Multi-unmanned aerial vehicle flight path planning method for radiation source mapping
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
- Filing Date
- 2023-09-28
- Publication Date
- 2026-06-26
AI Technical Summary
Existing technologies struggle to efficiently map multiple UAV radiation sources in unknown environments and cannot optimize flight paths to reduce energy consumption and flight time.
By acquiring initialization parameters, performing spectrum data sampling and clustering, generating an initial path, and using acceptance criteria for multiple optimizations, the optimal flight path is finally generated to complete the radiation source mapping.
It enables the mapping of high-precision radiation sources in unknown environments with shorter energy consumption and flight time, avoiding local optima problems and saving on drone battery consumption.
Smart Images
Figure CN117330075B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of application technology of UAVs in spectrum mapping, specifically involving a multi-UAV flight path planning method for radiation source mapping. Background Technology
[0002] With the rapid development of the information age, the electromagnetic spectrum, as an important resource, has attracted much attention and is crucial to the nation's informatization development. However, with the increase in users, electromagnetic spectrum resources are becoming increasingly scarce, and illegal frequency use has brought severe challenges to electromagnetic spectrum management. In order to effectively utilize spectrum resources and strengthen the control of illegal frequency use, it is necessary to monitor, manage, and control radio frequencies in real time to understand spectrum activities and resource utilization throughout the region, and to promptly detect and deal with illegal radio frequency use. Therefore, constructing accurate radiation source maps is crucial.
[0003] Traditional radiation source mapping involves randomly placing sensors within a region and reconstructing the map using completion algorithms based on the discrete data collected by the sensors. However, for scenarios where it is inconvenient to place fixed sensors, dynamic data acquisition methods are required for radiation source mapping. In this regard, drones (UAVs) have significant advantages due to their mobility and data acquisition capabilities, making them a new direction for radiation source mapping development. The more radiation source sampling points there are, the more difficult it becomes to plan the flight paths of multiple UAVs. Therefore, in the field of multi-UAV radiation source mapping, how to efficiently complete radiation source mapping and how to execute data acquisition tasks with optimal flight paths are important problems that need to be solved. Summary of the Invention
[0004] The technical problem to be solved by the present invention is to address the shortcomings of the prior art by providing a multi-UAV flight path planning method for radiation source mapping, which can map high-precision radiation source maps with less energy consumption and flight time, and is suitable for radiation source mapping in unknown environments.
[0005] To achieve the above-mentioned technical objectives, the technical solution adopted by the present invention is as follows:
[0006] A multi-UAV flight path planning method for radiation source mapping includes:
[0007] Step 1: Obtain initialization parameters, including the area to be tested, the drone's flight altitude, the number of optimization attempts N, and the number of drones;
[0008] Step 2: Sample and cluster the spectrum data of the area to be tested according to the flight altitude of the UAV, and obtain the optimized spectrum data sampling points based on the clustering results;
[0009] Step 3: Generate the initial path of the drones based on the number of drones and the optimized spectrum data sampling points to be tested;
[0010] Step 4: Optimize the initial path of the UAV generated in Step 3 N times to obtain the final path, and convert the final path information into control commands and send them to the UAV.
[0011] Step 5: After receiving the final path information, the UAV begins to fly according to the final path information and collects data in real time. It then sends the collected spectrum data to the computing and processing unit to complete the mapping of the radiation source.
[0012] To optimize the above technical solution, the specific measures also include:
[0013] The specific implementation steps of step 2 above are as follows:
[0014] 2.1) Randomly set sampling locations in the area to be tested, and use drones or other fixed acquisition equipment to collect spectrum data at the sampling locations;
[0015] 2.2) Divide the spectral signal intensity of the spectrum data into m levels, and calculate the difference between the spectrum data obtained in 2.1) and the signal intensity of the m levels;
[0016] Cluster analysis is performed based on the differences to obtain m clusters. For each of the m clusters, the point with the smallest mean intensity between the point in the cluster and other points in the cluster is taken as the cluster center point.
[0017] The spectral map is divided into m regions based on the clusters. Sampling points are assigned according to the cluster size or the distance from the cluster center, resulting in the positions of K sampling points to be measured, denoted as . K i =1,2,...,K.
[0018] The method for calculating the difference mentioned in 2.2) above is as follows:
[0019] E i (x,y)=F(x,y)-f i (i=1,2,...,m) (1)
[0020] Among them, E i (x, y) represents the signal strength F(x, y) at sampling position (x, y) and the signal strength f of the i-th level. i The difference in signal strength between them;
[0021] In step 3 above, when the number of drones is 1, the initial path is formed by the optimized spectrum data sampling points to be tested;
[0022] When the number of drones is greater than 1, the specific steps for generating the initial path are as follows:
[0023] 3.1) Set the starting position of the U-shaped UAV to (x0, y0), where the values of x0 and y0 are usually 0;
[0024] 3.2) Calculate the distances between the sampling points to be measured, and construct the distance matrix D. mat ;
[0025] 3.3) Calculate the number of breakpoints B of the U-shaped UAV. brk and the breakpoint array B nbrk Based on the obtained breakpoints, the optimized initial path R is calculated. rte (1) The remaining initial paths R are obtained by random generation. rte (2) ~R rte (8) .
[0026] In step 2, the K sampling points obtained in step 2 are coded and sorted into numbers 1 to K in any order. The distance D between any two sampling points is calculated as follows:
[0027]
[0028] The above 3.3) uses formulas (3) and (4) to calculate the number of breakpoints B of the U-shaped UAV. brk and the breakpoint array B nbrk :
[0029] Number of points B brk The calculation method is as follows:
[0030] B brk =u-1 (3)
[0031] Breakpoint array B nbrk Including B nbrk (1) and B nbrk (2) ~B nbrk (8) :
[0032] B nbrk (1) =u×n(n=1,2,…,u-1) (4).
[0033] The optimized initial path R is calculated using formulas (5) to (8) in section 3.3 above. rte (1) :
[0034] D min =min[D mat (Pnow [x, x)](x=1,2,…,n)(D min >0) (5)
[0035] D mat (y,x)=0(y=1,2,…,n) (6)
[0036] P now =x (7)
[0037] R rte (1) (i)=P now (i = 1, 2, ..., n) (8)
[0038] Among them, D min D represents mat The size of the minimum value in any row of the matrix, P now This indicates that the current location of the drone is at point D. mat The index of a row in the matrix is initialized to 1, x represents the index of the column containing the minimum value in that row, and y represents D. mat Any row in the matrix, R rte (1) (i) Given a defined set of initial paths, use it as the initial path set for the iterative process; the remaining initial paths R rte (2) ~R rte (n) Randomly generated.
[0039] The specific steps for generating the final path in step 4 above are as follows:
[0040] 4.1) Using R rte (1) ~R rte (n) The paths are candidate paths. For each candidate path, its distance matrix D is used to... mat The values of each element in the algorithm are used to obtain the final flight distance d for the corresponding candidate path, which is the total flight distance. By comparing the total flight distances of each candidate path with the total flight distances of the other paths, the path with the shortest total distance is obtained, which is the new shortest path R. opR ;
[0041] 4.2) Calculate the acceptance criterion parameter A:
[0042]
[0043] D val =R opR (i)-R opR (i-1)(i>1) (9)
[0044] Where T is the test parameter;
[0045] 4.3) Determine whether A is greater than any random number between 0 and 1. If so, the new shortest route is better than the old shortest route, and the new shortest route replaces the old shortest route. Otherwise, retain the original shortest path.
[0046] 4.4) Using the shortest path determined in 4.3) as a sample, randomly combine them. The resulting new path is used as a candidate path in step 4.1). Repeat the following steps until the set number of optimization iterations N is reached, completing the path optimization process and obtaining the final path R. end .
[0047] The present invention has the following beneficial effects:
[0048] The multi-UAV flight path planning method proposed in this invention for radiation source mapping optimizes the randomly generated initial path, reduces the time required for the optimization process, and can find the optimal path for multi-UAV flight planning more quickly.
[0049] This invention proposes a multi-UAV flight path planning method. Based on the acceptance criterion, it avoids the problem of getting trapped in local optima during stochastic optimization, making the flight paths of multiple UAVs more optimal. In practical applications, this helps to save on UAV battery consumption. Attached Figure Description
[0050] Figure 1 This is a flowchart of the multi-UAV flight path planning method for radiation source mapping according to the present invention;
[0051] Figure 2 This is a flowchart illustrating the generation of the initial path in the multi-UAV flight path planning method of the present invention;
[0052] Figure 3 This refers to the sampling point location information set in this embodiment of the invention;
[0053] Figure 4 This is an optimized path planning diagram for a single unmanned aerial vehicle according to an embodiment of the present invention;
[0054] Figure 5 This is an optimized path planning diagram for three drones according to an embodiment of the present invention;
[0055] Figure 6 This is an optimized path planning diagram for four drones according to an embodiment of the present invention. Detailed Implementation
[0056] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.
[0057] Although the steps in this invention are arranged by reference numerals, this is not intended to limit the order of the steps. Unless the order of the steps is explicitly stated or the execution of a step requires other steps as a basis, the relative order of the steps can be adjusted. It is understood that the term "and / or" as used herein refers to and covers any and all possible combinations of one or more of the associated listed items.
[0058] like Figure 1 As shown, the present invention provides a multi-UAV flight path planning method for radiation source mapping, comprising the following steps:
[0059] Step 1: Obtain initialization parameters, including the area to be tested, the drone's flight altitude, the number of optimization attempts N, and the number of drones;
[0060] The user inputs initial parameters into the computer. The initial parameters specifically include the length L and width W of the area to be tested, the drone's flight altitude h, the number of optimization attempts N, and the number of drones u.
[0061] The specific parameters in the embodiments are shown in Table 1.
[0062] Table 1 User Input Parameters
[0063] Length L of the region to be measured 1000m Width W of the area to be measured 1000m Number of drones u 3 Optimization times N 20000 Drone flight altitude h 50m
[0064] Step 2: Sample and cluster the spectrum data of the area to be tested according to the flight altitude of the UAV, and obtain the optimized spectrum data sampling points based on the clustering results;
[0065] Sampling locations are randomly set in the area to be tested. Spectral intensity information is collected using a drone or other fixed acquisition equipment. The collected data is analyzed in real time to understand the approximate intensity distribution in the area. More sampling points are allocated to areas with higher spectral intensity, and fewer sampling points are allocated to areas with lower spectral intensity. The locations of K sampling points are obtained, denoted as K. K i =1,2,...,K.
[0066] The locations of the 50 sampling points to be tested obtained in the example are denoted as follows: K i =1,2,...,50.
[0067] Specifically, the steps for obtaining the location information of the sampling point to be tested in step 2 are as follows:
[0068] 2.1) Randomly set sampling locations in the area to be tested, and use drones or other fixed acquisition equipment to collect spectral data at the sampling locations. Analyze the collected data in real time to understand the approximate intensity distribution in the area to be tested.
[0069] 2.2) Divide the spectral signal intensity of the spectral data into m levels: L1, L2, L3, ..., L m The difference between the spectral data obtained in 2.1) and the signal strength of the m levels is calculated using the following method:
[0070] E i (x,y)=F(x,y)-f i (i=1,2,...,m) (1)
[0071] Among them, E i (x, y) represents the signal strength F(x, y) at sampling position (x, y) and the signal strength f of the i-th level. i The difference in signal strength between them.
[0072] Cluster analysis is performed based on the differences to obtain m clusters. For each of the m clusters after classification, the point with the smallest mean intensity between the point in the cluster and other points in the cluster is calculated as the cluster center point.
[0073] The spectral map is divided into m regions based on the cluster. More sampling points are allocated to regions with higher spectral intensity, and fewer sampling points are allocated to regions with lower spectral intensity. This can be adjusted based on factors such as cluster size or distance from the cluster center, resulting in the positions of K sampling points to be measured, denoted as . K i =1,2,...,K.
[0074] In this embodiment, the spectrum data is divided into three levels, L1, L2, and L3. The difference between the obtained spectrum data and the signal intensity of the three levels is calculated using formula (1). Cluster analysis is performed based on the difference to obtain three clusters. For each of the three clusters, the point with the smallest mean intensity of other points within the cluster is used as the new standard. The spectrum map is divided into three regions based on the clusters. More sampling points are allocated to regions with higher spectrum intensity, and fewer sampling points are allocated to regions with lower spectrum intensity. This can be adjusted according to factors such as cluster size or distance from the cluster center point to obtain the positions of 50 sampling points to be tested, denoted as L1, L2, and L3. K i =1,2...,5.0 The sampling points are set as follows: Figure 3 As shown.
[0075] Step 3: Generate the initial path of the drones based on the number of drones and the optimized spectrum data sampling points to be tested;
[0076] The initial path of the drone is generated based on the initial parameters input by the user and the obtained sampling point information, specifically as follows: Figure 2 As shown.
[0077] Specifically, when the number of drones is 1, the initial path is formed by the optimized spectrum data sampling points to be tested;
[0078] When the number of drones is greater than 1, the specific steps for generating the initial path are as follows:
[0079] 3.1) Set the starting position of the U-shaped UAV to (x0, y0), where the values of x0 and y0 are usually 0;
[0080] In this embodiment, the coordinate system used is a Cartesian coordinate system, the initial position of the UAV is set to (0,0), and the initial movement direction is the sampling point closest to the initial position;
[0081] 3.2) The 50 sampling points are coded and sorted into numbers 1 to 50 in any order. The distance D between any two sampling points is calculated using formula (2). Thus, the distance matrix D between the sampling points can be obtained. mat .
[0082]
[0083] 3.3) Calculate the number of breakpoints B of the U-shaped UAV using formulas (3) and (4). brk and the breakpoint array B nbrk Based on the obtained breakpoints, the data collection tasks of the UAVs are arranged, and it is determined how many data collection points each UAV needs to go to for data collection. This serves as the basis for dividing the initial path. The optimized initial path R is calculated using formulas (5) to (8). rte (1) The remaining initial paths R are obtained by random generation. rte (2) ~R rte (8) .
[0084] The breakpoint array B mentioned above nbrk This indicates the division of drone mission paths. For example, if there are three drones, and the numbers stored in this array are 10 and 20, it means that the flight mission sequence numbers of the first drone are 1-10. Similarly, the flight mission sequence numbers of the second and third drones are 11-20 and 21-50, respectively. This provides more options for the diversity of drone flight missions and the random selection by the genetic algorithm, making it easier to find the optimal solution.
[0085] For cases where u is greater than or equal to 2, the sampling points to be tested need to be divided. Multiple paths are determined based on the number of drones u. Breakpoints are designed between these sampling points, and these breakpoints are used to divide the number of sampling points to be tested for each drone.
[0086] Number of breakpoints B brkThe calculation method is as follows:
[0087] B brk =u-1 (3)
[0088] After obtaining the number of breakpoints, create an array B to store the breakpoints. nbrk Including B nbrk (1) and B nbrk (2) ~B nbrk (8) :
[0089] B nbrk (1) =u×n(n=1,2,…,u-1) (4)
[0090] The breakpoints of the first group of paths are obtained by formula (4), and the remaining breakpoints are B. nbrk (2) ~B nbrk (8) The number of elements is denoted as B, and is generated randomly. brk ;
[0091] An optimized initial path is obtained using breakpoint calculations, as follows:
[0092] D min =min[D mat (P now [x, x)](x=1,2,…,n)(D min >0) (5)
[0093] D mat (y,x)=0(y=1,2,…,n) (6)
[0094] P now =x (7)
[0095] R rte (1) (i)=P now (i = 1, 2, ..., n) (8)
[0096] Among them, D min D represents mat The size of the minimum value in any row of the matrix, P now This indicates that the current location of the drone is at point D. mat The index of a row in the matrix is initialized to 1, x represents the index of the column containing the minimum value in that row, and y represents D. mat Any row in the matrix, R rte (1)(i) Given a defined set of initial paths, use it as the initial path set for the iterative process; the remaining initial paths R rte (2) ~R rte (n) Randomly generated.
[0097] Step 4: Based on the initial path of the UAV generated in Step 3, perform N optimizations to obtain the final path, and convert the final path information into control commands and send them to the UAV.
[0098] Specifically, the steps for generating the final path in step 4 are as follows:
[0099] 4.1) Using R rte (1) ~R rte (n) The paths are candidate paths. For each candidate path, its distance matrix D is used to... mat The values of each element in the algorithm are used to obtain the final flight distance d for the corresponding candidate path, which is the total flight distance. By comparing the total flight distances of each candidate path with the total flight distances of the other paths, the path with the shortest total distance is obtained, which is the new shortest path R. opR ;
[0100] 4.2) Since random search may lead to the trap of local optima, it is necessary to determine whether to accept the new solution. The specific implementation method is as follows:
[0101]
[0102] Judge D val If the value is less than 0, then the new route is superior to the old route, and the new route replaces the old route; if D... val If the value is greater than 0, then the following processing is required:
[0103]
[0104] Where T is the test parameter used, which gradually decreases as the number of optimizations increases. When T is lower, it is less likely to accept the new route. The greater the difference in the total distance between the old and new routes, the less likely it is to accept the new solution.
[0105] 4.3) Determine whether to accept the new shortest route based on probability. Determine whether A is greater than any random number between 0 and 1. If the condition is met, the new shortest route is better than the old shortest route, and the new shortest route replaces the old shortest route. Otherwise, the new solution is not accepted, and the original shortest path is retained.
[0106] 4.4) Using the shortest path determined in 4.3) as a sample, perform various processing steps, including random combination. The resulting new path is then used as a candidate path in step 4.1). Repeat these steps until the set number of optimization iterations N is reached, completing the path optimization process and obtaining the final path R. end .
[0107] For the case where u equals 1, steps 1 to 4 are executed normally, without needing to execute the step in step 3 where breakpoints are generated, and the final path R can be obtained. end .
[0108] In this embodiment, the final output flight path diagram is as follows: Figure 5 As shown. Figure 4 and Figure 6 These are the optimized flight path maps using a single drone and four drones, respectively.
[0109] Step 5: After receiving the final path information, the UAV begins to fly according to the final path information and collects data in real time. It then sends the collected spectrum data to the computing and processing unit to complete the mapping of the radiation source.
[0110] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.
[0111] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.
Claims
1. A multi-UAV flight path planning method for radiation source mapping, characterized in that, include: Step 1: Obtain initialization parameters, including the test area, UAV flight altitude, and number of optimization attempts. And the number of drones; Step 2: Sample and cluster the spectrum data of the area to be tested according to the flight altitude of the UAV, and obtain the optimized spectrum data sampling points based on the clustering results; Step 3: Generate the initial path of the drones based on the number of drones and the optimized spectrum data sampling points to be tested; In step 3, when the number of drones is 1, the initial path is formed by the optimized spectrum data sampling points to be tested. When the number of drones is greater than 1, the specific steps for generating the initial path are as follows: 3.1) Let The starting position of the drone is set to ,in , The value is 0; 3.2) Calculate the distances between the sampling points to be measured, and construct a distance matrix. ; 3.3) Calculation Number of breakpoints of drones and the array of breakpoints Based on the obtained breakpoints, the optimized initial path is calculated. The remaining initial paths are obtained by random generation. ; The optimized initial path is calculated using formulas (5) to (8) in section 3.3). : （5） （6） （7） （8） in, express The size of the minimum value in any row of the matrix. Indicates the current location of the drone. The index value of the row in the matrix, initially set to 1. This indicates the index value of the column containing the minimum value in that row. express Any row in the matrix, Given a defined set of initial paths, use these as the initial paths for the iterative process; the remaining initial paths... Randomly generated; Step 4: Based on the initial path of the drone generated in Step 3, perform... The optimization process yields the final path, and the final path information is converted into control commands and sent to the drone. Step 5: After receiving the final path information, the UAV begins to fly according to the final path information and collects data in real time. It then sends the collected spectrum data to the computing and processing unit to complete the mapping of the radiation source.
2. The multi-UAV flight path planning method for radiation source mapping according to claim 1, characterized in that, The specific implementation steps of step 2 are as follows: 2.1) Randomly set sampling locations in the area to be tested, and use drones or other fixed acquisition equipment to collect spectrum data at the sampling locations; 2.2) Divide the spectral signal intensity of the spectral data into... Each level, for the spectral data obtained in 2.1), calculate its correlation with... The difference between signal strength levels; Cluster analysis was performed based on the differences to obtain... Each cluster, for For each cluster, the point with the smallest mean intensity among all points within the cluster is selected as the cluster center. Based on the cluster, the spectrum map is divided into Each region is assigned sampling points based on cluster size or distance from the cluster center. The location of each sampling point to be tested is denoted as... .
3. The multi-UAV flight path planning method for radiation source mapping according to claim 2, characterized in that, 2.2) The method for calculating the difference is as follows: （1） in, The signal strength at sampling location (x, y) With the signal strength of the i-th level The difference in signal strength between them.
4. The multi-UAV flight path planning method for radiation source mapping according to claim 1, characterized in that, In step 2, the result obtained in step 2 is described in section 3.2). The sampling points are encoded and sorted in any order as follows: The distance between any two sampling points The calculation method is as follows: （2）。 5. The multi-UAV flight path planning method for radiation source mapping according to claim 1, characterized in that, The calculation in section 3.3) uses formulas (3) and (4). Number of breakpoints of drones and the array of breakpoints : Number of points The calculation method is as follows: （3） Array of breakpoints ,include and : （4）。 6. The multi-UAV flight path planning method for radiation source mapping according to claim 1, characterized in that, The specific steps for generating the final path in step 4 are as follows: 4.1) with The paths are candidate paths. For each candidate path, its distance matrix is used to... The values of each element in the algorithm are used to obtain the final flight distance of the corresponding candidate path. The total distance traveled is calculated by comparing it with the total distances required for each candidate path, and the path with the shortest total distance is selected as the new shortest path. ; 4.2) Calculate the acceptance criterion parameters : （10） （9） in, For test parameters; 4.3) Judgment If the new shortest path is greater than any random number between 0 and 1, then the new shortest path is better than the old shortest path, and the new shortest path is used to replace the old shortest path; otherwise, the original shortest path is retained. 4.4) Using the shortest path determined in 4.3) as a sample, randomly combine them. The resulting new path is used as a candidate path in step 4.1). Repeat the following steps until the set number of optimizations is reached. Complete the path optimization process to obtain the final path. .