Multi-unmanned aerial vehicle cooperation method and system for electromagnetic spectrum detection

Abstract

The invention provides a multi-unmanned aerial vehicle cooperation method and system for electromagnetic spectrum detection, and belongs to the field of electromagnetic spectrum detection and multi-unmanned aerial vehicle cooperation. According to the method, firstly, collaborative detection is carried out on an overall task area to obtain a sparse electromagnetic spectrum map, then a global dense map is obtained through interpolation, and then clustering is carried out on the map to obtain a key task area. And finally, through a hybrid algorithm of a differential evolution algorithm and a wolf pack algorithm, completing a detection task on a key area and updating and correcting the previous electromagnetic map to obtain a final electromagnetic spectrum map. Meanwhile, on the collaborative method, the thought of the differential evolution algorithm is applied to the wolf pack algorithm, the problem that the wolf pack algorithm is prone to falling into local optimum is solved, the global search capacity is improved, and the search precision is improved.

Description

Technical Field

[0001] The present invention relates to the field of electromagnetic spectrum detection and multi-UAV cooperation, and specifically to a multi-UAV cooperation method and system for electromagnetic spectrum detection. Background Art

[0002] Electromagnetic spectrum detection is a visualization method that combines the geographical space and the electromagnetic spectrum space, and is a means to display the electromagnetic spectrum intensity and distribution in a specific area in detail. However, in the process of spectrum detection, the UAV swarm often faces a complex environment, and there are often problems such as difficult task allocation and path planning, and low accuracy. Traditional task allocation and cooperation methods are mainly divided into two categories: one is traditional mathematical programming methods, mainly including dynamic programming method, mixed integer linear programming method, Hungarian algorithm, etc.; the other is intelligent optimization algorithms, mainly including tabu search algorithm, simulated annealing algorithm, particle swarm algorithm, wolf pack algorithm, etc. However, each algorithm has its own deficiencies and needs to be improved. Summary of the Invention

[0003] In view of the problems of insufficient accuracy and low efficiency of the existing electromagnetic spectrum detection methods, the present invention proposes a multi-UAV cooperation method and system for electromagnetic spectrum detection.

[0004] A multi-UAV cooperation method for electromagnetic spectrum detection includes:

[0005] Step 1: Determine the task area that needs to be detected for the electromagnetic spectrum; determine the number of UAVs; determine the state of the UAVs;

[0006] Step 2: Divide the task area into multiple sub-areas, and dispatch an unmanned cluster to each sub-area;

[0007] Step 3: Plan the search path of each unmanned cluster in its respective sub-area;

[0008] Step 4: According to the search path, the unmanned cluster detects the electromagnetic spectrum of its respective sub-area, and summarizes and stitches the electromagnetic spectrum conditions of each obtained sub-area to obtain a sparse electromagnetic map of the entire task area;

[0009] Step 5: Interpolate the obtained sparse electromagnetic map to obtain a global dense electromagnetic map;

[0010] Step 6: Determine the core area that needs to be investigated key points according to the global dense electromagnetic map;

[0011] Step 7: Based on the core area to be investigated key points, re-divide the sub-areas of the task, and at the same time re-form the UAVs according to the situation of the re-divided sub-areas;

[0012] Step 8: According to the reorganized unmanned cluster, re-detect the electromagnetic spectrum of their respective sub-regions and correct the global dense electromagnetic map.

[0013] Preferably, in the said Step 8, the method of re-detecting the electromagnetic spectrum of their respective sub-regions includes:

[0014] Step 8.1: Initialize the artificial wolf pack, where the artificial wolf pack corresponds to the unmanned cluster, and each artificial wolf in the artificial wolf pack corresponds to each unmanned aerial vehicle in the unmanned cluster;

[0015] Step 8.2: Roaming behavior, including:

[0016] According to the intensity of the electromagnetic spectrum at the positions of the initialized unmanned aerial vehicles, select the unmanned aerial vehicle at the position with the strongest electromagnetic spectrum intensity as the lead wolf, and select multiple unmanned aerial vehicles with the second-strongest electromagnetic spectrum intensity as scout wolves in the search state, and the remaining ones as fierce wolves; the scout wolves execute the roaming behavior, that is, start to detect the intensity of the electromagnetic spectrum. If the intensity of the electromagnetic spectrum detected by the scout wolf is weaker than the previous position, it retreats. If it is stronger than the previous position, it continues to explore around with a set step size. And if the prey smell of the scout wolf is greater than that of the lead wolf at this time, this scout wolf is called the new lead wolf, and the roles of the artificial wolves are reallocated; this step is iterated continuously until the roaming behavior ends.

[0017] Step 8.3: Hunting behavior, including:

[0018] When the average prey smell detected by all scout wolves is greater than the set upper bound of prey smell Y H or the number of roaming steps is greater than the set maximum number of roaming steps max kr , then the lead wolf howls, and the fierce wolves approach the lead wolf for hunting behavior, that is, detect the electromagnetic spectrum situation near the lead wolf, and update the original electromagnetic map according to the detection situation.

[0019] Step 8.4: Mutation operation, including:

[0020] Randomly generate three distinct integers r 1 , r 2 , r 3 , and it is required that i, r 1 , r 2 and r 3 these four numbers are distinct from each other, and then calculate where N is the scale of the artificial wolf pack;

[0021] where F is the mutation scale, and respectively represent the ones numbered r 1 , r 2 , r 3The electromagnetic spectrum intensity of an individual; Let v i As individual X i The detected electromagnetic spectrum intensity, thus completing the mutation of individual X i At the same time, find an individual X in the wolf pack whose electromagnetic spectrum intensity is closest to v i ; j ;

[0022] Step 8.5, Crossover operation, including:

[0023] Set intermediate variables: Where CR is the crossover probability, and rand(0, 1) represents a random number between 0 and 1;

[0024] Step 8.6, Selection operation, including:

[0025] Based on the intermediate variables, use the following relational expression to select the mutated individual and the original individual to obtain the individual that is more adaptable to iteration between the two: Where, E j Represents the electromagnetic spectrum intensity at individual X j ;

[0026] Step 8.7, Evaluate the advantages and disadvantages of each UAV, eliminate some artificial wolves according to the strategy of survival of the fittest, and select a new leader wolf. After the elimination is completed, iterate according to steps 8.2 to 8.6;

[0027] Step 8.8, When the iteration stop condition is reached, end the iteration to complete the search of the area and reconstruct the electromagnetic map.

[0028] Preferably, the method for planning the search path of each unmanned cluster in its respective sub-region includes:

[0029] Divide the sub-region into grids, and use η to represent the uncertainty at each grid; among them, the uncertainty is a value between 0 and 1. For a grid that has not been detected at all, the uncertainty is 1, and as the number of detections increases, the uncertainty η decreases accordingly;

[0030] For a certain grid, define the search value function as J = ω 1 γJ 1 +ω 2 γJ 2 +ω 3 J 3 , where, J 1 , J 2 , J 3 Represents the value standard of the route selection point; ω 1 , ω 2 And ω 3are the weights of the corresponding paths, and the sum of the three is 1; γ is the importance factor, and the importance of the path trajectory selection point is reflected by adjusting γ, and the value criterion therein is determined by uncertainty and whether it is a no-fly zone. Specifically:

[0031] For the value criterion J 1 , it is determined by the uncertainty of the grid. The specific relationship is J 1 = 20lnαη, where α is the adjustment importance parameter;

[0032] For the value criterion J 2 , it is determined by the increased coverage rate that can be obtained by selecting this grid. The specific relationship is J 2 = ln△S, where △S is the increased detection area;

[0033] For the value criterion J 3 , if it is a no-fly zone, the value criterion J 3 takes -65535, and if it is not a no-fly zone, the value criterion J 3 takes 1;

[0034] Maximize the search value function of the path as the constraint of path search to obtain the search path.

[0035] Preferably, the method for determining the core area that needs to be key-probed includes:

[0036] Set a distance threshold esp and a minimum sample point threshold min_sample; for any grid detected, within a circular area with a radius of the distance threshold esp, determine the number of sample points where the electromagnetic spectrum intensity is greater than the lowest electromagnetic spectrum threshold. If the number of sample points is greater than the minimum sample point threshold min_sample, then this circular area is the core area.

[0037] Preferably, the method for redefining the sub-regions of the task includes:

[0038] Each connected core area is used as a sub-region, and this sub-region is regarded as the key-probing area; the connected non-core areas are regarded as one sub-region.

[0039] Preferably, the method for re-forming the UAV formation according to the redefined sub-region situation includes:

[0040] The number of UAVs allocated to each sub-region is where α max is the maximum electromagnetic intensity of all regions, α h is the maximum electromagnetic intensity of the h-th sub-region, N s represents the total number of UAVs, S sum represents the area of all sub-regions, S hrepresents the area of the h-th sub-region; m and k are adjustable parameters, γ is determined by whether it is a core region, taking 1 if it is and 0 if not, and satisfies where N is the total number of divided sub-regions.

[0041] Preferably, the electromagnetic map of each iteration is retained. When the electromagnetic spectrum intensity detected each time is different, the arithmetic mean method is used to update the electromagnetic spectrum intensity.

[0042] A multi-UAV cooperative system for electromagnetic spectrum detection includes a first module, a second module, a third module, a fourth module, a fifth module, a sixth module, a seventh module, and an eighth module;

[0043] Among them, the first module is used to determine the task area that needs to conduct electromagnetic spectrum detection; determine the number of UAVs; determine the status of UAVs;

[0044] The second module divides the task area into multiple sub-regions and dispatches unmanned clusters to each sub-region;

[0045] The third module plans the search paths of each unmanned cluster within its respective sub-region;

[0046] The fourth module detects the electromagnetic spectrum of each sub-region where the unmanned cluster is located, summarizes and stitches the electromagnetic spectrum conditions of each obtained sub-region to obtain a sparse electromagnetic map of the entire task area;

[0047] The fifth module interpolates the obtained sparse electromagnetic map to obtain a globally dense electromagnetic map;

[0048] The sixth module determines the core regions that need to be key explored according to the globally dense electromagnetic map;

[0049] The seventh module re-divides the sub-regions of the task based on the core regions that need to be key explored, and at the same time re-forms the UAVs according to the situation of the re-divided sub-regions;

[0050] The eighth module re-detects the electromagnetic spectrum of its respective sub-region according to the re-formed unmanned cluster and corrects the globally dense electromagnetic map.

[0051] The present invention has the following beneficial effects:

[0052] A multi-UAV cooperation method and system for electromagnetic spectrum detection provided by the present invention first conducts cooperative detection on the overall task area to obtain a sparse electromagnetic spectrum map, then obtains a global dense map through interpolation, and then clusters the map to obtain key task areas. Finally, a hybrid algorithm of differential evolution algorithm and wolf pack algorithm is used to complete the detection task of key areas and update and correct the previous electromagnetic map to obtain the final electromagnetic spectrum map. At the same time, in the cooperation method, the idea of differential evolution algorithm is applied to the wolf pack algorithm, which solves the problem that the wolf pack algorithm is easy to fall into local optimum, improves the global search ability, and improves the search accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is the overall idea framework diagram of a multi-UAV cooperation method for electromagnetic spectrum detection of the present invention.

[0054] Figure 2 It is the flow chart of the hybrid algorithm of differential evolution algorithm and wolf pack algorithm adopted by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0055] The following takes embodiments in conjunction with the drawings and describes the present invention in detail.

[0056] The present invention provides a multi-UAV cooperation method for electromagnetic spectrum detection, which specifically includes the following steps:

[0057] Step 1: Define the task area, the number of UAVs, and the equipment status of the UAVs, and record the above information in the decision-making system. The equipment status includes: the maximum mileage r that the UAV can navigate f , the maximum communication radius r of the UAV c , the maximum radius r of the electromagnetic spectrum that the UAV can detect d , the maximum turning angle θ that can be physically controlled, and the initial position of the i-th UAV in the d-dimensional space.

[0058] Step 2: Divide the task area into several sub-areas, and dispatch UAV clusters to each sub-interval. The sub-areas are regarded as a finite number of square grids with side length a, and the no-fly zones in the area are abstracted into convex obstacles and concave obstacles. In addition, η is used to represent the uncertainty at each grid; among them, the uncertainty is a value between 0 and 1. For a grid that has not been detected at all, the uncertainty is 1. As the number of detections increases, the uncertainty η(x, y) will decrease accordingly.

[0059] Step 3: Based on the search value function, plan the search path for each UAV cluster. For a certain grid, the search value function is defined as J = ω 1 γJ 1 + ω 2 γJ2 +ω 3 J 3 , in the above formula, J 1 、J 2 、J 3 represent the value criteria of waypoint selection; ω 1 、ω 2 and ω 3 are the weights of the corresponding paths respectively, and the sum of the three is 1; the importance of the value criteria is achieved by adjusting the weights; γ is the importance factor, and the importance of the path trajectory selection point is reflected by adjusting γ, and the value criteria therein are determined by uncertainty and whether it is a no-fly zone. Specifically:

[0060] For the value criterion J 1 , it is determined by the uncertainty of the grid. The specific relationship is J 1 = 20lnαη, where α is the adjustment importance parameter;

[0061] For the value criterion J 2 , it is determined by the increased coverage rate that can be obtained by selecting this grid. The specific relationship is J 2 = ln△S, where △S is the increased area;

[0062] For the value criterion J 3 , if it is a no-fly zone, the value criterion J 3 takes -65535, and if it is not a no-fly zone, the value criterion J 3 takes 1.

[0063] Maximize the search value function of the path as the constraint of path search to obtain the search path. The specific method for path planning based on the search value function is that when selecting the i-th path, the total search value of the UAV is V i sum = ∑J ij , that is, the sum of the search value functions of the selected grids when the UAV selects this path. Finally, the total search value function of the planned path satisfies V = max{V i sum}.

[0064] Step 4: The UAV swarm detects the electromagnetic spectrum of its respective sub-region, and summarizes and stitches the electromagnetic spectrum situations obtained by each UAV cluster to obtain a sparse electromagnetic map of the entire mission area;

[0065] Step 5: Use Kriging interpolation method to interpolate the sparse electromagnetic map to obtain a globally dense electromagnetic map. The calculation method for interpolating the electromagnetic map is In the formula, P(m 0 ) is the interpolation result of the point to be interpolated; λ i is each surrounding point mi is the weight coefficient, and it is related to the semivariogram; N is the total number of sampling points adjacent to m 0 surrounding. The semivariogram is 0.5 times the variance of the difference between the attribute variable P(m) corresponding to the sample points at spatial positions m i and m i +h.

[0066] Step 6: For the globally dense electromagnetic map, based on the DBSCN algorithm, obtain the core areas that need to be key explored. The DBSCN algorithm is a density-based clustering algorithm, and the parameters in the clustering process are the set distance threshold esp and the minimum number of sample points threshold min_sample; for any grid to be detected, within the circular area with a radius of the distance threshold esp, determine the number m of sample points whose electromagnetic spectrum intensity is greater than the lowest electromagnetic spectrum threshold. If m is greater than the minimum number of sample points threshold min_sample, then this circular area is the core area and this grid is the core point; when two core areas are adjacent to each other, it is considered that these two areas are connected and can be regarded as one core area.

[0067] Step 7: Redefine the sub-regions of the task, and the UAV formation re-detects the sub-regions: Each connected core area is used as a sub-region, and this sub-region is regarded as the key exploration area; the connected non-core areas are regarded as one sub-region.

[0068] Determine the distances of each UAV to each key exploration area at this time; the number of UAVs allocated to each sub-region is where α max is the maximum electromagnetic intensity of all regions, α h is the maximum electromagnetic intensity of the h-th sub-region, N s represents the total number of UAVs, S sum represents the areas of all sub-regions, S h represents the area of the h-th sub-region; m and k are adjustable parameters, γ is determined by whether it is a core area, taking 1 if it is and 0 if it is not, and satisfies where N is the total number of divided sub-regions.

[0069] Step 8: Re-detect the electromagnetic spectrum based on the hybrid algorithm of the differential evolution algorithm and the wolf pack algorithm, and correct the globally dense electromagnetic map, which specifically includes the following steps:

[0070] Step 8.1: Initialize the artificial wolf pack, where the artificial wolf pack corresponds to the UAV cluster, and each artificial wolf in the artificial wolf pack corresponds to each UAV in the UAV cluster. Initialize the following parameters, including: the number n of individuals in the artificial wolf pack, the maximum number of iterations max k , the number q of artificial wolves in the search state, the search area h, the maximum number of wandering steps maxkr , the walking step length step a , the hunting step length step b and the position X of the i-th artificial wolf i , the upper bound Y of the prey smell during roaming H , the upper bound Y of the prey smell at the end HE ; where the prey smell represents the intensity of the electromagnetic spectrum; and the Euclidean distance d between each UAV should satisfy d min < d < min{r c , r d}, d min is the minimum allowable distance between UAVs.

[0071] Step 8.2, Roaming behavior. According to the intensity of the electromagnetic spectrum at the grid where each UAV is located in the initial position, select the UAV at the position with the strongest electromagnetic spectrum intensity as the alpha wolf, and select another q UAVs with the second-strongest electromagnetic spectrum intensity as the exploring wolves in the search state, and the remaining ones as the fierce wolves. The exploring wolves perform the roaming behavior, that is, start to detect the intensity of the electromagnetic spectrum. The step length for an exploring wolf to move in a certain direction is step a , if the intensity of the electromagnetic spectrum detected by the exploring wolf is weaker than the previous position, it retreats, and if it is stronger than the previous position, it continues to explore around with the step length step a , and if the prey smell of this exploring wolf is greater than that of the alpha wolf at this time, this exploring wolf is called the new alpha wolf and the roles are reallocated. This step is iterated continuously until the roaming behavior ends.

[0072] Step 8.3, Hunting behavior. When the average prey smell detected by all exploring wolves is greater than the upper bound Y of the prey smell H or the number of walking steps is greater than the maximum walking step max kr , then the alpha wolf howls, and the fierce wolves approach the alpha wolf with the step length step b to perform the hunting behavior, that is, detect the electromagnetic spectrum situation near the alpha wolf, update the original electromagnetic spectrum map according to the detection situation, and save the electromagnetic map of this round of iteration.

[0073] Step 8.4, Mutation operation. Perform differential mutation on the wolf pack. For each individual X i in the population, the intensity E of the detected electromagnetic spectrum i , randomly generate three distinct integers r 1 , r 2 , r 3 ∈{1, 2, …, N}, and it is required that the four numbers i, r 1 , r 2 and r 3 are distinct from each other, and then calculate

[0074] Where F is the mutation scale, and respectively represent the electromagnetic spectrum intensity of the individuals numbered r 1 , r 2 , r 3 ; Taking v i as the electromagnetic spectrum intensity detected by the individual X i , thus completing the mutation of the individual X i ; At the same time, find an individual X i in the wolf pack with the electromagnetic spectrum intensity closest to v j ;

[0075] Step 8.5, Crossover operation. In this step, the overall crossover of the mutation vector and the target vector is performed. The formula for completing the crossover operation is where CR is the crossover probability, and rand(0, 1) represents a random number between 0 and 1.

[0076] Step 8.6, Selection operation. The following relationship is used to select between the mutated individual and the original individual to obtain the individual that is more adaptable to iteration: where, E j represents the electromagnetic spectrum intensity at the individual X j ;

[0077] Step 8.7, Evaluate the advantages and disadvantages of each UAV. Eliminate some artificial wolves according to the strategy of survival of the fittest, and select a new lead wolf. The evaluation criteria mainly include the maximum mileage r f that the UAV can navigate, and the intensity of the electromagnetic spectrum that can be detected at the current position. After the elimination, iterate according to steps 8.2 to 8.6.

[0078] Step 8.8, When the prey smell is greater than the upper bound Y HE of the prey smell or the number of iterations is greater than the maximum number of iterations max k , end the iteration and complete the search of the area. Reconstruct the entire electromagnetic map. If there are several different iteration situations, use the arithmetic mean method to update the electromagnetic intensity.

[0079] Based on the above multi-UAV cooperation method for electromagnetic spectrum detection, the present invention also provides a multi-UAV cooperation system for electromagnetic spectrum detection, including a first module, a second module, a third module, a fourth module, a fifth module, a sixth module, a seventh module, and an eighth module.

[0080] Among them, the first module is used to determine the task area that needs to perform electromagnetic spectrum detection; determine the number of UAVs; determine the state of the UAVs;

[0081] The second module divides the mission area into a plurality of sub-areas, and dispatches a drone cluster to each sub-area;

[0082] The third module plans the search path of each drone cluster in its respective sub-area;

[0083] The unmanned cluster of the fourth module detects the electromagnetic spectrum of each sub-area, summarizes and splices the electromagnetic spectrum conditions of each sub-area, and obtains a sparse electromagnetic map of the entire mission area;

[0084] The fifth module interpolates the obtained sparse electromagnetic map to obtain a global dense electromagnetic map;

[0085] The sixth module determines the core area that needs to be explored based on the global dense electromagnetic map;

[0086] The seventh module re-defines the sub-areas of the mission based on the core area of ​​key exploration, and re-forms the drones according to the re-demarcated sub-areas;

[0087] The eighth module re-detects the electromagnetic spectrum of each sub-area according to the re-formed unmanned cluster and corrects the global dense electromagnetic map.

[0088] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-UAV collaboration method for electromagnetic spectrum detection, characterized in that: include: Step 1: Determine the mission area where electromagnetic spectrum detection is required; determine the number of drones; determine the status of the drones; Step 2: dividing the mission area into multiple sub-areas, and dispatching an unmanned cluster to each sub-area; Step 3: Plan the search path for each unmanned cluster in its respective sub-area; Step 4: According to the search path, the unmanned cluster detects the electromagnetic spectrum of each sub-area, summarizes and splices the electromagnetic spectrum of each sub-area, and obtains a sparse electromagnetic map of the entire mission area; Step 5: interpolate the obtained sparse electromagnetic map to obtain a global dense electromagnetic map; Step 6: Determine the core area that needs to be explored based on the global dense electromagnetic map; Step 7: Based on the core area of ​​key exploration, redefine the sub-areas of the mission, and re-form the drones according to the re-delineated sub-areas; Step 8: Based on the reorganized unmanned swarm, re-detect the electromagnetic spectrum of each sub-area and correct the global dense electromagnetic map.

2. The multi-UAV collaboration method for electromagnetic spectrum detection according to claim 1 is characterized in that: In step 8, the method of re-detecting the electromagnetic spectrum of each sub-area includes: Step 8.1, initialize the artificial wolf group, wherein the artificial wolf group corresponds to the unmanned cluster, and each artificial wolf in the artificial wolf group corresponds to each drone in the unmanned cluster; Step 8.2: Wandering behavior, including: According to the intensity of the electromagnetic spectrum at the locations of the initialized drones, the drone at the location with the strongest electromagnetic spectrum intensity is selected as the alpha wolf, and multiple drones with the second strongest electromagnetic spectrum intensity are selected as scout wolves in the search state, and the rest are fierce wolves; the scout wolves perform wandering behavior, that is, start to detect the intensity of the electromagnetic spectrum. If the electromagnetic spectrum intensity detected by the scout wolf is weaker than the previous step position, it will retreat. If it is stronger than the previous step position, it will continue to explore the surroundings with a set step length. If the smell of the prey detected by the scout wolf is stronger than that of the alpha wolf at this time, this scout wolf is called the new alpha wolf, and the artificial wolf role is reallocated; this step is iterated until the wandering behavior ends; Step 8.3: Hunting behavior, including: When the average prey odor detected by all wolves is greater than the set upper limit Y of prey odor H Or the number of steps is greater than the maximum number of steps max kr , the leader wolf howls, and the fierce wolves approach the leader wolf to hunt, that is, the electromagnetic spectrum situation near the leader wolf is detected, and the original electromagnetic map is updated according to the detection situation; Step 8.4: mutation operation, including: Randomly generate three different integers r1, r2, and r3 in the range of 1 to N, and require that i, r1, r2, and r3 are different from each other, and then calculate Where N is the size of the artificial wolf pack; Where F is the variation scale, and Respectively represent the electromagnetic spectrum intensity of individuals numbered r1, r2, and r3; v i As an individual X i The electromagnetic spectrum intensity detected, thus completing the individual X i Variation; at the same time, an electromagnetic spectrum intensity with v was found in the wolf pack i The closest individual X j ; Step 8.5: Crossover operation, including: Set up intermediate variables: Where CR is the crossover probability, rand(0,1) represents a random number between 0 and 1; Step 8.6, select an operation, including: Based on the intermediate variables, the following relationship is used to select the mutant individual and the original individual to obtain the individual that is more suitable for iteration: Among them, E j Represents individual X j The intensity of the electromagnetic spectrum at Step 8.7, evaluate the quality of each drone, eliminate some artificial wolves according to the strategy of survival of the fittest, and select a new leader wolf. After the elimination is completed, iterate according to steps 8.2 to 8.6; Step 8.8: When the iteration stop condition is reached, the iteration is terminated to complete the search of the area and the electromagnetic map is reconstructed.

3. The multi-UAV collaboration method for electromagnetic spectrum detection according to claim 2 is characterized in that , ,The electromagnetic map of each iteration is retained. When the electromagnetic spectrum intensity detected in each iteration is different, the electromagnetic spectrum intensity is updated using the arithmetic average method.

4. A multi-UAV collaboration method for electromagnetic spectrum detection according to claim 1, 2 or 3, characterized in that: The method for planning the search path of each unmanned cluster in the respective sub-area includes: The sub-area is divided into grids, and the uncertainty at each grid is represented by η; where uncertainty is a value between 0 and 1. For a grid that has not been detected at all, the uncertainty is 1. As the number of detections increases, the uncertainty η decreases. For a certain grid, the search value function is defined as J = ω1γJ1+ω2γJ2+ω3J3, where J1, J2, and J3 represent the value criteria of the route selection points; ω1, ω2, and ω3 are the weights of the corresponding paths, and the sum of the three is 1; γ is the importance factor, and the importance of the path trajectory selection point is reflected by adjusting γ, and the value criteria are determined by uncertainty and whether it is a no-fly zone, specifically: For the value standard J1, it is determined by the uncertainty of the grid, and the specific relationship is J1 = 20lnαη, where α is the adjustment importance parameter; The value criterion J2 is determined by the coverage that can be increased by selecting this grid, and the specific relationship is J2 = ln△S, where △S is the increased detection area; For the value standard J3, if it is a no-fly zone, the value standard J3 is -65535, and if it is not a no-fly zone, the value standard J3 is 1; The search value function of the path is maximized as the constraint of the path search to obtain the search path.

5. The multi-UAV collaboration method for electromagnetic spectrum detection according to claim 4 is characterized in that: The method for determining the core area that needs to be explored includes: The set distance threshold esp and minimum sample point threshold min_sample; for any detected grid, within the circular area with the distance threshold esp as the radius, determine the number of sample points whose electromagnetic spectrum intensity is greater than the minimum threshold of the electromagnetic spectrum. If the number of sample points is greater than the minimum sample point threshold min_sample, the circular area is the core area.

6. The multi-UAV collaboration method for electromagnetic spectrum detection according to claim 5 is characterized in that: The method for re-defining the sub-areas of the task comprises: Each connected core area is regarded as a sub-area, and the sub-area is regarded as a key exploration area; each connected non-core area is regarded as a sub-area.

7. The multi-UAV collaboration method for electromagnetic spectrum detection according to claim 6 is characterized in that: The method for re-forming the drones according to the re-defined sub-areas includes: The number of drones allocated to each sub-area is α in the formula max is the maximum electromagnetic intensity in all regions, α h is the maximum electromagnetic intensity of the hth sub-region, N s represents the total number of drones, S sum represents the area of ​​all sub-regions, S h represents the area of ​​the hth sub-region; m and k are adjustable parameters, and γ is determined by whether it is a core region, which is 1 or 0, and satisfies Where N is the total number of sub-regions divided.

8. A multi-UAV cooperative system for electromagnetic spectrum detection, characterized in that: It includes a first module, a second module, a third module, a fourth module, a fifth module, a sixth module, a seventh module and an eighth module; The first module is used to determine the mission area where electromagnetic spectrum detection is required; determine the number of drones; and determine the status of drones; The second module divides the mission area into a plurality of sub-areas, and dispatches an unmanned cluster to each sub-area; The third module plans the search path of each unmanned cluster in its respective sub-area; The unmanned cluster of the fourth module detects the electromagnetic spectrum of each sub-area, summarizes and splices the electromagnetic spectrum conditions of each sub-area, and obtains a sparse electromagnetic map of the entire mission area; The fifth module interpolates the obtained sparse electromagnetic map to obtain a global dense electromagnetic map; The sixth module determines the core area that needs to be explored based on the global dense electromagnetic map; The seventh module re-defines the sub-areas of the mission based on the core area of ​​key exploration, and re-forms the drones according to the re-demarcated sub-areas; The eighth module re-detects the electromagnetic spectrum of each sub-area according to the re-formed unmanned cluster and corrects the global dense electromagnetic map.

Images