CN117556979A - Unmanned plane platform and load integrated design method based on group intelligent search - Google Patents
Unmanned plane platform and load integrated design method based on group intelligent search Download PDFInfo
- Publication number
- CN117556979A CN117556979A CN202410038924.4A CN202410038924A CN117556979A CN 117556979 A CN117556979 A CN 117556979A CN 202410038924 A CN202410038924 A CN 202410038924A CN 117556979 A CN117556979 A CN 117556979A
- Authority
- CN
- China
- Prior art keywords
- cluster
- unmanned aerial
- aerial vehicle
- index
- search
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 102
- 238000013461 design Methods 0.000 title claims abstract description 26
- 238000005457 optimization Methods 0.000 claims abstract description 167
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 152
- 239000002245 particle Substances 0.000 claims abstract description 47
- 238000010206 sensitivity analysis Methods 0.000 claims abstract description 21
- 230000000694 effects Effects 0.000 claims abstract description 20
- 230000035945 sensitivity Effects 0.000 claims description 54
- 230000006870 function Effects 0.000 claims description 42
- 230000008569 process Effects 0.000 claims description 35
- 239000003016 pheromone Substances 0.000 claims description 27
- 230000003993 interaction Effects 0.000 claims description 19
- 230000006399 behavior Effects 0.000 claims description 15
- 238000013459 approach Methods 0.000 claims description 12
- 230000009194 climbing Effects 0.000 claims description 12
- 238000010845 search algorithm Methods 0.000 claims description 11
- 238000012546 transfer Methods 0.000 claims description 10
- 230000000007 visual effect Effects 0.000 claims description 9
- 238000000354 decomposition reaction Methods 0.000 claims description 5
- 230000019637 foraging behavior Effects 0.000 claims description 5
- 238000003239 susceptibility assay Methods 0.000 claims description 5
- 241000257303 Hymenoptera Species 0.000 claims description 4
- 238000004220 aggregation Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000001550 time effect Effects 0.000 claims description 4
- 101100532584 Clostridium perfringens (strain 13 / Type A) sspC1 gene Proteins 0.000 claims description 3
- 101100410359 Drosophila melanogaster Patronin gene Proteins 0.000 claims description 3
- 102100032055 Elongation of very long chain fatty acids protein 1 Human genes 0.000 claims description 3
- 102100032050 Elongation of very long chain fatty acids protein 2 Human genes 0.000 claims description 3
- 101000771413 Homo sapiens Aquaporin-9 Proteins 0.000 claims description 3
- 101000921370 Homo sapiens Elongation of very long chain fatty acids protein 1 Proteins 0.000 claims description 3
- 101000921368 Homo sapiens Elongation of very long chain fatty acids protein 2 Proteins 0.000 claims description 3
- 101100256651 Homo sapiens SENP6 gene Proteins 0.000 claims description 3
- 101100095550 Homo sapiens SENP7 gene Proteins 0.000 claims description 3
- 101150038317 SSP1 gene Proteins 0.000 claims description 3
- 101150098865 SSP2 gene Proteins 0.000 claims description 3
- 101100125012 Saccharomyces cerevisiae (strain ATCC 204508 / S288c) ECM10 gene Proteins 0.000 claims description 3
- 101100125020 Schizosaccharomyces pombe (strain 972 / ATCC 24843) pss1 gene Proteins 0.000 claims description 3
- 101100018019 Schizosaccharomyces pombe (strain 972 / ATCC 24843) ssc1 gene Proteins 0.000 claims description 3
- 102100023645 Sentrin-specific protease 3 Human genes 0.000 claims description 3
- 102100023713 Sentrin-specific protease 6 Human genes 0.000 claims description 3
- 102100031406 Sentrin-specific protease 7 Human genes 0.000 claims description 3
- 108010092928 Stomoxys serine protease 3 Proteins 0.000 claims description 3
- 230000002860 competitive effect Effects 0.000 claims description 3
- 230000003247 decreasing effect Effects 0.000 claims description 3
- 230000007717 exclusion Effects 0.000 claims description 3
- 238000003384 imaging method Methods 0.000 claims description 3
- 230000011273 social behavior Effects 0.000 claims description 3
- 241000251468 Actinopterygii Species 0.000 claims 1
- 238000012544 monitoring process Methods 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 16
- 238000011160 research Methods 0.000 description 6
- 230000032683 aging Effects 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 5
- 241001137251 Corvidae Species 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000001174 ascending effect Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000002431 foraging effect Effects 0.000 description 2
- 238000013508 migration Methods 0.000 description 2
- 230000005012 migration Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 235000001968 nicotinic acid Nutrition 0.000 description 1
- 238000013433 optimization analysis Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/231—Hierarchical techniques, i.e. dividing or merging pattern sets so as to obtain a dendrogram
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06313—Resource planning in a project environment
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Artificial Intelligence (AREA)
- Educational Administration (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biodiversity & Conservation Biology (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention discloses an unmanned aerial vehicle platform and load integrated design method based on group intelligent search, which aims to optimize unmanned aerial vehicle group task planning and platform-load performance index matching and improve timeliness and accuracy of task planning. According to the invention, a cluster search task planning algorithm framework is established, and an improved sequential clustering algorithm, an ant colony algorithm, a particle swarm optimization algorithm, a crow search optimization algorithm and other multi-algorithm combined optimization strategies are fused, so that scientific decisions of the number of unmanned aerial vehicles, optimal task scheduling and index matching in two-dimensional and three-dimensional spaces are realized. According to the invention, the performance indexes of the unmanned aerial vehicle and the load are finely adjusted by combining sensitivity analysis and a multi-objective optimization function, so that the completion of the optimal multi-objective search task is ensured. The method is suitable for large-scale cluster tasks, can remarkably improve the application efficiency and the operation effect of the unmanned aerial vehicle cluster, and has important application value in the fields of urban logistics, ocean monitoring, border security, regional search and rescue and the like.
Description
Technical Field
The invention belongs to the technical field of unmanned aerial vehicle task planning, and relates to an unmanned aerial vehicle platform and load integrated design method, in particular to an unmanned aerial vehicle platform and load integrated design method based on a group intelligent search algorithm.
Background
Unmanned aerial vehicle Cluster (Unmanned Aerial Vehicle, UAV, cluster) refers to a collaborative operation system composed of a plurality of unmanned aerial vehicles, and is realized through advanced communication and algorithm technology, so that centralized control and collaborative work of a plurality of unmanned aerial vehicles are realized to complete tasks. The unmanned aerial vehicle cluster has wide application prospect in the fields of urban logistics, ocean monitoring, border security, regional search and rescue and the like due to the advantages of large quantity, large scale, group elasticity and the like. Particularly, aiming at some regional search tasks with strong aging requirements and a large number of targets, a large number of low-cost unmanned aerial vehicles can be dispatched to form a cluster to a task region, the number advantage is exerted, the task region is effectively covered, and the task is completed rapidly. However, despite the advantages of unmanned aerial vehicle clusters, they still face a number of challenges in practical operation. One of the key problems is the complexity of task planning, and particularly under the conditions of high task aging requirements and high target quantity, the method not only relates to path planning of each unmanned aerial vehicle, but also comprises optimal scheduling and management of the whole cluster, and relates to the aspects of unmanned aerial vehicle quantity determination, route planning, platform-load index matching and the like. The existing task planning method often cannot fully utilize the potential of the cluster, so how to perform task planning is a problem worthy of research.
Aiming at the problems of demonstration and matching of unmanned plane platform performance and photoelectric load (Optical-Electronic Payload) index performance, the prior art generally simplifies the processing of the unmanned plane platform performance and the photoelectric load (Optical-Electronic Payload) index performance, and does not consider the load index and the task matching capability. From task planning to unmanned aerial vehicle design, the load is regarded as accessories to be selected and installed, the matching performance of the load and the tasks is not fully considered, the optimal index based on the cluster task is difficult to realize, and the potential of the cluster cannot be fully utilized. Aiming at large-scale cluster tasks, only the load index capacity and the tasks and unmanned aerial vehicle platforms are comprehensively researched, so that the cluster efficiency is better exerted. At present, less researches are carried out on photoelectric load, unmanned aerial vehicle performance and task requirements, the load capacity-task requirements are modeled generally based on an optimization idea, and optimal indexes under targets such as aging, energy consumption and the like are solved. The work makes active researches on the unmanned aerial vehicle cluster task planning problem, and achieves great achievements.
However, unmanned aerial vehicle mission planning problems remain a point of continued research. In the face of strong aging, multitasking targets, most research is based on a given number of unmanned aerial vehicles, however, the number of unmanned aerial vehicles is itself a problem worth studying. In the face of numerous targets, how many unmanned aerial vehicles to dispatch, how to follow each unmanned aerial vehicle on the target route is not a simple problem, and too many or too few unmanned aerial vehicles can affect task completion quality and efficiency. In addition, in the two-dimensional level, the optimal task scheduling is not enough to meet the actual requirements. On the basis of two dimensions, the unmanned aerial vehicle performance and the photoelectric load performance are considered, the problem is raised to three dimensions, the influence of unmanned aerial vehicle platforms and photoelectric load performance indexes on task completion capacity is analyzed, and an optimal index analysis result is given, so that the problem is more in line with reality. Therefore, in order to efficiently utilize the unmanned aerial vehicle cluster, how to determine the scale, route and platform-load index matching and the like of the unmanned aerial vehicle cluster according to the task requirements and the target number, the deep research and the solution of the problems are key points for improving the application efficiency and the effect of the unmanned aerial vehicle cluster, and are also technical problems to be solved urgently.
Disclosure of Invention
Object of the invention
Aiming at the defects and shortcomings in the prior art, in order to solve the problems of how to determine the scale, route and platform-load index matching of unmanned aerial vehicle clusters according to task requirements and target quantity, the invention aims to provide an unmanned aerial vehicle platform and load integrated design method based on intelligent group search. The method can effectively improve timeliness and accuracy of task planning, simultaneously realize optimal matching of the unmanned aerial vehicle platform and load performance indexes, enhance cluster efficiency of unmanned aerial vehicle clusters when complex tasks are executed, and has important significance for improving operation efficiency of urban logistics, ocean monitoring, border security, regional search and rescue and the like.
(II) technical scheme
In order to achieve the aim of the invention, the invention adopts the following technical scheme:
the unmanned aerial vehicle platform and load integrated design method based on group intelligent search is characterized by at least comprising the following steps of:
SS1, in the initialization stage of task planning, carrying out sequential clustering processing on target points of a task area, according to geographic coordinate information of the target points,applying a sequential clustering algorithm to divide the disordered target point clusters into a plurality of cluster circles, wherein each cluster circle corresponds to the task allocation of one unmanned aerial vehicle, and introducing in the clustering processdAndΔdtwo optimization parameters are used for adjusting the clustering process, and the optimization parameters are used fordThe diameter of the cluster circle is represented, the number of the cluster circles and the number of target points which can be covered by each cluster are directly influenced, and parameters are optimizedΔdRepresenting the minimum position deviation between the initial point of a new cluster and the center of the previous cluster in the cluster transfer process and directly influencing the relative position between the cluster circles, and adjusting the size and the relative position of the cluster circles by optimizing the two parameters so as to optimize the spatial distribution of the unmanned aerial vehicle clusters and the uniform coverage of a target area and target points in the target area, and simultaneously ensuring the geographic position of the target points in each cluster circle to be close to avoid unnecessary track overlapping, thereby reducing the transfer distance of the unmanned aerial vehicle between the target points and improving the task efficiency;
SS2. For the set of target points determined in each cluster circle, applying a single-machine optimal track algorithm based on an ant colony optimization algorithm (Ant Colony Optimization Algorithm) to plan a shortest two-dimensional track for each unmanned aerial vehicle to approach an optimal solution of a traveler problem (Traveling Salesman Problem, TSP), wherein the ant colony optimization algorithm simulates the shortest path principle of natural ant foraging behavior, and iteratively searches the shortest possible track of the unmanned aerial vehicle from a starting point to each target point and then returns to the starting point;
SS3. Combining the application of the cluster task planning optimization algorithm, and clustering the diameter of the circle by a multi-objective optimization functiondAnd cluster circle positioning distanceΔdOptimizing to optimize the number of unmanned aerial vehicles, the average two-dimensional track length, the longest two-dimensional track length and the standard deviation of the two-dimensional track lengths of all unmanned aerial vehicles so as to realize global optimization of the number of unmanned aerial vehicles and the overall two-dimensional track configuration scheme, wherein the expression of the multi-objective optimization function is as follows:
in the method, in the process of the invention,Nis the number of unmanned aerial vehicles in the cluster and is equal to the number of unmanned aerial vehicles in the clusterThe number of cluster circles is determined,L i for the two-dimensional track length of each single unmanned aerial vehicle in the cluster, i=1, 2, …,N，mean{L i the average two-dimensional track length of each single unmanned aerial vehicle in the cluster is max {L i The maximum two-dimensional track length of each single unmanned aerial vehicle in the cluster,σ 1 two-dimensional track length for each single unmanned aerial vehicle in clusterL i Is set in the standard deviation of (2),w 1 、w 2 、w 3 、w 4 the weighting coefficient of each corresponding index is represented and can be adjusted according to different task requirements;
SS4. In three-dimensional space route planning, comprehensively considering the platform performance and photoelectric load performance indexes of the unmanned aerial vehicle, analyzing the imaging requirement of the photoelectric load and the flight characteristics of the unmanned aerial vehicle, and at least selecting the climbing and descending angle of the unmanned aerial vehicleαVisible distance of photoelectric load L gd And the viewing angle of the photoelectric loadθ gd Three index parameters are used as main indexes for influencing three-dimensional search paths and task time of unmanned aerial vehicle, wherein the climbing descending angle of the unmanned aerial vehicleαRepresenting the visual distance of photoelectric load of unmanned plane from cruising altitude to plane flying altitude or from plane flying altitude to cruising altitudeL gd Representing the maximum horizontal distance that the photoelectric load can detect the target point under the flat flying height, and the visible angle of the photoelectric loadθ gd Representing the maximum elevation angle at which the photoelectric load can detect the target point at the fly height;
SS5. Use of Sobol sensitivity assayα、L gd Andθ gd task time sensitivity analysis is carried out on the three index parameters, global sensitivity index and first-order sensitivity index of each index parameter are calculated based on variance decomposition, and each single index parameter and each index parameter combination pair optimization objective function are analyzed according to the global sensitivity index and the first-order sensitivity indexf(α, L gd,, θ gd ) Judging the influence degree of each index parameter on the total time effect of the task and determining the key point of the index parameter optimization according to the influence degree, wherein the higher the first-order sensitivity index is, the single index parameter isThe larger the number impact, the larger the global sensitivity index indicates a greater interaction of the index parameter with other index parameters;
SS6. Based on the result of Sobol sensitivity analysis in step SS5, adopting a clustered task planning optimization algorithm to optimize the objective function f(α, L gd,, θ gd ) For a pair ofα、L gd And/orθ gd Optimizing three index parameters to optimize the average search task time and the maximum search task time of each single unmanned aerial vehicle in the cluster, wherein the optimization objective functionf(α, L gd,, θ gd ) The expression of (2) is:
in the method, in the process of the invention,T i the search task time for each single-machine drone in the cluster is represented by i=1, 2, …,N，Nfor the number of unmanned aerial vehicles in the cluster, mean {T i The average task searching time of each single unmanned aerial vehicle in the cluster,w 5 、w 6 the weighting coefficient is represented and can be adjusted according to different task requirements;
and SS7, determining the optimal unmanned aerial vehicle cluster scale, the flight path of each unmanned aerial vehicle and the unmanned aerial vehicle platform and load performance index according to the optimization results of the steps SS3 and SS6, and ensuring that the unmanned aerial vehicle cluster can complete the multi-target search task in the shortest time.
Preferably, in step SS1, the sequential clustering algorithm includes at least the following sub-steps when implemented:
SS1.1. Input coordinates { of the set of target pointsP tar { set point coordinates }, {P jjd ' and reference point coordinatesP 0 ；
SS1.2 initializes algorithm parameters including, but not limited to, cluster circle diameterdAnd cluster circle positioning distance deltad；
SS1.3 determination of distance reference pointP 0 The nearest pointP ini i- And take it as the firstiInitial points of cluster circles and ensure P ini i- To any existing cluster center pointCIs greater than or equal to the cluster circle positioning distance deltad；
SS1.4 at the initial pointP ini i- As a starting point, search at diameterdTarget point within range furthest from itP far-i And determine the first according to the aboveiCircle center of each clusterC i Coordinates of (c), wherein；
SS1.5 toC i Is the center of a circle and is based on the diameter of the cluster circledDrawing a cluster circle from a target point set {P tar Searching and recording target points falling in the cluster circle according to the distance to form a target point cluster set {P ci，j And is formed into the firstiCluster circles, whereinjIs the firstiSequence numbers of target points in the clustering circles;
SS1.6 at the firstiIndividual target point cluster set {P ci，j In } for each cluster centerC i Selecting the nearest target point as the first target pointiAggregation points within a cluster circleP jjd i, ；
SS1.7 { at target point setP tar Exclusion of the clustering set { that has been partitioned into target points }P ci，j Target points in the set, and updating the remaining set of target points;
SS1.8 repeatedly executes steps SS 1.3-SS 1.7 until the target point set {P tar And the target points are all traversed and properly distributed into the clustering circles, so that all the clustering circles and the target points and the collecting points inside the clustering circles are obtained.
Further, in the step SS1.3, the distance reference point is determined P 0 The nearest pointP ini i- When using nearest neighbor search algorithm (Nearest Neighbor Search Algorithm), according to each targetGeographical coordinate information of points, calculating target points and reference pointsP 0 Is determined as the Euclidean distance (Euclidean Distance) and the target point corresponding to the smallest distance is selected asP ini i- ；
In the above step SS1.4, the diameter isdIn-range search distance initial pointP ini i- Furthest target pointP far-i At the time, using the furthest neighbor search algorithm (Farthest Neighbor Search Algorithm), calculating the target point and the initial point according to the geographic coordinate information of the target pointP ini i- And selecting the target point corresponding to the maximum distance as the Euclidean distanceP far-i ；
In the above step SS1.5, the first search is performed according to the distance and the record is madeiIndividual target point cluster set {P ci，j When in use, a radius search algorithm (Radius Search Algorithm) is used to calculate the circle centers of the target points and the clusters according to the geographic coordinate information of the target pointsC i And choose from the Euclidean distance of (2) less than or equal todThe target point corresponding to the distance of/2 is taken as {P ci，j Elements of };
in step SS1.8, the iteration is performed until the target point set {P tar When the target points are empty and all target points are traversed, a dynamic updating algorithm (Dynamic Update Algorithm) is used for updating the target point set { in real time according to the geographic coordinate information of the target points P tar Elements of { whenever a target point is divided into a cluster circle, from {P tar Delete the target point until {P tar No element in }.
Preferably, in the step SS2, for the set of target points determined in each cluster circle, an ant colony optimization algorithm is adopted to solve the shortest track path of the unmanned aerial vehicle, so as to obtain a two-dimensional track of each unmanned aerial vehicle, and the specific flow of the ant colony optimization algorithm at least comprises the following substeps:
SS2.1. Initializing a number of artificial ants, each representing a possible track path, and randomly assigning them to respective target points within a cluster circle;
SS2.2. Each artificial ant selects the next target point to be accessed in the cluster circle according to a certain probability, the selection basis is the distance between the target points and the concentration of the pheromone, and the pheromone is released in the access process, the selection basis of the target point to be accessed is the distance between the target points and the concentration of the pheromone, the closer the distance is, the more the pheromone is, and the larger the selection probability is;
SS2.3. Each artificial ant returns to the starting point after accessing all the target points in the cluster circle, forming a complete track path, and calculating the length of the track path;
SS2.4. Updating the pheromone concentration, increasing and decreasing the pheromone on the path of each path travelled by the artificial ant according to the path length and the pheromone volatilization coefficient, and releasing more pheromones when the path is shorter;
SS2.5. Pheromones on each track path are gradually volatilized over time, so that old track path information is gradually desalted, and space is provided for new track path information;
and SS2.6. Repeating the steps SS 2.1-SS 2.5 until the preset maximum iteration times are reached or the shortest path meeting the conditions is found, and outputting the shortest track path.
Preferably, in the steps SS3 and SS6, the objective optimization function is optimized by using the clustered task planning optimization algorithmf(d, Δd) Or (b)f(α, L gd,, θ gd ) Optimizing to obtain the optimal cluster circle diameterdAnd cluster circle positioning distanceΔdOr optimallyα、L gd And/orθ gd And the cluster task planning optimization algorithm is a particle swarm optimization algorithm PSO and/or a crow search optimization algorithm CSO.
Further, the specific flow of the particle swarm optimization algorithm PSO at least comprises the following substeps:
SSP1. Initializing a number of particles, each particle representing one possible solution, i.e. a set of cluster diametersdAnd cluster circle positioning room Distance from each otherΔdValues or a set of values of (2)α、L gd And/orθ gd Randomly assigned to a solution space;
SSP2, calculating the fitness value of each particle, namely the value of a multi-objective optimization function, and determining an individual optimal solution and a global optimal solution of each particle according to the fitness value;
SSP3, updating the speed and the position of each particle according to the speed and the position of the particle, the individual optimal solution and the global optimal solution, and moving towards the direction of the optimal solution;
SSP4, repeating the steps SSP 1-SSP 3 until the preset iteration times or convergence conditions are reached, and outputting a global optimal solution.
Further, the specific flow of the crow search optimization algorithm CSO comprises the following sub-steps:
SSC1. Initializing a number of crow's, each crow representing a possible solution, i.e. a set of cluster diametersdAnd cluster circle positioning distanceΔdValues or a set of values of (2)α、L gd And/orθ gd Randomly assigning values to the solution space and assigning an initial position and velocity to each crow;
SSC2. Set maximum iteration timesT max And fitness thresholdθ fit And initializing a global optimal solutiong best A larger value;
SSC3, in each generation of circulation, calculating and updating the fitness value and the optimal solution of each crow, updating the speed and the position of each crow according to a certain strategy and probability, carrying out cooperation and competition behaviors, and judging whether a termination condition is reached;
SSC4. Returning to Global optimal solutiong best I.e. optimal cluster diameterdAnd cluster circle positioning distanceΔdOr the optimum value of (2)α、L gd And/orθ gd Is a value of (2).
Further, in the above step SSC3, for each cycle, the following is performedt=1 toT max At least the following operations are performedThe steps are as follows:
SSC3.1 for each crow, calculate its fitness value, i.e., the value of the multi-objective optimization function, based on its current location and update its individual optimal solution based on its fitness valuep best ；
SSC3.2. If an individual best solution for crow' sp best Is superior to the global optimal solutiong best Will theng best Updated top best ；
SSC3.3 updating the velocity and position of each crow to a globally optimal solution based on inertial weights, learning factors, stochastic factors, current velocity, location of individual optimal solutions and/or globally optimal solutionsg best Or an individual optimal solutionp best Approaching;
SSC3.4. Carrying out cooperative behavior, for each crow, selecting the currently known optimal solution in the surrounding crow, namely the solution with the highest fitness value, as the reference of the search strategy, and updating the position of the solution to be the position of the optimal solution or the position near the optimal solution according to a certain probability;
SSC3.5. Performing competitive behaviors, namely selecting the currently known worst solution, namely the solution with the lowest fitness value, in surrounding crow's as the opponent of a search strategy according to a certain probability, and if a solution which is better than the worst solution, namely the solution with the higher fitness value, is found, updating the position of the solution to be the position of the better solution or the position near the better solution;
SSC3.6. Judging whether the termination condition is reached, i.e. the iteration number reaches the maximum iteration numberT max Or the fitness value of the globally optimal solution is lower than a threshold valueθ fit If so, the cycle is ended, and if not, the next generation cycle is continued.
Further, the clustered task planning optimization algorithm simultaneously uses a particle swarm optimization algorithm PSO and a crow search optimization algorithm CSO to optimize a multi-objective optimization function, the particle swarm optimization algorithm PSO simulates the group behaviors of shoves or shoves in nature, the speed and the position of each particle are iteratively updated to enable the particle to approach to a global optimal solution or an individual optimal solution, the crow search optimization algorithm CSO simulates the cooperation and competition behaviors of crows in nature, the speed and the position of each crow are updated by probability to enable the particle to approach to or depart from the optimal solution or the worst solution in surrounding crows, if the results obtained by the two algorithms are the same, the original problem is proved to be converged, otherwise iteration is continued until termination conditions are met.
Preferably, in the above step SS5, a Sobol sensitivity assay method is used for the reactionα、L gd Andθ gd when the task time sensitivity analysis is carried out on the three index parameters, the implementation flow at least comprises the following substeps:
SS5.1. Will optimize the objective functionf(α, L gd,, θ gd ) Expressed as a function of a single index parameter and a combination of index parameters, namely:
wherein,X 1 =α，X 2 =L gd ，X 3 =θ gd ，f 0 is a constant term which is used to determine the degree of freedom,f i (X i ) For the first order effect term, the influence of a single index parameter on the objective function is represented,f ij (X i ,X j ) For the second order interaction effect term, the effect of interaction between two index parameters on the objective function is represented,f 123 (X 1 ,X 2 ,X 3 ) The three-order interaction effect item represents the influence of interaction among three index parameters on an objective function;
SS5.2. Calculation of the optimization objective functionf(α, L gd,, θ gd ) Is the total variance of (2)DBias of each effect termD i 、D ij AndD 123 wherein:
SS5.3. Calculating the first-order sensitivity index for each index parameterS i Second order sensitivity indexS ij And a third order sensitivity indexS 123 Wherein:
SS5.4. Calculating the Global sensitivity index for each index parameterTS i Wherein:
SS5.5. According to the first order sensitivity indexS i Global sensitivity indexTS i Judging the size of each index parameter pair to optimize the objective functionf(α, L gd,, θ gd ) Is the total variance of (2)DIs the first order sensitivity indexS i The larger the impact representing a single index parameter, the larger the global sensitivity indexTS i The larger the index parameter, the greater the interaction with other index parameters.
(III) technical effects
Compared with the prior art, the unmanned aerial vehicle platform and load integrated design method based on group intelligent search has the following beneficial and remarkable technical effects:
(1) The invention provides an unmanned aerial vehicle platform and load integrated design method based on group intelligent search, which is characterized in that a group search task planning algorithm framework is established, and a multi-algorithm combined optimization strategy such as a sequential clustering algorithm, an ant colony algorithm, a particle swarm optimization algorithm, a crow search optimization algorithm and the like is combined, so that scientific decision on the number of unmanned aerial vehicles, and optimal task scheduling and index matching on two-dimensional and three-dimensional layers are realized. The method can effectively cope with the regional search task of the strong aging and multitasking targets, realize the integrated design of the scale, the route and the platform-load index of the unmanned aerial vehicle cluster, improve the spatial distribution of the unmanned aerial vehicle cluster and the uniform coverage of the target region, reduce the transfer distance of the unmanned aerial vehicle between the target points, shorten the search task time of the unmanned aerial vehicle, and improve the application efficiency and the effect of the unmanned aerial vehicle cluster.
(2) According to the invention, the performance of the unmanned aerial vehicle platform and the photoelectric load performance index are taken as important factors influencing the task completion capacity, the problem is raised to a three-dimensional space, three index parameters such as the climbing down angle, the photoelectric load visual distance and the photoelectric load visual angle of the unmanned aerial vehicle are comprehensively considered, and the index parameters are optimized by using a Sobol sensitivity analysis method and a cluster task planning optimization algorithm, so that the optimal matching of the unmanned aerial vehicle platform and the load performance index is realized, the timeliness and the accuracy of task planning are improved, and the cluster efficiency of the unmanned aerial vehicle cluster is enhanced.
(3) According to the method, a sequential clustering method based on position coordinates is adopted for clustering, two optimization amounts of cluster circle diameter and cluster circle positioning distance are introduced, the size and the relative position of the cluster circles are adjusted through optimizing the two optimization amounts, so that the spatial distribution of unmanned aerial vehicle clusters and the uniform coverage of target areas and target points in the target areas are optimized, meanwhile, the geographic positions of the target points in each cluster circle are ensured to be close, unnecessary track overlapping is avoided, and therefore the transfer distance of unmanned aerial vehicles among the target points is reduced, and the task efficiency is improved. The method also adopts a single-machine optimal track algorithm based on an ant colony optimization algorithm, plans the shortest two-dimensional track for each unmanned aerial vehicle to approach the optimal solution of the problem of the tourist, simulates the shortest path principle of the natural ant foraging behavior, and iteratively searches the shortest possible track of the unmanned aerial vehicle from the starting point to each target point and then returns to the starting point.
(4) According to the invention, by adopting a cluster task planning optimization algorithm which uses a particle swarm optimization algorithm and a crow search optimization algorithm simultaneously, index parameters such as cluster circle diameter, cluster circle positioning distance or climbing descending angle of unmanned aerial vehicles, photoelectric load visual distance and photoelectric load visual angle are optimized through a multi-objective optimization function, so that the number of unmanned aerial vehicles, average two-dimensional track length, longest two-dimensional track length, standard deviation of each two-dimensional track length or average search task time and maximum search task time are optimized, and global optimization of unmanned aerial vehicle number and overall two-dimensional track configuration scheme or unmanned aerial vehicle platform and load performance index is realized. The algorithm utilizes the global searching capability of the particle swarm optimization algorithm and the local searching capability of the crow search optimization algorithm, and through cooperation and competition behaviors, the problem of local optimization caused by a single algorithm is avoided, and the reliability and stability of an optimization result are improved.
(5) According to the method, the task time sensitivity analysis is carried out on the unmanned aerial vehicle platform and the load performance index by adopting a Sobol sensitivity analysis method, the global sensitivity index and the first-order sensitivity index of each index parameter are calculated based on variance decomposition, the contribution of each single index parameter and each index parameter combination to the total variance of the optimization objective function is analyzed according to the global sensitivity index and the first-order sensitivity index, the influence degree of each index parameter on the total time effect of the task is judged, and the key point of index parameter optimization is determined according to the influence degree.
Drawings
FIG. 1 is a frame diagram of an unmanned aerial vehicle platform and load integrated design method based on group intelligent search;
FIG. 2 is a schematic diagram of a target point and a departure point in the sequential clustering method of the present invention;
fig. 3 is a schematic diagram of solving the shortest path problem of the unmanned aerial vehicle based on an ant colony optimization algorithm;
FIG. 4 shows the clustering diameter in the present inventiond=30 km and cluster circle positioning pitchΔdSchematic of validation test performed=0 km;
FIG. 5 shows the followingd=40 km, clusteringΔdSchematic diagram of validation test =5 km;
Fig. 6 is a schematic diagram of a drone employing a classical search strategy;
FIG. 7 is a schematic diagram of fitness function in the optimization process of CSO algorithm and PSO algorithm;
fig. 8 is a schematic diagram of the number of unmanned aerial vehicles in the optimization process of the CSO algorithm and the PSO algorithm;
FIG. 9 is a schematic diagram showing the average path length during optimization of the CSO algorithm and the PSO algorithm;
FIG. 10 is a schematic diagram of the longest path length in the optimization of the CSO algorithm and the PSO algorithm;
FIG. 11 is a schematic diagram of standard deviation of each machine path in the optimization process of CSO algorithm and PSO algorithm;
FIG. 12 is a schematic diagram of the optimization paths of each machine in the optimization process of the CSO algorithm and the PSO algorithm;
FIG. 13 is a schematic diagram showing the sensitivity analysis results;
FIG. 14 is a schematic diagram showing fitness functions in the optimization process of the CSO algorithm and the PSO algorithm;
FIG. 15 is a schematic diagram showing average time in the optimization process of the CSO algorithm and the PSO algorithm;
fig. 16 is a schematic diagram showing the longest time in the optimization process of the CSO algorithm and the PSO algorithm.
Detailed Description
For a better understanding of the present invention, the following examples are set forth to illustrate the present invention. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some, but not all, embodiments of the invention. The embodiments described below by referring to the drawings are illustrative and intended to explain the present invention and should not be construed as limiting the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention. The following describes the structure and technical scheme of the present invention in detail with reference to the accompanying drawings, and an embodiment of the present invention is given.
Example 1
Fig. 1 is a frame diagram of an unmanned aerial vehicle platform and load integrated design method based on group intelligent search. As shown in fig. 1, the unmanned aerial vehicle platform and load integrated design method based on group intelligent search at least comprises the following steps:
SS1, in the initialization stage of task planning, carrying out sequential clustering treatment on target points of a task area, applying a sequential clustering algorithm according to geographic coordinate information of the target points, dividing disordered target point clusters into a plurality of cluster circles, wherein each cluster circle corresponds to task allocation of one unmanned aerial vehicle, and introducing in the clustering processdAndΔdtwo optimization parameters are used for adjusting the clustering process, and the optimization parameters are used fordThe diameter of the cluster circle is represented, the number of the cluster circles and the number of target points which can be covered by each cluster are directly influenced, and parameters are optimizedΔdRepresenting the minimum position deviation between the initial point of a new cluster and the center of the previous cluster in the cluster transfer process and directly influencing the relative position between the cluster circles, and adjusting the size and the relative position of the cluster circles by optimizing the two parameters so as to optimize the spatial distribution of the unmanned aerial vehicle clusters and the uniform coverage of a target area and target points in the target area, and simultaneously ensuring the geographic position of the target points in each cluster circle to be close to avoid unnecessary track overlapping, thereby reducing the transfer distance of the unmanned aerial vehicle between the target points and improving the task efficiency;
SS2, for the set of target points determined in each cluster circle, a single-machine optimal track algorithm based on an ant colony optimization algorithm is applied to plan a shortest two-dimensional track for each unmanned aerial vehicle so as to approach an optimal solution of a tourist problem, the ant colony optimization algorithm simulates a shortest path principle of natural ant foraging behaviors, and the shortest possible track of the unmanned aerial vehicle from a starting point to each target point and returning to the starting point is searched through iteration;
SS3. Combining the application of the cluster task planning optimization algorithm, and clustering the diameter of the circle by a multi-objective optimization functiondAnd cluster circle positioning distanceΔdOptimizing to optimize the number of unmanned aerial vehicles, the average two-dimensional track length, the longest two-dimensional track length and the standard deviation of the two-dimensional track lengths of all unmanned aerial vehicles so as to realize global optimization of the number of unmanned aerial vehicles and the overall two-dimensional track configuration scheme, wherein the expression of the multi-objective optimization function is as follows:
in the method, in the process of the invention,Nfor the number of drones in the cluster and equal to the number of cluster circles,L i for the two-dimensional track length of each single unmanned aerial vehicle in the cluster, i=1, 2, …,N，mean{L i the average two-dimensional track length of each single unmanned aerial vehicle in the cluster is max {L i The maximum two-dimensional track length of each single unmanned aerial vehicle in the cluster, σ 1 Two-dimensional track length for each single unmanned aerial vehicle in clusterL i Is set in the standard deviation of (2),w 1 、w 2 、w 3 、w 4 the weighting coefficient of each corresponding index is represented and can be adjusted according to different task requirements;
SS4. In three-dimensional space route planning, comprehensively considering the platform performance and photoelectric load performance indexes of the unmanned aerial vehicle, analyzing the imaging requirement of the photoelectric load and the flight characteristics of the unmanned aerial vehicle, and at least selecting the climbing and descending angle of the unmanned aerial vehicleαVisible distance of photoelectric loadL gd And the viewing angle of the photoelectric loadθ gd Three index parameters are used as main indexes for influencing three-dimensional search paths and task time of unmanned aerial vehicle, wherein the climbing descending angle of the unmanned aerial vehicleαRepresenting the visual distance of photoelectric load of unmanned plane from cruising altitude to plane flying altitude or from plane flying altitude to cruising altitudeL gd Representing the maximum horizontal distance that the photoelectric load can detect the target point under the flat flying height, and the visible angle of the photoelectric loadθ gd Representing the maximum elevation angle at which the photoelectric load can detect the target point at the fly height;
SS5. Use of Sobol sensitivity assayα、L gd Andθ gd task time sensitivity analysis is carried out on the three index parameters, global sensitivity index and first-order sensitivity index of each index parameter are calculated based on variance decomposition, and each single index parameter and each index parameter combination pair optimization objective function are analyzed according to the global sensitivity index and the first-order sensitivity index f(α, L gd,, θ gd ) Judging the influence degree of each index parameter on the total time effect of the task and determining the key point of the index parameter optimization according to the influence degree, wherein the larger the first-order sensitivity index is, the larger the influence of a single index parameter is, and the larger the global sensitivity index is, the larger the interaction between the index parameter and other index parameters is;
SS6. Based on the result of Sobol sensitivity analysis in step SS5, adopting a clustered task planning optimization algorithm to optimize the objective functionf(α, L gd,, θ gd ) For a pair ofα、L gd And/orθ gd Optimizing three index parameters to optimize the average search task time and the maximum search task time of each single unmanned aerial vehicle in the cluster, wherein the optimization objective functionf(α, L gd,, θ gd ) The expression of (2) is:
in the method, in the process of the invention,T i the search task time for each single-machine drone in the cluster is represented by i=1, 2, …,N，Nfor the number of unmanned aerial vehicles in the cluster, mean {T i The average task searching time of each single unmanned aerial vehicle in the cluster,w 5 、w 6 the weighting coefficient is represented and can be adjusted according to different task requirements;
and SS7, determining the optimal unmanned aerial vehicle cluster scale, the flight path of each unmanned aerial vehicle and the unmanned aerial vehicle platform and load performance index according to the optimization results of the steps SS3 and SS6, and ensuring that the unmanned aerial vehicle cluster can complete the multi-target search task in the shortest time.
Example 2
In this embodiment, on the basis of embodiment 1, the related steps are further refined and perfected.
In step SS1 of the present invention, the sequential clustering algorithm includes at least the following sub-steps:
SS1.1. Input coordinates { of the set of target pointsP tar Coordinates of the aggregation point{P jjd ' and reference point coordinatesP 0 ；
SS1.2 initializes algorithm parameters including, but not limited to, cluster circle diameterdAnd cluster circle positioning distance deltad；
SS1.3 determination of distance reference pointP 0 The nearest pointP ini i- And take it as the firstiInitial points of cluster circles and ensureP ini i- To any existing cluster center pointCIs greater than or equal to the cluster circle positioning distance deltad；
SS1.4 at the initial pointP ini i- As a starting point, search at diameterdTarget point within range furthest from itP far-i And determine the first according to the aboveiCircle center of each clusterC i Coordinates of (c), wherein；
SS1.5 toC i Is the center of a circle and is based on the diameter of the cluster circledDrawing a cluster circle from a target point set {P tar Searching and recording target points falling in the cluster circle according to the distance to form a target point cluster set {P ci，j And is formed into the firstiCluster circles, whereinjIs the firstiSequence numbers of target points in the clustering circles;
SS1.6 at the firstiIndividual target point cluster set {P ci，j In } for each cluster center C i Selecting the nearest target point as the first target pointiAggregation points within a cluster circleP jjd i, ；
SS1.7 { at target point setP tar Exclusion of the clustering set { that has been partitioned into target points }P ci，j Target points in the set, and updating the remaining set of target points;
SS1.8 repeatedly executes steps SS 1.3-SS 1.7 until the target point set {P tar Null, ensuring that all target points are traversed and properly traversedAnd distributing the cluster circles to obtain all the cluster circles and the target points and the collecting points in the cluster circles.
In the above sub-step SS1.3, the distance reference point is determinedP 0 The nearest pointP ini i- When the method is used, a nearest neighbor search algorithm is used, and the target point and the reference point are calculated according to the geographic coordinate information of each target pointP 0 And selecting the target point corresponding to the minimum distance from the Euclidean distanceP ini i- ；
In the above step SS1.4, the diameter isdIn-range search distance initial pointP ini i- Furthest target pointP far-i When the method is used, the furthest neighbor searching algorithm is used, and the target point and the initial point are calculated according to the geographic coordinate information of the target pointP ini i- And selecting the target point corresponding to the maximum distance as the Euclidean distanceP far-i ；
In the above step SS1.5, the first search is performed and recordediIndividual target point cluster set {P ci，j When in use, a radius search algorithm is used to calculate the circle centers of the target points and the clusters according to the geographic coordinate information of the target points C i And choose from the Euclidean distance of (2) less than or equal todThe target point corresponding to the distance of/2 is taken as {P ci，j Elements of };
in step SS1.8, the iteration is performed until the target point set {P tar In the process that all target points are traversed, a dynamic updating algorithm is used for updating the target point set { in real time according to the geographic coordinate information of the target pointsP tar Elements of { whenever a target point is divided into a cluster circle, from {P tar Delete the target point until {P tar No element in }.
In step SS2 of the present invention, for the set of target points determined in each cluster circle, an ant colony optimization algorithm is adopted to solve the shortest track path of the unmanned aerial vehicle, so as to obtain a two-dimensional track of each unmanned aerial vehicle, and the specific flow of the ant colony optimization algorithm at least comprises the following sub-steps:
SS2.1. Initializing a number of artificial ants, each representing a possible track path, and randomly assigning them to respective target points within a cluster circle;
SS2.2. Each artificial ant selects the next target point to be accessed in the cluster circle according to a certain probability, the selection basis is the distance between the target points and the concentration of the pheromone, and the pheromone is released in the access process, the selection basis of the target point to be accessed is the distance between the target points and the concentration of the pheromone, the closer the distance is, the more the pheromone is, and the larger the selection probability is;
SS2.3. Each artificial ant returns to the starting point after accessing all the target points in the cluster circle, forming a complete track path, and calculating the length of the track path;
SS2.4. Updating the pheromone concentration, increasing and decreasing the pheromone on the path of each path travelled by the artificial ant according to the path length and the pheromone volatilization coefficient, and releasing more pheromones when the path is shorter;
SS2.5. Pheromones on each track path are gradually volatilized over time, so that old track path information is gradually desalted, and space is provided for new track path information;
and SS2.6. Repeating the steps SS 2.1-SS 2.5 until the preset maximum iteration times are reached or the shortest path meeting the conditions is found, and outputting the shortest track path.
In the steps SS3 and SS6 of the present invention, the objective optimization function is optimized by using the clustered task planning optimization algorithmf(d, Δd) Or (b)f(α, L gd,, θ gd ) Optimizing to obtain the optimal cluster circle diameterdAnd cluster circle positioning distanceΔdOr optimallyα、L gd And/orθ gd And the cluster task planning optimization algorithm is a particle swarm optimization algorithm PSO and/or a crow search optimization algorithm CSO.
The specific flow of the particle swarm optimization algorithm PSO at least comprises the following sub-steps:
SSP1. Initializing a number of particles, each particle representing one possible solution, i.e. a set of cluster diametersdAnd cluster circle positioning distanceΔdValues or a set of values of (2)α、L gd And/orθ gd Randomly assigned to a solution space;
SSP2, calculating the fitness value of each particle, namely the value of a multi-objective optimization function, and determining an individual optimal solution and a global optimal solution of each particle according to the fitness value;
SSP3, updating the speed and the position of each particle according to the speed and the position of the particle, the individual optimal solution and the global optimal solution, and moving towards the direction of the optimal solution;
SSP4, repeating the steps SSP 1-SSP 3 until the preset iteration times or convergence conditions are reached, and outputting a global optimal solution.
The specific flow of the crow search optimization algorithm CSO comprises the following substeps:
SSC1. Initializing a number of crow's, each crow representing a possible solution, i.e. a set of cluster diametersdAnd cluster circle positioning distanceΔdValues or a set of values of (2)α、L gd And/orθ gd Randomly assigning values to the solution space and assigning an initial position and velocity to each crow;
SSC2. Set maximum iteration timesT max And fitness thresholdθ fit And initializing a global optimal solution g best A larger value;
SSC3, in each generation of circulation, calculating and updating the fitness value and the optimal solution of each crow, updating the speed and the position of each crow according to a certain strategy and probability, carrying out cooperation and competition behaviors, and judging whether a termination condition is reached;
SSC4. Returning to Global optimal solutiong best I.e. optimal cluster diameterdAnd cluster circle positioning distanceΔdOr the optimum value of (2)α、L gd And/orθ gd Is a value of (2).
In the above sub-step SSC3, for each cycle, the sequence is selected fromt=1 toT max At least the following operating steps are performed:
SSC3.1 for each crow, calculate its fitness value, i.e., the value of the multi-objective optimization function, based on its current location and update its individual optimal solution based on its fitness valuep best ；
SSC3.2. If an individual best solution for crow' sp best Is superior to the global optimal solutiong best Will theng best Updated top best ；
SSC3.3 updating the velocity and position of each crow to a globally optimal solution based on inertial weights, learning factors, stochastic factors, current velocity, location of individual optimal solutions and/or globally optimal solutionsg best Or an individual optimal solutionp best Approaching;
SSC3.4. Carrying out cooperative behavior, for each crow, selecting the currently known optimal solution in the surrounding crow, namely the solution with the highest fitness value, as the reference of the search strategy, and updating the position of the solution to be the position of the optimal solution or the position near the optimal solution according to a certain probability;
SSC3.5. Performing competitive behaviors, namely selecting the currently known worst solution, namely the solution with the lowest fitness value, in surrounding crow's as the opponent of a search strategy according to a certain probability, and if a solution which is better than the worst solution, namely the solution with the higher fitness value, is found, updating the position of the solution to be the position of the better solution or the position near the better solution;
SSC3.6. Judging whether the termination condition is reached, i.e. the iteration number reaches the maximum iteration numberT max Or the fitness value of the globally optimal solution is lower than a threshold valueθ fit If so, the cycle is ended, and if not, the next generation cycle is continued.
The cluster task planning optimization algorithm of the invention simultaneously uses a particle swarm optimization algorithm PSO and a crow search optimization algorithm CSO to respectively optimize a multi-objective optimization function, wherein the particle swarm optimization algorithm PSO simulates the swarm behaviors of shoals or shoals in nature, the speed and the position of each particle are iteratively updated to enable each particle to approach to a global optimal solution or an individual optimal solution, the crow search optimization algorithm CSO simulates the cooperation and competition behaviors of crows in nature, the speed and the position of each crow are updated by probability to enable each crow to approach to or depart from the optimal solution or the worst solution in surrounding crows, if the results obtained by the two algorithms are the same, the original problem is proved to be converged, otherwise iteration is continued until the termination condition is met.
In step SS5 of the present invention, a Sobol sensitivity assay method is used for the reactionα、L gd Andθ gd when the task time sensitivity analysis is carried out on the three index parameters, the implementation flow at least comprises the following substeps:
SS5.1. Will optimize the objective functionf(α, L gd,, θ gd ) Expressed as a function of a single index parameter and a combination of index parameters, namely:
wherein,X 1 =α，X 2 =L gd ，X 3 =θ gd ，f 0 is a constant term which is used to determine the degree of freedom,f i (X i ) For the first order effect term, the influence of a single index parameter on the objective function is represented,f ij (X i ,X j ) For the second order interaction effect term, the effect of interaction between two index parameters on the objective function is represented,f 123 (X 1 ,X 2 ,X 3 ) The three-order interaction effect item represents the influence of interaction among three index parameters on an objective function;
SS5.2. Calculation of the optimization objective functionf(α, L gd,, θ gd ) Is the total variance of (2)DBias of each effect termD i 、D ij AndD 123 wherein:
SS5.3. Calculating the first-order sensitivity index for each index parameterS i Second order sensitivity indexS ij And a third order sensitivity indexS 123 Wherein:
SS5.4. Calculating the Global sensitivity index for each index parameterTS i Wherein:
SS5.5. According to the first order sensitivity indexS i Global sensitivity indexTS i Judging the size of each index parameter pair to optimize the objective functionf(α, L gd,, θ gd ) Is the total variance of (2)DIs the first order sensitivity index S i The larger the impact representing a single index parameter, the larger the global sensitivity indexTS i The larger the index parameter, the greater the interaction with other index parameters.
Example 3
In the technical scheme adopted by the invention, the first part is mainly used for researching the quantity allocation of unmanned aerial vehicles and the optimal path problem of each single machine in the cluster based on an optimization algorithm and a clustering algorithm. The whole algorithm comprises a sequential clustering algorithm, a single machine optimal track algorithm and a cluster task planning optimization algorithm.
The clustering algorithm is a classification algorithm, and aims to group target points which are seemingly disordered, so that the positions of the target points in the same group are similar, the single frame machine is ensured to save long-distance transfer, and meanwhile, the positions of the target points in different groups are as far as possible, so that the overlapping of the position areas of the unmanned aerial vehicles is avoided, and the resource waste is caused. The classical clustering algorithm is represented by a K-means clustering algorithm, which is an iterative solution algorithm and has the core that data are divided into K groups, K objects are randomly selected as initial clustering centers, then the distance between each object and each seed clustering center is calculated, and each object is allocated to the closest clustering center. This approach is highly exploratory, but often comes with high uncertainty and high iterative computation time costs.
Because the invention introduces multiple optimizations in the proposed algorithm, in the early stage of calculation, the uncertainty in the optimization process is hoped to be reduced, and the result can be interpreted and optimized as much as possible. Therefore, the invention provides a sequential clustering method for clustering based on position coordinates based on a classical clustering method and taking the position of a coordinate point as a basis. In the clustering process, the diameter of a clustering circle is introduceddAnd cluster circle positioning distanceΔdTwo optimization variables. Cluster circle diameterdThe diameter of the clustering circle is represented, and the number of clusters and the number of clustering parcel targets are affected; clustering circle positioning distanceΔdThe minimum position deviation between the initial point of a new cluster and the circle center of the previous cluster in the cluster transfer process affects the positions among the cluster circles to adjust the relative positions among the cluster circles, thereby improving the uniformity degree of the clustering of the dense target area and reducing the space crossing of the cluster circles, as shown in figure 2.
In the sequential clustering process, the target points contained in each cluster correspond to tasks allocated by each unmanned aerial vehicle and need to be traversed. This problem is a typical traveler problem TSP. The traveller problem is one of the well-known problems in the mathematical arts. Each unmanned aerial vehicle needs to visit n target points, the unmanned aerial vehicle needs to select a path to be taken, and the limitation of the path is that each target can only visit once and finally returns to the original starting position. The path is selected with the goal that the required path length is the minimum among all paths. TSP is a typical combinatorial optimization problem and is an NP-complete challenge, and no perfect solution is currently found.
As shown in fig. 3, the method solves the shortest path problem of the unmanned aerial vehicle based on the ant colony optimization algorithm, and approximates to the optimal solution. The fundamental principle of the ant colony algorithm is derived from the shortest path principle of foraging in nature. According to the observation of the entomologist, ants can find the shortest path from the food source to the nest without any prompt, and can still search the best path when the environment is complex.
Before being optimized, the diameter of the clusters is clusteredd =30 km, cluster circle positioning pitchΔd =0 kmd =40 km, clusteringΔd =5 km, searching the shortest path through sequential clustering and ant colony optimization algorithm, and starting with the starting point of the unmanned aerial vehicle and ending with the last target point of each machine when calculating the path length, wherein the results are shown in fig. 4 and 5. At the position ofd = 40 km、Δd =At 0 km, searching for a cluster of 18 unmanned aerial vehicle units to be dispatched for 100 targets, and searching for the longest path length max {L i The shortest path length is 34.52 km, and the average path length mean { is 116.53 kmL i After 70.50 and km, the standard deviation of the path length of each machine is 23.58. At the position ofd = 40 km、Δd =5 km, searching for a cluster of 100 targets, which requires dispatching 13 unmanned aerial vehicle sets altogether, and the longest path length max {L i 141.87 km, a shortest path length of 26.49 km, and an average path length mean { L i After 86.55 km, each machine path lengthσ 1 Is 33.97. From the test results shown in Table 1, it can be seen that different cluster diametersdDistance between cluster circlesΔdThe selection of the number of unmanned aerial vehicles in the cluster, the longest path length and the like are obviously influenced, and a large optimization space is provided.
/>
On the basis of the algorithm, an optimization algorithm is introduced, optimization is carried out by combining task requirements, and an optimization objective function is as follows
，
Wherein,Nfor the number of drones, equal to the number of clusters,L i i=1, 2, …, N for the path length of each individual machine in the cluster;σ 1 standard deviation of path length of each machine in the cluster;w 1 、w 2 、w 3 andw 4 the representation weighting coefficients can be adjusted according to different task requirements. The optimization quantity is the cluster diameterdAnd cluster circle positioning distanceΔd 。
In the optimization process, in order to improve the explanatory performance of an optimization result and avoid the problem of local optimization caused by a single algorithm, the particle swarm optimization algorithm PSO and the crow selection search optimization algorithm are adopted, and meanwhile, the objective function is optimized. The particle swarm optimization (Particle Swarm Optimization, PSO) algorithm is an optimization algorithm based on the principle of bionics. It finds the optimal solution of the problem by simulating the behavior of group collaboration in a shoal or shoal. The PSO algorithm consists of a population of individuals called particles. Each particle represents one possible solution in the solution space. They search for the optimal solution by iteratively adjusting their own position and velocity. The crow search optimization algorithm (Crow Search Optimization, CSO) is a heuristic optimization algorithm based on the behavior of a shoal, and simulates a search strategy of crow in the foraging process. The algorithm finds the optimal solution by simulating the foraging behavior of the crow. The basic idea of the crow search optimization algorithm is to consider the problem to be optimized as a search space, the crow group represents the solution in the solution space by the position in the search space, and the quality of the solution is optimized by the actions of cooperation and competition.
In the technical scheme adopted by the invention, the second part is unmanned plane platform-photoelectric load index analysis and optimization considering three-dimensional route duration. In the task, the unmanned aerial vehicle adopts a classical search strategy, and after the unmanned aerial vehicle starts from a collection point, the speed is kept at 100 km/h, and the unmanned aerial vehicle cruises along a fixed heightHAfter reaching the vicinity of the target point, the target point is slid down, then the target point is imaged in a short distance while the target point is kept flying at a lower height, and the target point is slid up after imagingThe next cut into cruising altitude, until the next target, as shown in figure 6.
For the photoelectric load of the unmanned aerial vehicle, the visual range of the target point is a sphere top cone, and the vertex of the top cone surface and the sphere center are coincided. After the unmanned aerial vehicle approaches the target from the cruising height, the unmanned aerial vehicle cuts into the round edge at the top of the ball cone rapidly, namely, the target is seen, then the unmanned aerial vehicle keeps flying flatly, after reaching the position right above the target, the unmanned aerial vehicle deflects the course to turn to the next target point to fly flatly, after reaching the round edge at the top of the ball cone, the vertical horse starts climbing, rapidly reaches the cruising height, and continues flying to the next target. If the distance from the next target is insufficient to reach the cruising altitude, the target is turned into a dive in advance, and the target is visited. In the invention, in order to simplify the problem, the climbing and descending angle of the unmanned aerial vehicle is selected αVisible distance of angle photoelectric loadL gd And the viewing angle of the photoelectric loadθ gd Analysis was performed. The three parameters are independent of each other, but all have influence on the three-dimensional search path of the unmanned aerial vehicle, so that the task time of the cluster is influenced.
In order to explore the significance degree of the influence of three indexes on task time, the sensitivity of the three indexes is analyzed by using a Sobol sensitivity analytical analysis method. The Sobol sensitivity analysis method is a representative global sensitivity analysis method and is based on a variance decomposition method, and the central idea is to set a model as a function of a single parameter and the mutual combination of the parameters, so as to decompose the total variance of an objective function into the first-order variances of the parameters in the function and the second-order and higher-order variances formed by the interaction between the parameters.
Example 4
In the optimization process of the invention, when using a PSO algorithm, the population particle number is 40, the maximum iteration is 100 rounds, the learning factors are 1.5 and 1.5 respectively, and the maximum value and the minimum value of the inertia weight are 0.8 and 0.4 respectively. When the CSO algorithm is used, the crow number is set to 40, the maximum iteration is 100, the flight distance is 2, and the exploration probability is 0.2. Cluster diameterDThe value range of (2) is defined as [20, 100 ]]Clustering circle positioning distance ΔdThe value range of (2) is set as 0, 50]，w 1 、w 2 、w 3 Andw 4 taking 1, 0.5 and 0.5 respectively. The optimization process and the result are shown in fig. 7-12 and table 2.
Based on the optimization process and the results shown in fig. 7-12 and table 2, both the PSO and CSO methods can achieve convergence rapidly in terms of trend, and the PSO algorithm achieves fewer algebra for convergence. Comparing fitness, number, average path length, longest path length and standard deviation of path length of each machine, and obtaining the same result, which indicates that the original problem has reached convergence. Aiming at the original optimization task, the cluster is formed by dispatching 27 sets of unmanned aerial vehicles, and the multi-target search can be completed approximately fastest. The two optimization methods are used, the obtained paths of all machines are the same, and the stability of the solution of the original problem obtained by using the algorithm is high and is less influenced by the algorithm. In addition, in the optimization results of the two solving methods, the values of the optimal cluster diameters are almost equal, the values of the cluster migration distances are not different, and finally, no difference exists in each result, which also shows that the cluster migration distances in the original algorithm have a certain tolerance to be overlarge.
Task time sensitivity analysis is carried out on three indexes by using a Sobol sensitivity analysis method, and the unmanned aerial vehicle climbs down at a descending angle αVisible distance of photoelectric loadL gd And the viewing angle of the photoelectric loadθ gd The intervals of (2) are respectively 0, 45](degree), [1,3](km) and [0, 60](degree), the parameter sample space was set to 10000, and the sensitivity analysis result is shown in fig. 13. From the result analysis, unmanned aerial vehicle climbs angle of descentαThe global sensitivity and the first-order sensitivity of (2) are respectively 0.73 and 0.97, the influence on task time is most direct, and the visible distance of the photoelectric load isL gd And the viewing angle of the photoelectric loadθ gd The global sensitivity and the first-order sensitivity of (a) are 0.0.02, 0.18 and 0.02, 0.17 respectively, the influence on task time is relatively small, but cannot be ignored, and the global sensitivity and the first-order sensitivity should be considered in index optimization demonstration.
After sensitivity analysis, an optimization algorithm is adopted, task time is calculated on the basis of the cluster multi-target distribution and the two-dimensional optimization track obtained in the previous step, and optimization analysis is carried out on three indexes. Optimizing an objective function as
，
Wherein the method comprises the steps ofT i Represents the search task time of each individual machine in the cluster,w 5 、w 6 representing the weighting coefficients. The optimized quantity is the climbing descending angle of the unmanned aerial vehicleαVisible distance of photoelectric loadL gd And the viewing angle of the photoelectric loadθ gd . The optimization process and the result are shown in fig. 14-16 and table 3.
Based on the optimization process of fig. 14-16 and the optimization results shown in table 3, both methods can quickly achieve convergence from the trend. Comparing the results of the two methods, the optimization result of the PSO algorithm is trapped in a local optimal value near the beginning interval of the variable, and the final whole-process longest task time is 64.75 min. The CSO algorithm has stronger exploratory property, achieves a relatively better result, and the final whole-process optimal task time is 64.39 min. The optimization results of the two algorithms are similar to each other aiming at the ascending and descending angles of the photoelectric load and are near the interval 1 km, so that the indexes of the photoelectric load can meet the basic requirements without being too high aiming at the task. The indexes optimized by the two algorithms are not higher than 20 degrees aiming at the ascending and descending angles of the unmanned aerial vehicle, which indicates that the unmanned aerial vehicle cannot quickly climb and descend in the scene, and basically keeps maneuvering in the pitching direction and walks the shortest task route when flying in the task area. Aiming at the photoelectric visual angle, the optimized indexes have larger difference and can be in coupling relation with the aircraft climbing descending angle index, and in practical application, the demonstration can be carried out based on a more accurate task model.
The object of the present invention is fully effectively achieved by the above-described embodiments. Those skilled in the art will appreciate that the present invention includes, but is not limited to, those illustrated in the drawings and described in the foregoing detailed description. While the invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not limited to the disclosed embodiment, but on the contrary, is intended to cover various modifications and equivalent arrangements included within the scope of the appended claims.
Claims (10)
1. The unmanned aerial vehicle platform and load integrated design method based on group intelligent search is characterized by at least comprising the following steps of:
SS1, in the initialization stage of task planning, carrying out sequential clustering treatment on target points of a task area, applying a sequential clustering algorithm according to geographic coordinate information of the target points, dividing disordered target point clusters into a plurality of cluster circles, wherein each cluster circle corresponds to task allocation of one unmanned aerial vehicle, and introducing in the clustering processdAndΔdtwo optimization parameters are used for adjusting the clustering process, and the optimization parameters are used fordThe diameter of the cluster circle is represented, the number of the cluster circles and the number of target points which can be covered by each cluster are directly influenced, and parameters are optimized ΔdRepresenting the minimum position deviation between the initial point of a new cluster and the center of the previous cluster in the cluster transfer process and directly influencing the relative position between the cluster circles, and adjusting the size and the relative position of the cluster circles by optimizing the two parameters so as to optimize the spatial distribution of the unmanned aerial vehicle clusters and the uniform coverage of a target area and target points in the target area, and simultaneously ensuring the geographic position of the target points in each cluster circle to be close to avoid unnecessary track overlapping, thereby reducing the transfer distance of the unmanned aerial vehicle between the target points and improving the task efficiency;
SS2, for the set of target points determined in each cluster circle, a single-machine optimal track algorithm based on an ant colony optimization algorithm is applied to plan a shortest two-dimensional track for each unmanned aerial vehicle so as to approach an optimal solution of a tourist problem, the ant colony optimization algorithm simulates a shortest path principle of natural ant foraging behaviors, and the shortest possible track of the unmanned aerial vehicle from a starting point to each target point and returning to the starting point is searched through iteration;
SS3. Combining the application of the cluster task planning optimization algorithm, and clustering the diameter of the circle by a multi-objective optimization functiondAnd cluster circle positioning distanceΔdOptimizing to optimize the number of unmanned aerial vehicles, the average two-dimensional track length, the longest two-dimensional track length and the standard deviation of the two-dimensional track lengths of all unmanned aerial vehicles so as to realize global optimization of the number of unmanned aerial vehicles and the overall two-dimensional track configuration scheme, wherein the expression of the multi-objective optimization function is as follows:
In the method, in the process of the invention,Nfor the number of drones in the cluster and equal to the number of cluster circles,L i for the two-dimensional track length of each single unmanned aerial vehicle in the cluster, i=1, 2, …,N，mean{L i the average two-dimensional track length of each single unmanned aerial vehicle in the cluster is max {L i The maximum two-dimensional track length of each single unmanned aerial vehicle in the cluster,σ 1 two-dimensional track length for each single unmanned aerial vehicle in clusterL i Is set in the standard deviation of (2),w 1 、w 2 、w 3 、w 4 the weighting coefficient of each corresponding index is represented and can be adjusted according to different task requirements;
SS4. In three-dimensional space route planning, comprehensively considering the platform performance and photoelectric load performance indexes of the unmanned aerial vehicle, analyzing the imaging requirement of the photoelectric load and the flight characteristics of the unmanned aerial vehicle, and at least selecting the climbing and descending angle of the unmanned aerial vehicleαVisible distance of photoelectric loadL gd And the viewing angle of the photoelectric loadθ gd Three index parameters are used as main indexes for influencing three-dimensional search paths and task time of unmanned aerial vehicle, wherein the climbing descending angle of the unmanned aerial vehicleαRepresenting the visual distance of photoelectric load of unmanned plane from cruising altitude to plane flying altitude or from plane flying altitude to cruising altitudeL gd Representing the maximum horizontal distance that the photoelectric load can detect the target point under the flat flying height, and the visible angle of the photoelectric load θ gd Representing the maximum elevation angle at which the photoelectric load can detect the target point at the fly height;
SS5. Use of Sobol sensitivity assayα、L gd Andθ gd task time sensitivity analysis is carried out on the three index parameters, global sensitivity index and first-order sensitivity index of each index parameter are calculated based on variance decomposition, and each single index parameter and each index parameter combination pair optimization objective function are analyzed according to the global sensitivity index and the first-order sensitivity indexf(α, L gd,, θ gd ) Judging the influence degree of each index parameter on the total time effect of the task and determining the key point of the index parameter optimization according to the influence degree, wherein the larger the first-order sensitivity index is, the larger the influence of a single index parameter is, and the larger the global sensitivity index is, the larger the interaction between the index parameter and other index parameters is;
SS6. Based on the result of Sobol sensitivity analysis in step SS5, adopting a clustered task planning optimization algorithm to optimize the objective functionf(α, L gd,, θ gd ) For a pair ofα、L gd And/orθ gd Optimizing three index parameters to optimize the average search task time and the maximum search task time of each single unmanned aerial vehicle in the cluster, wherein the optimization objective functionf(α, L gd,, θ gd ) The expression of (2) is:
in the method, in the process of the invention,T i the search task time for each single-machine drone in the cluster is represented by i=1, 2, …,N，Nfor the number of unmanned aerial vehicles in the cluster, mean { T i The average task searching time of each single unmanned aerial vehicle in the cluster,w 5 、w 6 the weighting coefficient is represented and can be adjusted according to different task requirements;
and SS7, determining the optimal unmanned aerial vehicle cluster scale, the flight path of each unmanned aerial vehicle and the unmanned aerial vehicle platform and load performance index according to the optimization results of the steps SS3 and SS6, and ensuring that the unmanned aerial vehicle cluster can complete the multi-target search task in the shortest time.
2. The unmanned aerial vehicle platform and load integrated design method based on group intelligent search according to claim 1, wherein in step SS1, the sequential clustering algorithm comprises at least the following sub-steps:
SS1.1. Input coordinates { of the set of target pointsP tar { set point coordinates }, {P jjd ' and reference point coordinatesP 0 ；
SS1.2 initializes algorithm parameters including, but not limited to, cluster circle diameterdAnd cluster circle positioning distance deltad；
SS1.3 determination of distance reference pointP 0 The nearest pointP ini i- And take it as the firstiInitial points of cluster circles and ensureP ini i- To any existing cluster center pointCIs greater than or equal to the cluster circle positioning distance deltad；
SS1.4 at the initial pointP ini i− As a starting point, search at diameterdTarget point within range furthest from itP far-i And determine the first according to the above iCircle center of each clusterC i Coordinates of (c), wherein；
SS1.5 toC i Is the center of a circle and is based on the diameter of the cluster circledDrawing a cluster circle from a target point set {P tar Searching and recording target points falling in the cluster circle according to the distance to form a target point cluster set {P ci，j And is formed into the firstiCluster circles, whereinjIs the firstiSequence numbers of target points in the clustering circles;
SS1.6 at the firstiIndividual target point cluster set {P ci，j In } for each cluster centerC i Selecting the nearest target point as the first target pointiAggregation points within a cluster circleP jjd i, ；
SS1.7 { at target point setP tar Exclusion of the clustering set { that has been partitioned into target points }P ci，j Target points in the set, and updating the remaining set of target points;
SS1.8 repeatedly executes steps SS 1.3-SS 1.7 until the target point set {P tar And the target points are all traversed and properly distributed into the clustering circles, so that all the clustering circles and the target points and the collecting points inside the clustering circles are obtained.
3. The unmanned aerial vehicle platform and load integrated design method based on group intelligent search according to claim 2, wherein in the step SS1.3, the distance reference point is determinedP 0 The nearest pointP ini i- When the method is used, a nearest neighbor search algorithm is used, and the target point and the reference point are calculated according to the geographic coordinate information of each target point P 0 And selecting the target point corresponding to the minimum distance from the Euclidean distanceP ini i- ；
In the above step SS1.4, the diameter isdIn-range search distance initial pointP ini i- Furthest target pointP far-i When the method is used, the furthest neighbor searching algorithm is used, and the target point and the initial point are calculated according to the geographic coordinate information of the target pointP ini i- And selecting the target point corresponding to the maximum distance as the Euclidean distanceP far-i ；
In the above step SS1.5, the first search is performed and recordediIndividual target point cluster set {P ci，j When in use, a radius search algorithm is used to calculate the circle centers of the target points and the clusters according to the geographic coordinate information of the target pointsC i And choose from the Euclidean distance of (2) less than or equal todThe target point corresponding to the distance of/2 is taken as {P ci，j Elements of };
in step SS1.8, the iteration is performed until the target point set {P tar In the process that all target points are traversed, a dynamic updating algorithm is used for updating the target point set { in real time according to the geographic coordinate information of the target pointsP tar Elements of { whenever a target point is divided into a cluster circle, from {P tar Delete the target point until {P tar No element in }.
4. The unmanned aerial vehicle platform and load integrated design method based on group intelligent search according to claim 1, wherein in the step SS2, for the set of target points determined in each cluster circle, an ant colony optimization algorithm is adopted to solve the shortest track path of the unmanned aerial vehicle, so as to obtain a two-dimensional track of each unmanned aerial vehicle, and the specific flow of the ant colony optimization algorithm at least comprises the following substeps:
SS2.1. Initializing a number of artificial ants, each representing a possible track path, and randomly assigning them to respective target points within a cluster circle;
SS2.2. Each artificial ant selects the next target point to be accessed in the cluster circle according to a certain probability, the selection basis is the distance between the target points and the concentration of the pheromone, and the pheromone is released in the access process, the selection basis of the target point to be accessed is the distance between the target points and the concentration of the pheromone, the closer the distance is, the more the pheromone is, and the larger the selection probability is;
SS2.3. Each artificial ant returns to the starting point after accessing all the target points in the cluster circle, forming a complete track path, and calculating the length of the track path;
SS2.4. Updating the pheromone concentration, increasing and decreasing the pheromone on the path of each path travelled by the artificial ant according to the path length and the pheromone volatilization coefficient, and releasing more pheromones when the path is shorter;
SS2.5. Pheromones on each track path are gradually volatilized over time, so that old track path information is gradually desalted, and space is provided for new track path information;
and SS2.6. Repeating the steps SS 2.1-SS 2.5 until the preset maximum iteration times are reached or the shortest path meeting the conditions is found, and outputting the shortest track path.
5. The unmanned aerial vehicle platform and load integrated design method based on group intelligent search according to claim 1, wherein in the steps SS3 and SS6, a group task planning optimization algorithm is used to optimize a function for a targetf(d, Δd)Or (b)f(α, L gd,, θ gd ) Optimizing to obtain the optimal cluster circle diameterdAnd cluster circle positioning distanceΔdOr optimallyα、L gd And/orθ gd And the cluster task planning optimization algorithm is a particle swarm optimization algorithm PSO and/or a crow search optimization algorithm CSO.
6. The unmanned aerial vehicle platform and load integrated design method based on group intelligent search according to claim 5, wherein the specific flow of the particle swarm optimization algorithm PSO at least comprises the following sub-steps:
SSP1. Initializing a number of particles, each particle representing one possible solution, i.e. a set of cluster diametersdAnd cluster circle positioning distanceΔdValues or a set of values of (2)α、L gd And/orθ gd Randomly assigned to a solution space;
SSP2, calculating the fitness value of each particle, namely the value of a multi-objective optimization function, and determining an individual optimal solution and a global optimal solution of each particle according to the fitness value;
SSP3, updating the speed and the position of each particle according to the speed and the position of the particle, the individual optimal solution and the global optimal solution, and moving towards the direction of the optimal solution;
SSP4, repeating the steps SSP 1-SSP 3 until the preset iteration times or convergence conditions are reached, and outputting a global optimal solution.
7. The unmanned aerial vehicle platform and load integrated design method based on group intelligent search according to claim 5, wherein the specific flow of the crow search optimization algorithm CSO comprises the following sub-steps:
SSC1. Initializing a number of crow's, each crow representing a possible solution, i.e. a set of cluster diametersdAnd cluster circle positioning distanceΔdValues or a set of values of (2)α、L gd And/orθ gd Randomly assigning values to the solution space and assigning an initial position and velocity to each crow;
SSC2. Set maximum iteration timesT max And fitness thresholdθ fit And initializing a global optimal solutiong best A larger value;
SSC3, in each generation of circulation, calculating and updating the fitness value and the optimal solution of each crow, updating the speed and the position of each crow according to a certain strategy and probability, carrying out cooperation and competition behaviors, and judging whether a termination condition is reached;
SSC4. Returning to Global optimal solutiong best I.e. optimal cluster diameterdAnd cluster circle positioning distanceΔdOr the optimum value of (2)α、L gd And/orθ gd Is a value of (2).
8. The unmanned aerial vehicle platform and load integrated design method based on group intelligent search according to claim 6, wherein in the step SSC3, for each generation of cycle, the unmanned aerial vehicle platform and load integrated design method is characterized by comprising the following steps of t=1 toT max At least the following operating steps are performed:
SSC3.1 for each crow, calculate its fitness value, i.e., the value of the multi-objective optimization function, based on its current location and update its individual optimal solution based on its fitness valuep best ；
SSC3.2. If a certain crowIndividual optimal solutionp best Is superior to the global optimal solutiong best Will theng best Updated top best ；
SSC3.3 updating the velocity and position of each crow to a globally optimal solution based on inertial weights, learning factors, stochastic factors, current velocity, location of individual optimal solutions and/or globally optimal solutionsg best Or an individual optimal solutionp best Approaching;
SSC3.4. Carrying out cooperative behavior, for each crow, selecting the currently known optimal solution in the surrounding crow, namely the solution with the highest fitness value, as the reference of the search strategy, and updating the position of the solution to be the position of the optimal solution or the position near the optimal solution according to a certain probability;
SSC3.5. Performing competitive behaviors, namely selecting the currently known worst solution, namely the solution with the lowest fitness value, in surrounding crow's as the opponent of a search strategy according to a certain probability, and if a solution which is better than the worst solution, namely the solution with the higher fitness value, is found, updating the position of the solution to be the position of the better solution or the position near the better solution;
SSC3.6. Judging whether the termination condition is reached, i.e. the iteration number reaches the maximum iteration numberT max Or the fitness value of the globally optimal solution is lower than a threshold valueθ fit If so, the cycle is ended, and if not, the next generation cycle is continued.
9. The unmanned aerial vehicle platform and load integrated design method based on group intelligent search according to claim 5, wherein the group task planning optimization algorithm uses a particle swarm optimization algorithm PSO and a crow search optimization algorithm CSO to optimize multi-objective optimization functions respectively, the particle swarm optimization algorithm PSO simulates the group behaviors of shoves or fish shoves in nature, the speed and the position of each particle are updated iteratively to enable each particle to approach to a global optimal solution or an individual optimal solution, the crow search optimization algorithm CSO simulates the cooperation and competition behaviors of crow in nature, the speed and the position of each crow are updated by probability to enable each crow to approach to or depart from an optimal solution or a worst solution in surrounding crow, if the results obtained by the two algorithms are the same, the original problem is converged, otherwise iteration is continued until a termination condition is met.
10. The unmanned aerial vehicle platform and load integrated design method based on group intelligent search according to claim 1, wherein in the step SS5, a Sobol sensitivity analysis method is adopted for the unmanned aerial vehicle platform and load integrated design method α、L gd Andθ gd when the task time sensitivity analysis is carried out on the three index parameters, the implementation flow at least comprises the following substeps:
SS5.1. Will optimize the objective functionf(α, L gd,, θ gd ) Expressed as a function of a single index parameter and a combination of index parameters, namely:
wherein,X 1 =α，X 2 =L gd ，X 3 =θ gd ，f 0 is a constant term which is used to determine the degree of freedom,f i (X i ) For the first order effect term, the influence of a single index parameter on the objective function is represented,f ij (X i ,X j ) For the second order interaction effect term, the effect of interaction between two index parameters on the objective function is represented,f 123 (X 1 ,X 2 ,X 3 ) The three-order interaction effect item represents the influence of interaction among three index parameters on an objective function;
SS5.2. Calculation of the optimization objective functionf(α, L gd,, θ gd ) Is the total variance of (2)DBias of each effect termD i 、D ij AndD 123 wherein:
SS5.3. Calculating the first-order sensitivity index for each index parameterS i Second order sensitivity indexS ij And a third order sensitivity indexS 123 Wherein:
SS5.4. Calculating the Global sensitivity index for each index parameterTS i Wherein:
SS5.5. According to the first order sensitivity indexS i Global sensitivity indexTS i Judging the size of each index parameter pair to optimize the objective functionf(α, L gd,, θ gd ) Is the total variance of (2)DIs the first order sensitivity indexS i The larger the impact representing a single index parameter, the larger the global sensitivity index TS i The larger the index parameter, the greater the interaction with other index parameters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410038924.4A CN117556979B (en) | 2024-01-11 | 2024-01-11 | Unmanned plane platform and load integrated design method based on group intelligent search |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410038924.4A CN117556979B (en) | 2024-01-11 | 2024-01-11 | Unmanned plane platform and load integrated design method based on group intelligent search |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117556979A true CN117556979A (en) | 2024-02-13 |
| CN117556979B CN117556979B (en) | 2024-03-08 |
Family
ID=89818955
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410038924.4A Active CN117556979B (en) | 2024-01-11 | 2024-01-11 | Unmanned plane platform and load integrated design method based on group intelligent search |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117556979B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107562072A (en) * | 2017-10-11 | 2018-01-09 | 湖北工业大学 | A kind of unmanned plane optimum path planning method based on self-adaptive genetic operator |
| CN108549402A (en) * | 2018-03-19 | 2018-09-18 | 哈尔滨工程大学 | Unmanned aerial vehicle group method for allocating tasks based on quantum crow group hunting mechanism |
| WO2018176595A1 (en) * | 2017-03-31 | 2018-10-04 | 深圳市靖洲科技有限公司 | Unmanned bicycle path planning method based on ant colony algorithm and polar coordinate transformation |
| CN109144102A (en) * | 2018-09-19 | 2019-01-04 | 沈阳航空航天大学 | A kind of Path Planning for UAV based on improvement bat algorithm |
| CN113776532A (en) * | 2021-07-27 | 2021-12-10 | 昆明理工大学 | Unmanned aerial vehicle inspection line flight path planning method based on grouping hybrid optimization group search algorithm |
| CN115420294A (en) * | 2022-09-21 | 2022-12-02 | 江苏科技大学 | Unmanned aerial vehicle path planning method and system based on improved artificial bee colony algorithm |
| CN115657721A (en) * | 2022-10-30 | 2023-01-31 | 天翼电子商务有限公司 | Space environment unmanned aerial vehicle trajectory planning method based on improved ant colony algorithm |
-
2024
- 2024-01-11 CN CN202410038924.4A patent/CN117556979B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018176595A1 (en) * | 2017-03-31 | 2018-10-04 | 深圳市靖洲科技有限公司 | Unmanned bicycle path planning method based on ant colony algorithm and polar coordinate transformation |
| CN107562072A (en) * | 2017-10-11 | 2018-01-09 | 湖北工业大学 | A kind of unmanned plane optimum path planning method based on self-adaptive genetic operator |
| CN108549402A (en) * | 2018-03-19 | 2018-09-18 | 哈尔滨工程大学 | Unmanned aerial vehicle group method for allocating tasks based on quantum crow group hunting mechanism |
| CN109144102A (en) * | 2018-09-19 | 2019-01-04 | 沈阳航空航天大学 | A kind of Path Planning for UAV based on improvement bat algorithm |
| CN113776532A (en) * | 2021-07-27 | 2021-12-10 | 昆明理工大学 | Unmanned aerial vehicle inspection line flight path planning method based on grouping hybrid optimization group search algorithm |
| CN115420294A (en) * | 2022-09-21 | 2022-12-02 | 江苏科技大学 | Unmanned aerial vehicle path planning method and system based on improved artificial bee colony algorithm |
| CN115657721A (en) * | 2022-10-30 | 2023-01-31 | 天翼电子商务有限公司 | Space environment unmanned aerial vehicle trajectory planning method based on improved ant colony algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117556979B (en) | 2024-03-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | An adaptive clustering-based algorithm for automatic path planning of heterogeneous UAVs | |
| CN109254588B (en) | Unmanned aerial vehicle cluster cooperative reconnaissance method based on cross variation pigeon swarm optimization | |
| CN108229719B (en) | Multi-objective optimization method and device for unmanned aerial vehicle formation task allocation and flight path planning | |
| CN110608743B (en) | Multi-unmanned aerial vehicle collaborative route planning method based on multi-population chaotic grayling algorithm | |
| CN108036790B (en) | Robot path planning method and system based on ant-bee algorithm in obstacle environment | |
| CN108983823B (en) | Plant protection unmanned aerial vehicle cluster cooperative control method | |
| CN107807665B (en) | Unmanned aerial vehicle formation detection task cooperative allocation method and device | |
| US8260485B1 (en) | Adaptive multi-vehicle area coverage optimization system and method | |
| Wu | A survey on population-based meta-heuristic algorithms for motion planning of aircraft | |
| CN111609864B (en) | Multi-policeman cooperative trapping task allocation and path planning method under road network constraint | |
| CN110428111A (en) | Multi-Tasking method for planning track when UAV/UGV collaboration is long | |
| CN112230675B (en) | Unmanned aerial vehicle task allocation method considering operation environment and performance in collaborative search and rescue | |
| CN107392388A (en) | A kind of method for planning no-manned plane three-dimensional flight path using artificial fish-swarm algorithm is improved | |
| CN112469050B (en) | WSN three-dimensional coverage enhancement method based on improved wolf optimizer | |
| CN113485409B (en) | Geographic fairness-oriented unmanned aerial vehicle path planning and distribution method and system | |
| CN112733251B (en) | Collaborative flight path planning method for multiple unmanned aerial vehicles | |
| CN108985516A (en) | Indoor paths planning method based on cellular automata | |
| CN113504798A (en) | Unmanned aerial vehicle cluster cooperative target searching method imitating biological group negotiation behaviors | |
| CN109885082A (en) | The method that a kind of lower unmanned aerial vehicle flight path of task based access control driving is planned | |
| Sai et al. | A comprehensive survey on artificial intelligence for unmanned aerial vehicles | |
| CN117556979B (en) | Unmanned plane platform and load integrated design method based on group intelligent search | |
| Liu et al. | A survey, taxonomy and progress evaluation of three decades of swarm optimisation | |
| CN117434965A (en) | Unmanned aerial vehicle multi-machine collaborative natural pasture intelligent management method and management system | |
| Cuevas et al. | A swarm optimization algorithm for multimodal functions and its application in multicircle detection | |
| Li | Some problems of deployment and navigation of civilian aerial drones |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |