storage medium can address the technical issue of poor search and rescue efficiency in unmanned swarm mission planning schemes in complex search and rescue scenarios, which impacts the success rate of search and rescue missions. Compared to the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the unmanned swarm search and rescue mission planning method provided in the aforementioned embodiments, and are not further elaborated here.

[0202] The present application also provides a computer program product, including a computer program, which implements the steps of the unmanned cluster search and rescue mission planning method as described above when the computer program is executed by a processor.

[0203] The computer program product provided in this application can address the technical issue of poor search and rescue efficiency in unmanned swarm mission planning schemes in complex search and rescue scenarios, which impacts the success rate of search and rescue missions. Compared to the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the unmanned swarm search and rescue mission planning method provided in the aforementioned embodiments, and are not further elaborated here.

[0204] The above are only some embodiments of the present application and are not intended to limit the patent scope of the present application. All equivalent structural transformations made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for planning an unmanned swarm search and rescue mission, characterized in that: The method comprises: Analyze multiple distress signals in the search and rescue area to determine the accident point information; Dynamically updating the target discovery probability of each grid sub-area in the search and rescue area based on the environmental feature information and target movement feature information of the search and rescue area; Determining a current mission planning strategy based on the number of types of unmanned equipment in the unmanned swarm and the accident point information, wherein when the number of types of unmanned equipment is greater than or equal to a preset type number threshold, adopting a mission planning strategy based on a competition mechanism as the current mission planning strategy, the preset type number threshold is 2, and the unmanned equipment includes at least two of a drone, an unmanned ship, a robot dog, and an unmanned underwater vehicle; Determining a regional search and rescue priority list for each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area; Tasks are allocated based on the regional search and rescue priority list, scenario adaptability, and execution cost of each unmanned equipment to determine the target search and rescue area, execution sequence, and search and rescue path of each unmanned equipment. When a task planning strategy based on a competition mechanism is adopted, all unmanned equipment participates in task allocation, so that unmanned equipment that is heading to the target search and rescue area or searching the target search and rescue area is adjusted in real time to the grid sub-area with the highest search and rescue priority in the regional search and rescue priority list; The accident point information includes at least the number of accident points and the scene type of the accident points. The step of analyzing the multi-source distress signals in the search and rescue area to determine the accident point information includes: Extracting spatiotemporal features from multi-source distress signals in the area to be searched and rescued; Inputting the spatiotemporal characteristics into an accident point quantity recognition model for recognition, and determining the number of accident points; extracting scene features from multi-source distress signals in the area to be searched and rescued; Inputting the scene features into the accident point scene classification model for classification to determine the accident point scene type; The dynamic updating of the target discovery probability of each grid sub-area in the search and rescue area specifically includes: The calculation formula of the target discovery probability of the grid sub-area at the moment is determined The target discovery probability of each grid sub-area at the moment, The calculation formula for the target discovery probability of the grid sub-area at a certain moment is: Indicates The time coordinate is The target discovery probability of the grid sub-area is Indicates The time coordinate is The target discovery probability of the adjacent grid sub-area is, Represents the convolution kernel coefficient, reflecting the distress target from the adjacent grid sub-region Drift to grid sub-region The probability of the distressed target moving to the surrounding area is calculated as follows: The vector that causes the distress target to move fixedly due to external environmental factors. The vector that randomly moves the target in distress, vector At a radius of Evenly distributed within the circle, Indicates the side length of the grid sub-region.

2. The method according to claim 1, wherein The step of determining the current mission planning strategy based on the types and number of unmanned equipment in the unmanned cluster and the accident point information includes: When the number of types of the unmanned equipment is greater than or equal to a preset type number threshold, determining that the current mission planning strategy is a mission planning strategy based on a competition mechanism; When the number of types of the unmanned equipment is less than a preset type number threshold and the accident point information does not conform to a multiple accident point scenario, determining that the current task planning strategy is a task planning strategy based on probability sorting; When the number of types of the unmanned equipment is less than a preset type number threshold and the accident point information meets the multi-accident point scenario, the current task planning strategy is determined to be a task planning strategy based on probability and speed sorting.

3. The method according to claim 1, wherein The step of determining the regional search and rescue priority list of each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area includes: When the current mission planning strategy is a mission planning strategy based on probability and speed sorting, obtaining a correspondence between target discovery probability, area distance, speed of the unmanned equipment, and search and rescue priority; Determining the search and rescue priority of the unassigned area corresponding to each unassigned area based on the target discovery probability of the unassigned area in the grid sub-area, the speed of the unmanned equipment, the area distance between the unmanned equipment and the unassigned area, and the corresponding relationship; The search and rescue priorities of the unassigned areas corresponding to each unmanned equipment are sorted in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

4. The method according to claim 1, wherein The step of determining the regional search and rescue priority list of each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area includes: When the current mission planning strategy is a mission planning strategy based on a competition mechanism, obtaining a corresponding relationship between the target discovery probability, the area distance, the speed of the unmanned equipment, and the search and rescue priority; Determining the search and rescue priority of the grid sub-area corresponding to each unmanned device based on the target discovery probability of the grid sub-area, the speed of the unmanned device, the regional distance between the unmanned device and the grid sub-area, and the corresponding relationship; The search and rescue priorities of the grid sub-areas corresponding to each unmanned equipment are sorted in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

5. The method according to claim 1, wherein The step of determining the regional search and rescue priority list of each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area includes: When the current mission planning strategy is a mission planning strategy based on probability sorting, the target discovery probability of the unassigned area in the grid sub-area is used as the search and rescue priority of the corresponding unassigned area; The search and rescue priorities of the unassigned areas corresponding to each unmanned equipment are sorted in descending order to obtain a regional search and rescue priority list for each unmanned equipment.

6. The method according to any one of claims 1 to 5, characterized in that The steps of allocating tasks based on the regional search and rescue priority list, scene adaptability, and execution cost of each unmanned equipment, and determining the target search and rescue area, execution sequence, and search and rescue path of each unmanned equipment include: Based on the regional search and rescue priority list, mapping the grid sub-region into a task set and determining the weight of each task in the task set; Constructing a task allocation variable, and determining a comprehensive benefit function of the unmanned swarm based on the task allocation variable, task weights, scene adaptability, execution costs, and a balance coefficient; Based on the comprehensive benefit function and the corresponding constraints, the optimal allocation strategy is solved to obtain the target search and rescue area, execution sequence and search and rescue path of each unmanned equipment.

7. An unmanned swarm search and rescue mission planning system, characterized in that: The system comprises: The scene recognition module is used to analyze the multi-source distress signals in the search and rescue area and determine the accident point information; A probability calculation module, configured to dynamically update the target discovery probability of each grid sub-area in the search and rescue area based on the environmental characteristic information and target movement characteristic information of the search and rescue area; a strategy selection module, configured to determine a current mission planning strategy based on the number of types of unmanned equipment in the unmanned cluster and the accident point information, wherein when the number of types of unmanned equipment is greater than or equal to a preset type number threshold, a mission planning strategy based on a competition mechanism is used as the current mission planning strategy, the preset type number threshold is 2, and the unmanned equipment includes at least two of a drone, an unmanned ship, a robot dog, and an unmanned underwater vehicle; A mission planning module, configured to determine a regional search and rescue priority list for each unmanned equipment based on the current mission planning strategy and the target discovery probability of the grid sub-area; The task planning module is further used to allocate tasks based on the regional search and rescue priority list, scenario adaptability, and execution cost of each unmanned equipment, and determine the target search and rescue area, execution sequence, and search and rescue path of each unmanned equipment. When a task planning strategy based on a competition mechanism is adopted, all unmanned equipment participates in task allocation, so that unmanned equipment that is heading to the target search and rescue area or searching the target search and rescue area is adjusted in real time to the grid sub-area with the highest search and rescue priority in the regional search and rescue priority list; The accident point information includes at least the number of accident points and the scene type of the accident points. The scene recognition module is further used to extract spatiotemporal features from the multi-source distress signals in the search and rescue area; Inputting the spatiotemporal characteristics into an accident point quantity identification model for identification, and determining the number of accident points; extracting scene features from multi-source distress signals in the area to be searched and rescued; Inputting the scene features into the accident point scene classification model for classification to determine the accident point scene type; The probability calculation module is also used based on The calculation formula of the target discovery probability of the grid sub-area at the moment is determined The target discovery probability of each grid sub-area at the moment, The calculation formula for the target discovery probability of the grid sub-area at a certain moment is: Indicates The time coordinate is The target discovery probability of the grid sub-area is Indicates The time coordinate is The target discovery probability of the adjacent grid sub-area is, Represents the convolution kernel coefficient, reflecting the distress target from the adjacent grid sub-region Drift to grid sub-region The probability of the distressed target moving to the surrounding area is calculated as follows: The vector that causes the distress target to move fixedly due to external environmental factors. The vector that randomly moves the target in distress, vector At a radius of Evenly distributed within the circle, Indicates the side length of the grid sub-region.

8. An unmanned swarm search and rescue mission planning device, characterized in that: The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the unmanned swarm search and rescue mission planning method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the unmanned cluster search and rescue mission planning method according to any one of claims 1 to 6 are implemented.

Images