CN116088577B - Unmanned cluster autonomous exploration method, unmanned cluster autonomous exploration system, electronic equipment and medium - Google Patents
Unmanned cluster autonomous exploration method, unmanned cluster autonomous exploration system, electronic equipment and medium Download PDFInfo
- Publication number
- CN116088577B CN116088577B CN202310163841.3A CN202310163841A CN116088577B CN 116088577 B CN116088577 B CN 116088577B CN 202310163841 A CN202310163841 A CN 202310163841A CN 116088577 B CN116088577 B CN 116088577B
- Authority
- CN
- China
- Prior art keywords
- unmanned aerial
- aerial vehicle
- center
- front edge
- position coordinates
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 75
- 230000010261 cell growth Effects 0.000 claims abstract description 31
- 230000004927 fusion Effects 0.000 claims description 15
- 238000004590 computer program Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000011218 segmentation Effects 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 claims description 6
- 230000012010 growth Effects 0.000 claims description 4
- 238000003860 storage Methods 0.000 claims description 4
- AZFKQCNGMSSWDS-UHFFFAOYSA-N MCPA-thioethyl Chemical compound CCSC(=O)COC1=CC=C(Cl)C=C1C AZFKQCNGMSSWDS-UHFFFAOYSA-N 0.000 claims description 3
- 238000005520 cutting process Methods 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 230000005540 biological transmission Effects 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 9
- 238000004422 calculation algorithm Methods 0.000 description 7
- 238000012549 training Methods 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 4
- 238000001514 detection method Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 101100400452 Caenorhabditis elegans map-2 gene Proteins 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 101150064138 MAP1 gene Proteins 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 238000010923 batch production Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
- G05D1/104—Simultaneous control of position or course in three dimensions specially adapted for aircraft involving a plurality of aircrafts, e.g. formation flying
-
- 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)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention discloses an unmanned cluster autonomous exploration method, an unmanned cluster autonomous exploration system, electronic equipment and a medium, and relates to the technical field of multi-unmanned-plane autonomous exploration. The method comprises the following steps: inputting the occupied grid map of the current environment into a trained U-net model to obtain the position coordinates of each front edge point; clustering the front edge points to obtain center position coordinates corresponding to the front edge point categories; taking initial cells as the center, and adopting a cell growth method to grow the unexplored area to obtain a plurality of divided areas; and taking the center corresponding to the front edge point type as a starting point of the unmanned aerial vehicle path, taking the minimum length of the unmanned aerial vehicle cluster to finish autonomous exploration path as a target, and obtaining the optimal path of each unmanned aerial vehicle according to the unmanned aerial vehicle, the center corresponding to each front edge point type and the position coordinates of the center of each division area. The invention can improve planning efficiency and solve the problems of collision among unmanned aerial vehicles and non-real-time message transmission.
Description
Technical Field
The invention relates to the technical field of multi-unmanned aerial vehicle autonomous exploration, in particular to an unmanned cluster autonomous exploration method, an unmanned cluster autonomous exploration system, electronic equipment and a medium.
Background
Autonomous exploration is an important technology for unmanned aerial vehicles to navigate under unexplored areas, and plans for unmanned aerial vehicle flight by carrying out map reconstruction and searching on unexplored areas. Autonomous exploration has important application value in the fields of post-disaster rescue, scene reconstruction and the like. A large number of unmanned aerial vehicle autonomous exploration work focuses on exploration by using a single unmanned aerial vehicle, so that the exploration speed and the exploration safety are improved, and the research on collaborative exploration by using a plurality of unmanned aerial vehicles is insufficient. The unmanned clusters are used for replacing a single unmanned aerial vehicle for exploration, and the advantage is that the speed of exploration is improved through cluster cooperation, and the robustness of the system is increased through the cooperation of a plurality of unmanned aerial vehicles, so that the safety of exploration tasks is improved.
However, the serious difficulty of using cluster collaborative autonomous exploration for a cluster system is that: the efficiency of cluster planning should avoid repeated exploration of a certain area by multiple unmanned aerial vehicles as far as possible; the method has the advantages that a plurality of unmanned aerial vehicles are subjected to path planning, higher requirements are set for the path planning of the unmanned aerial vehicles during cluster cooperation, and the unmanned aerial vehicles need to avoid static obstacles and collide with other dynamic unmanned aerial vehicles during path planning; the uncertainty of communication is hard to ensure that information can be transmitted at any time when the cluster is cooperated and autonomously explored, the cooperated information is simplified, and the high efficiency of sending the information is also an important problem in cooperated exploration.
Disclosure of Invention
The invention aims to provide an unmanned cluster autonomous exploration method, an unmanned cluster autonomous exploration system, electronic equipment and a medium, which can improve planning efficiency and solve the problems of collision among unmanned planes and non-real-time message transmission.
In order to achieve the above object, the present invention provides the following solutions:
an unmanned cluster autonomous exploration method, comprising:
inputting an occupied grid map of a current environment into a trained U-net model to obtain position coordinates of each front edge point corresponding to the occupied grid map; the front edge points occupy points on junctions of unexplored areas and explored areas in the grid map;
clustering the leading edge points according to the position coordinates of the leading edge points to obtain the center position coordinates corresponding to the categories of the leading edge points;
randomly selecting a set number of points in the unexplored region as initial cells in a cell growth method;
taking the initial cells as the center, and growing the unexplored region by adopting a cell growth method to obtain a plurality of segmented regions;
taking the center corresponding to the front edge point category as a starting point of the unmanned aerial vehicle path, taking the minimum distance of unmanned aerial vehicle cluster to finish autonomous exploration as a target, and obtaining the optimal path of each unmanned aerial vehicle according to the position coordinates of each unmanned aerial vehicle, the center position coordinates corresponding to each front edge point category and the position coordinates of the center of each division area; the centers corresponding to the front edge point categories are in one-to-one correspondence with the unmanned aerial vehicle paths.
Optionally, clustering each of the leading edge points according to the position coordinates of each of the leading edge points to obtain a center position coordinate corresponding to each of the leading edge point categories, which specifically includes:
calculating the distance between the front edge points according to the positions of the front edge points;
clustering the front edge points according to the distance between the front edge points to obtain front edge point sets corresponding to the front edge point categories; the number of the leading edge points in each leading edge point set is smaller than or equal to the preset number;
and obtaining the central position coordinates corresponding to the front edge point categories according to the front edge point set corresponding to the front edge point categories.
Optionally, the growing the unexplored area by using the initial cell as a center to obtain a plurality of segmented areas specifically includes:
under the current iteration times, respectively taking all the initial cells under the current iteration times as centers, and radiating a grid length to the periphery to obtain growth cells corresponding to all the initial cells under the current iteration times;
judging whether any two growing cells meet a fusion condition or not to obtain a first judgment result;
if the first judgment result is yes, two growing cells are communicated to obtain a communicated cell under the current iteration number, and then whether an iteration stopping condition is reached or not is judged to obtain a second judgment result;
If the first judgment result is negative, judging whether an iteration stop condition is reached or not to obtain a second judgment result;
if the second judgment result is yes, determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times form a plurality of segmentation areas;
if the second judgment result is negative, determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times are initial cells under the next iteration times, and entering the next iteration.
Optionally, the center corresponding to the leading edge point type is used as a starting point of the unmanned aerial vehicle path, the unmanned aerial vehicle cluster is used as a target with minimum autonomous exploration path length, and the optimal path of each unmanned aerial vehicle is obtained according to the position coordinates of each unmanned aerial vehicle, the center position coordinates corresponding to each leading edge point type and the position coordinates of the center of each divided area, specifically:
according to the formula
Obtaining an optimal path of the unmanned aerial vehicle, wherein Plan (X) represents the length of the autonomous exploration path of the unmanned aerial vehicle cluster, min represents the minimum, i represents the center of the ith divided area, V represents the total number of the divided areas, j represents the center of the jth divided area, K represents the total number of the unmanned aerial vehicle, K represents the total number of the unmanned aerial vehicle, A i,j Representing the cost of the flight distance from the center of the i-th divided area to the center of the j-th divided area,representing the result of the kth unmanned plane from the center of the ith divided area to the center of the jth divided area, u k Representing the position coordinates of the kth unmanned aerial vehicle, < ->Representing leading edge point category corresponding to kth unmanned plane pathCorresponding center position coordinates.
An unmanned cluster autonomous exploration system, comprising:
the front edge point determining module is used for inputting an occupied grid map of the current environment into a trained U-net model to obtain position coordinates of each front edge point corresponding to the occupied grid map; the front edge points occupy points on junctions of unexplored areas and explored areas in the grid map;
the clustering module is used for clustering the leading edge points according to the position coordinates of the leading edge points to obtain the central position coordinates corresponding to the categories of the leading edge points;
an initial cell center set determining module, configured to randomly select a set number of points in the unexplored region as initial cells in a cell growth method;
the cell growth module is used for growing the unexplored area by taking the initial cells as the center and adopting a cell growth method to obtain a plurality of divided areas;
The optimal path determining module is used for taking the center corresponding to the type of the leading edge point as the starting point of the unmanned aerial vehicle path, taking the minimum length of the unmanned aerial vehicle path for completing autonomous exploration by the unmanned cluster as a target, and obtaining the optimal path of each unmanned aerial vehicle according to the position coordinates of each unmanned aerial vehicle, the center position coordinates corresponding to the type of each leading edge point and the position coordinates of the center of each divided area; the centers corresponding to the front edge point categories are in one-to-one correspondence with the unmanned aerial vehicle paths.
Optionally, the clustering module specifically includes:
a distance calculating unit for calculating the distance between the front edge points according to the positions of the front edge points;
the clustering unit is used for clustering the leading edge points according to the distance between the leading edge points to obtain leading edge point sets corresponding to the categories of the leading edge points; the number of the leading edge points in each leading edge point set is smaller than or equal to the preset number;
and the central position coordinate calculation unit is used for obtaining the central position coordinate corresponding to each front edge point category according to the front edge point set corresponding to each front edge point category.
Optionally, the cell growth module specifically includes:
the growing unit is used for radiating a grid length to the periphery by taking all the initial cells under the current iteration number as the centers under the current iteration number to obtain growing cells corresponding to all the initial cells under the current iteration number;
The fusion judging unit is used for judging whether any two growing cells meet fusion conditions or not to obtain a first judging result;
the fusion unit is used for communicating the two growing cells to obtain a communicated cell under the current iteration number if the first judgment result is yes, and judging whether an iteration stop condition is reached to obtain a second judgment result;
the iteration stop judging unit is used for judging whether an iteration stop condition is reached or not to obtain a second judging result if the first judging result is negative;
the dividing region determining unit is used for determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times form a plurality of dividing regions if the second judging result is yes;
and the iteration updating unit is used for determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times are initial cells under the next iteration times if the second judgment result is negative, and entering the next iteration.
Optionally, the optimal path determining module includes:
an optimal path determining unit for determining the optimal path according to the formula
Obtaining an optimal path of the unmanned aerial vehicle, wherein Plan (X) represents the length of the autonomous exploration path of the unmanned aerial vehicle cluster, min represents the minimum, i represents the center of the ith divided area, V represents the total number of the divided areas, j represents the center of the jth divided area, K represents the kth unmanned aerial vehicle, and K represents the total number of the unmanned aerial vehicle Number A i,j Representing the cost of the flight distance from the center of the i-th divided area to the center of the j-th divided area,representing the result of the kth unmanned plane from the center of the ith divided area to the center of the jth divided area, u k Representing the position coordinates of the kth unmanned aerial vehicle, < ->And the center position coordinates corresponding to the leading edge point category corresponding to the kth unmanned aerial vehicle path are represented.
An electronic device, comprising:
the system comprises a memory and a processor, wherein the memory is used for storing a computer program, and the processor runs the computer program to enable the electronic equipment to execute the unmanned cluster autonomous exploration method.
A computer readable storage medium storing a computer program which when executed by a processor implements the unmanned cluster autonomous discovery method described above.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
according to the invention, a neural network U-net model is used for replacing traversing and searching the front edge on the basis of a traditional single unmanned aerial vehicle front edge exploration method, so that the processing time of the process of searching the front edge point is improved, the found front edge point is clustered by using a clustering method, the front edge point is divided into a plurality of groups for carrying out path starting point distribution on different unmanned aerial vehicles in a cluster, the unexplored area is divided into a plurality of cells which are independent of each other by initializing cell positions for carrying out operations such as growth and merging, all cells realize the full coverage of the unexplored area, the on-line optimization of path planning is realized on the basis of the cell center positions, the planned path which enables the unmanned aerial vehicle cluster to efficiently and autonomously explore is found, the task amount of each individual in the cluster is ensured to be balanced, the same area is not repeatedly explored, the efficiency is maximized, the problem of the planning efficiency of the cluster is solved, the problem of the path planning of the unmanned aerial vehicles cannot be collided among the unmanned aerial vehicles is solved, the tasks of each unmanned aerial vehicle are independently executed, and the communication problem of communication uncertainty is not needed to be solved among the unmanned aerial vehicles.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings that are needed in the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of an autonomous exploring method of an unmanned cluster provided by an embodiment of the present invention;
FIG. 2 is a general flow chart of an autonomous discovery method for an unmanned cluster provided by an embodiment of the invention;
FIG. 3 is a flow chart of a stand-alone decision making section;
FIG. 4 is a schematic view of occupying a grid map leading edge point;
FIG. 5 is a block diagram of a U-net model;
FIG. 6 is a schematic diagram of clustering results of leading edge points;
FIG. 7 is a flow chart of a cell growth method;
FIG. 8 is a schematic diagram of a map for verification;
FIG. 9 is a graph of example exploration completion time versus
FIG. 10 is a training flow diagram of the U-net model;
FIG. 11 is a task map;
FIG. 12 is a graph of the results of the cell growth algorithm processing of FIG. 11;
fig. 13 is a schematic diagram of an optimal path for each unmanned aerial vehicle.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. 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.
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description.
As shown in fig. 1 and fig. 2, the embodiment of the invention provides an unmanned cluster autonomous exploration method, and the general idea is to predict leading edge points occupied in a grid map by using a neural network and divide the leading edge points into a plurality of groups by clustering; covering and dividing the unexplored area by using a cell growth method based on the clustering result, and dividing the unexplored area into a plurality of groups of independent exploration tasks; task distribution is carried out on the exploration tasks, task load of each unmanned aerial vehicle in the cluster is balanced, and the method specifically comprises the following steps:
Step 101: inputting an occupied grid map of a current environment into a trained U-net model to obtain position coordinates of each front edge point corresponding to the occupied grid map; the leading edge points are points on the junctions of the unexplored area and the explored area in the occupied grid map, the occupied grid map is shown in fig. 4, and the U-net model structure is shown in fig. 5.
Step 102: and clustering the leading edge points according to the position coordinates of the leading edge points to obtain the center position coordinates corresponding to the categories of the leading edge points, wherein the clustering result is shown in fig. 6.
Step 103: a set number of points are randomly selected within the unexplored region as the starting cells in the cell growth method.
Step 104: and growing the unexplored region by using the initial cells as a center by adopting a cell growth method to obtain a plurality of segmented regions.
Step 105: taking the center corresponding to the front edge point category as a starting point of the unmanned aerial vehicle path, taking the minimum distance of unmanned aerial vehicle cluster to finish autonomous exploration as a target, and obtaining the optimal path of each unmanned aerial vehicle according to the position coordinates of each unmanned aerial vehicle, the center position coordinates corresponding to each front edge point category and the position coordinates of the center of each division area; the centers corresponding to the leading edge point categories are in one-to-one correspondence with the unmanned paths, as shown in fig. 13.
In practical applications, the goal of the four rotors in actively exploring the unexplored area is to build a complete map of the unexplored area. The creation of a map of unexplored areas is both a process and a goal for an active exploration task, where the goal is to be selected to maximize exploration when determining the target point, and where dangerous areas in the map are to be avoided when planning the trajectory. The four rotors acquire information of the environment through detection equipment carried by the four rotors, and the unmanned aerial vehicle converts the information into an occupied grid map which can be used for decision making through a map building process.
m t ，l t ＝f map (l t ，m t-1 ，data t )，
Wherein m is t ，l t Is an occupied grid map and an estimated pose which are established at the time t by an exploration module of the unmanned aerial vehicle, and data t The information acquired by the unmanned aerial vehicle cluster through the airborne detection equipment at the moment t is information for detecting surrounding environment information (obstacles, room structures and the like) where the unmanned aerial vehicle is located and information related to the unmanned aerial vehicle, and generally comprises radar information, image information, depth information, IMU information and the like, such as positions in a space where the unmanned aerial vehicle is located and attitude angles and speeds of the unmanned aerial vehicle. f (f) map (l t ，m t-1 ，data t ) Representing the mapping process of information acquired by the drone through the probe device to the occupied grid map.
The information obtained in the unmanned cluster is the sum of information obtained by multiple unmanned aerial vehicles, and the information obtained by the cluster can be described as:
wherein the method comprises the steps ofFor the information obtained by the ith unmanned aerial vehicle, each unmanned aerial vehicle in the cluster at the same timeThe information of the machines is summarized into cluster perception information. Therefore, the autonomous exploration target of the unmanned cluster is consistent with the autonomous exploration target of a single unmanned aerial vehicle, the shannon entropy of the map can be reduced more quickly, but different from the single unmanned aerial vehicle, a plurality of unmanned aerial vehicles capable of exploration cooperatively explore the area in the cluster.
The basis of the unmanned aerial vehicle for navigation in the space is to build a map of the environment in the space, and the map can be built in advance or built in an increment mode through sensor equipment carried by the unmanned aerial vehicle. Autonomous exploration is to build an incremental map and plan the flight of an unmanned aerial vehicle in a completely unknown environment. The storage mode of the map is crucial to decision planning, various objects in the real environment exist in the form of three-dimensional geometric shapes, because the environment information acquired by the onboard equipment is continuous information, such as the position of an obstacle at the front of 1.999 meters, but continuous information processing is too complex and inconvenient to store when the algorithm is processed, so that the detected environment information is discretized through grids, such as the grid (discretization) precision is set to be 0.01 meters, then the obstacle exists in the front 199 th and 200 th grids in the algorithm, and the two grids are set to be 1. The embodiment of the invention makes decisions on task allocation and flight planning based on the rasterized map, and saves the calculation cost by dispersing the map into a two-dimensional grid form. Specific process of discretization the process of discretizing the spatial map with precision ac can be expressed as:
Wherein V is occ Is the area of the obstacle in the four points { ac x i, ac x j, ac x (i+1), ac x (j+1) } on the map plane, ac represents the rasterization accuracy, C i，j Is the probability of occupancy of the grid. The higher the accuracy of the rasterization, the closer the map is to the actual situation, but the calculation cost of planning is increased at the same time. Whether there is an obstacle in each grid in the unexplored region is unknown, all grids C in the initial state i，j ＝0.5。
In practical application, incremental exploration of unexplored regions based on leading edge points is a popular method in unmanned aerial vehicle autonomous exploration, where leading edge points may define points at the boundary between unexplored and explored regions, and explore leading edge points front in an occupied grid map i，j The grid needs to meet the following conditions:
C i，j ＝0，
wherein C' represents a grid adjacent to the leading edge point and belongs to an unexplored grid, U (C i，j 2) represents C i，j Adjacent grid sets.
For the time cost of traversing the leading edge points of each grid on the map in the large-scale occupied grid map is overlarge, as shown in fig. 10, the leading edge point search in the large-scale occupied grid map is performed by using the neural network U-net model and the deep learning method, so that the exploration leading edge points in the current map can be efficiently and quickly searched, and decision basis is provided for cluster task allocation and flight planning. As shown in FIG. 5, the neural network U-net model structure used in the method has 3 layers, 3 times of downsampling and 3 times of upsampling are respectively carried out on the occupied grid map needing to be searched for the front edge point, the map in each layer is subjected to two times of 3 x 3 convolution operations, downsampling pooling or upsampling convolution is carried out on the map between the layers, so that the size of the map feature of each layer is half or twice of the original size, the left map feature and the right map feature obtained through upsampling are combined in the same layer in a copying and cutting mode, and the combined features are subjected to convolution operation to obtain the final prediction map. Before the trained U-net model is used, training the U-net model by adopting a neural network, training a network parameter model in a training stage, inputting occupation grid maps (occupation maps in fig. 10) under different exploration states into the model, and predicting a front edge point by the network model; and meanwhile, searching the leading edge points of the occupied grid map by a traversing method, finding out the accurate position of each leading edge point meeting the requirements to obtain a leading edge point map, comparing the found accurate position with a model prediction result, calculating a prediction error, and updating and lifting the model. In the using stage, model parameters are saved, an occupied grid map (occupied map in fig. 10) acquired by the cluster in the current state is input into the model, the model outputs a map (front point map) containing front point position information, and front points required by tasks are found.
The loss function used for calculating the error in model training is a probability BCE function, specifically:
l n ＝-[y n ·log(σ(x n ))+(1-y n )·(1-log(σ(x n ))]wherein, the method comprises the steps of, wherein, x n for the prediction result of whether the nth grid is a leading edge point, l n Representing the loss error of the nth occupied grid map, σ (x n ) Representing Sigmoid activation function, y n Whether the nth grid is a true result of the leading edge point.
When N map prediction results are simultaneously used as a batch process for training, the error Loss of the model can be expressed as:
on obtaining the current network parameter model theta t Error of (θ) t ) The Adam optimizer method is then used to gradient update the network parameters. Adam optimizer is a method for adaptively optimizing network parameters, and can adjust learning rate through adaptive matrix estimation to optimize a network model. The Adam optimization method can be specifically described as:
m t ＝β 1 m t-1 +(1-β 1 )g t ，
adam optimizer adjusts by adaptationTo update the adjustment parameter update, to increase the descent speed in the gradient update.
In practical application, the clustering of the leading edge points according to the position coordinates of the leading edge points to obtain the center position coordinates corresponding to the categories of the leading edge points specifically includes:
the distance between the leading edge points is calculated from the positions of the leading edge points.
Clustering the front edge points according to the distance between the front edge points to obtain front edge point sets corresponding to the front edge point categories; the number of the leading edge points in each leading edge point set is smaller than or equal to the preset number.
And obtaining the central position coordinates corresponding to the front edge point categories according to the front edge point set corresponding to the front edge point categories.
In practical application, as shown in fig. 7, the method for growing the unexplored region by using the initial cell as a center to obtain a plurality of segmented regions specifically includes:
and under the current iteration times, respectively taking all the initial cells under the current iteration times as the centers, radiating a grid length to the periphery to obtain the corresponding growing cells of each initial cell under the current iteration times, namely, growing the radius of all the initial cells at the same time by one grid.
And judging whether any two growing cells meet the fusion condition or not, and obtaining a first judgment result. The fusion condition is whether the number of increases of the current cell is greater than the fusion coefficient times the maximum number of increases of the cell.
If the first judgment result is yes, two growing cells are communicated to obtain a communicated cell under the current iteration number, and then whether an iteration stop condition is reached is judged to obtain a second judgment result. The iteration stop condition is that the cells have completed the full population of unexplored regions.
And if the first judgment result is negative, judging whether the iteration stop condition is reached or not to obtain a second judgment result.
If the second judgment result is yes, determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times form a plurality of segmentation areas.
If the second judgment result is negative, determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times are initial cells under the next iteration times, and entering the next iteration.
In practice, obstacles or map boundaries may be touched during cell growth, in which case growth will not continue.
In practical application, the center corresponding to the type of the leading edge point is used as a starting point of the unmanned aerial vehicle path, the minimum length of the autonomous exploration path completed by the unmanned aerial vehicle cluster is used as a target, and the optimal path of each unmanned aerial vehicle is obtained according to the position coordinates of each unmanned aerial vehicle, the center position coordinates corresponding to the type of each leading edge point and the position coordinates of the center of each divided area, specifically:
according to the formula
Obtaining an optimal path of the unmanned aerial vehicle, wherein Plan (X) represents the length of the autonomous exploration path of the unmanned aerial vehicle cluster, min represents the minimum, i represents the center of the ith divided area, V represents the total number of the divided areas, j represents the center of the jth divided area, K represents the total number of the unmanned aerial vehicle, K represents the total number of the unmanned aerial vehicle, A i，j Representing the cost of the flight distance from the center of the i-th divided area to the center of the j-th divided area,representing the result of the kth unmanned plane from the center of the ith divided area to the center of the jth divided area,/th>Is a zero-one matrix, if the kth unmanned aerial vehicle arrives at the center of the jth divided area from the center of the ith divided area,/the kth unmanned aerial vehicle is>The value is 1, otherwise 0, u k Representing the position coordinates of the kth unmanned aerial vehicle, < ->And the center position coordinates corresponding to the leading edge point category corresponding to the kth unmanned aerial vehicle path are represented.
The invention also provides a more specific embodiment for describing the unmanned cluster autonomous exploration method in detail:
the flow of this embodiment may be summarized into two parts, the first part is to find the map leading edge and cluster the leading edge; and the second part is used for dividing unexplored areas and constructing and distributing tasks, forming unexplored areas into a plurality of area tasks according to a front task result, distributing the tasks, and searching a globally optimal solution according to unmanned aerial vehicle conditions in the cluster.
And constructing an occupied grid map, as shown in fig. 11, and inputting the occupied grid map into a trained U-net model to obtain the positions of the corresponding leading edge points of the occupied grid map.
After the leading edge points are obtained, clustering is carried out on the leading edge points, so that the current leading edge points to be explored are divided into a plurality of tasks to be distributed to a plurality of unmanned aerial vehicles in a cluster, each class comprises a plurality of similar leading edge points, and the tasks of the leading edge points in the same class are only divided into one unmanned aerial vehicle to be completed. The principle of cluster allocation is that two sets of front points separated by an obstacle will be divided into two tasks for allocation, and a larger set of front points will be divided into a proper number of task categories so that the task allocation is as balanced as possible.
The embodiment uses a density-based front edge point clustering method to perform front edge point task segmentation, and the specific clustering segmentation process for the set F containing m front edge points can be described as follows:
|F i |≤MaxPts，
wherein distance (x) a ，x b ) Represents x a ，x b The distance between two leading edge points, ε is a set constant, representing the maximum distance condition that divides the two points into the same class, |F i I represents set F i The number of included leading edge points, maxPts, represents the maximum limit of the number of leading edge points in the set. All the obtained leading edge point classes are used as an initial exploration task, and the initial exploration task is used as a starting task of cluster task allocation in the follow-up embodiment, namely, a path starting point, to be allocated to the unmanned aerial vehicle.
The key to cluster task allocation in the multi-unmanned cooperative autonomous exploration problem is how to transform unexplored regions into a limited number of tasks that need to be allocated. In this embodiment, the unexplored region is divided by using the cell growth method, and the division structure is as shown in fig. 12, and the tasks are divided based on the division result and the leading edge classification result, so as to perform cluster task allocation. The idea of the cell growth method can be summarized in that a part of points are randomly selected as initial centers of cells in an unexplored area, the cells in the unexplored area are continuously expanded, the expanded cells automatically fill the unexplored area, and a plurality of unexplored areas with complete segmentation and area centers are obtained. And in the process of cell growth, two smaller cells can be fused into a new cell, the fusion condition is generally set to be that the cell growth times are less than 0.5 times of the cell size limit, and the flow chart of the cell growth method. The purpose of the random selection of the initial cell locations in the unexplored region is to generate a Thiessen polygon map in the unexplored region to effectively segment the unexplored region. The cells in the unexplored region are then matched to the results of the leading edge point clusters such that each path is guaranteed to pass through a leading edge point cluster center and cells in the unexplored region.
Important parameters of the cell growth method include an initial cell number and cell size limit Maxsize, which can be determined according to the image size and the detection range of the unmanned aerial vehicle, and the coverage of the cell growth method under different map sizes is shown in table 1.
TABLE 1 cell growth method coverage efficiency Table
It can be observed that the larger the Maxsize parameter setting, the better the coverage rate of the area and the number of the areas are, but the cost is that the calculation time is increased, so that it is important to select a proper Maxsize value according to the map size and the actual detection capability of the unmanned aerial vehicle.
On the basis of area coverage and division, task areas are allocated to all unmanned aerial vehicles in a cluster, an autonomous exploration problem is converted into a vehicle route problem, and for the problem, a plurality of unmanned aerial vehicles in the cluster are assigned to complete a certain task track according to routes, and any task is reduced as much as possible on the basis of completing all tasksAnd (5) a service path. The problem can be described as: u=u 0 ，u 1 ，...，u n ，V＝v 1 ，v 2 ，...，v m Wherein U is a point set, which is a set of unmanned aerial vehicle positions in the cluster, U n Representing the spatial position coordinates of the nth unmanned aerial vehicle, V is a set of points representing a set of central positions of the segmented regions in the unexplored region, V m Representing the spatial position coordinates of the mth target area.
A＝{(v i ，v j )|v i ，v j E T, i+.j }, where A is the adjacency matrix of V, A i，j Representation (v) i ，v j ) Distance of flight costs between.
The cluster autonomous exploration allocation task can be summarized as:
wherein the method comprises the steps ofRepresents the target path starting point of the kth unmanned aerial vehicle, one unmanned plane has K paths in total corresponding to one path, and the number of paths is +>And representing the distance from the kth unmanned aerial vehicle to the starting point of the target path.
By solving an X matrix that minimizes Plan (X), X includes kAnd obtaining an optimal solution of the optimal unmanned cluster autonomous exploration planning, wherein the optimal solution is the autonomous exploration flight path of the unmanned aerial vehicle in the cluster.
Therefore, on the premise of dividing the area, the cluster task allocation can be optimized, namely an optimal solution can be found so that the total path is minimum when all unmanned aerial vehicles in the cluster finish autonomous exploration. The solution for optimizing task allocation is thus found in this embodiment using an online improvement optimization method. The process of obtaining the optimal path of the unmanned aerial vehicle according to the formula can be summarized as follows: searching the shortest path of the cluster for completing all the divided areas.
(1) According to the result of the front clustering, the task areas (all area centers and the clustering centers) are divided into a plurality of groups with the same number as the front according to the distance, and each group is considered to be all areas needed to be explored by one unmanned aerial vehicle.
(2) And calculating the exploration optimal distance Plan (X) under the current grouping, judging whether the exploration distance of all unmanned aerial vehicles in the group for completing all target areas is smaller than the result of all previous groupings under the grouping condition, and if so, storing the task allocation condition.
(3) And repeatedly selecting two groups of adjacent task areas, exchanging a certain number of adjacent nodes in the task areas, calculating whether the optimal exploration distance of the two groups of task areas after the exchange is within an acceptable range, if so, updating the task areas which are not exchanged by using the task areas after the exchange, and if not, not replacing the previous task areas.
(4) Judging whether the condition of termination is reached (the set calculation time or the set number of calculation times is reached), and if the condition of termination is not reached, repeating the steps (2) - (3) until termination.
After the regional division is completed, each unmanned aerial vehicle in the cluster will receive an instruction to go to the distribution region for single machine exploration, as shown in fig. 3, task information, namely a starting point and an end point, is firstly obtained, then a task flight path is planned, whether the exploration track approaches other unmanned aerial vehicle targets is explored, if so, cluster planning is requested to be carried out again, and if not, the searching efficiency is maximized according to the task flight path.
The invention also provides an embodiment for verifying the method, which uses a plurality of room maps with different obstacles to verify the effectiveness of the method, wherein five different room maps and the initial positions of the unmanned aerial vehicle are shown in fig. 8, fig. 8 (a) is map 1 (30 m×40 m), fig. 8 (b) is map 2 (30 m×40 m), fig. 8 (c) is map 3 (30 m×40 m), fig. 8 (d) is map 4 (60 m×80 m), fig. 8 (e) is map 5 (60 m×80 m), the obstacles in the map are represented by black, the free areas are represented by white dots, and the departure positions of the unmanned clusters are marked. The present embodiment also verifies the efficiency of exploration of unmanned clusters in a map containing different numbers of unmanned robots.
In this embodiment, the parameters used by the algorithm are shown in table 2, and the exploration degree of the shannon entropy on the map is measured by using shannon entropy, which can be described as the sum of shannon entropy occupying each grid in the grid map and can be expressed as: h (m) = Σ i，j p(C i，j )logp(C i，j ) Wherein H (m) represents shannon entropy occupying the grid map, i.e., the exploration degree, p (C) i，j ) Representing the probability of the existence of the barrier in the (i, j) th grid, the embodiment visually represents the completion degree of the cluster to the task through the change rate of the shannon entropy on the basis of the shannon entropy of the map, and measures the superiority of the algorithm through the completion degree and the completion time.
Wherein, completion represents the completion rate of evaluation autonomous exploration, H (m) N ) Is the shannon entropy, H (m) 0 ) Is the shannon entropy of the map when the exploration is not started.
Table 2 related parameter setting table
The completion of this example is shown in table 3. When the number of unmanned aerial vehicles is 1, a clustering algorithm is not used, but a front-edge exploration method is used for comparison, and the efficiency of exploration is improved along with the improvement of the number of unmanned aerial vehicles by the clustering algorithm.
Table 3 cluster search completion table
Observing the cluster search completion level can find that the present embodiment completes all searches on different maps (there is an area in map 2 that cannot be searched by the unmanned aerial vehicle, so that 100% cannot be reached). The search completion times on the different maps are shown in fig. 9. The observation image can find that with the improvement of the number of unmanned aerial vehicles, the unmanned cluster autonomous exploration method provided by the invention can help the cluster to improve the exploration completion speed, and especially on a complex and large map, the efficiency of the cluster exploration method is improved by a plurality of times compared with the efficiency of a single machine.
The invention also provides an unmanned cluster autonomous exploration system corresponding to the method, which comprises the following steps:
The front edge point determining module is used for inputting an occupied grid map of the current environment into a trained U-net model to obtain position coordinates of each front edge point corresponding to the occupied grid map; the leading edge points are points occupying the junctions of the unexplored area and the explored area in the grid map.
And the clustering module is used for clustering the leading edge points according to the position coordinates of the leading edge points to obtain the central position coordinates corresponding to the categories of the leading edge points.
And the initial cell center set determining module is used for randomly selecting a set number of points in the unexplored area as initial cells in the cell growth method.
And the cell growth module is used for growing the unexplored area by taking the initial cells as the center and adopting a cell growth method to obtain a plurality of segmented areas.
The optimal path determining module is used for taking the center corresponding to the type of the leading edge point as the starting point of the unmanned aerial vehicle path, taking the minimum length of the unmanned aerial vehicle path for completing autonomous exploration by the unmanned cluster as a target, and obtaining the optimal path of each unmanned aerial vehicle according to the position coordinates of each unmanned aerial vehicle, the center position coordinates corresponding to the type of each leading edge point and the position coordinates of the center of each divided area; the centers corresponding to the front edge point categories are in one-to-one correspondence with the unmanned aerial vehicle paths.
As an optional implementation manner, the clustering module specifically includes:
and the distance calculation unit is used for calculating the distance between the front edge points according to the positions of the front edge points.
The clustering unit is used for clustering the leading edge points according to the distance between the leading edge points to obtain leading edge point sets corresponding to the categories of the leading edge points; the number of the leading edge points in each leading edge point set is smaller than or equal to the preset number.
And the central position coordinate calculation unit is used for obtaining the central position coordinate corresponding to each front edge point category according to the front edge point set corresponding to each front edge point category.
As an alternative embodiment, the cell growth module specifically includes:
and the growing unit is used for radiating a grid length to the periphery by taking all the initial cells under the current iteration number as the centers under the current iteration number to obtain growing cells corresponding to each initial cell under the current iteration number.
And the fusion judging unit is used for judging whether the two growing cells meet the fusion condition or not for any two growing cells, so as to obtain a first judging result.
And the fusion unit is used for connecting the two growing cells to obtain a connected cell under the current iteration number if the first judgment result is yes, and judging whether an iteration stop condition is reached to obtain a second judgment result.
And the iteration stop judging unit is used for judging whether the iteration stop condition is reached or not to obtain a second judging result if the first judging result is negative.
And the segmentation area determining unit is used for determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times form a plurality of segmentation areas if the second judging result is yes.
And the iteration updating unit is used for determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times are initial cells under the next iteration times if the second judgment result is negative, and entering the next iteration.
As an optional implementation manner, the optimal path determining module includes:
an optimal path determining unit for determining the optimal path according to the formula
Obtaining an optimal path of the unmanned aerial vehicle, wherein Plan (X) represents the length of the autonomous exploration path of the unmanned aerial vehicle cluster, min represents the minimum, i represents the center of the ith divided area, V represents the total number of the divided areas, j represents the center of the jth divided area, K represents the total number of the unmanned aerial vehicle, K represents the total number of the unmanned aerial vehicle, A i，j Representing the cost of the flight distance from the center of the i-th divided area to the center of the j-th divided area, Representing the result of the kth unmanned plane from the center of the ith divided area to the center of the jth divided area, u k Representing the position coordinates of the kth unmanned aerial vehicle, < ->And the center position coordinates corresponding to the leading edge point category corresponding to the kth unmanned aerial vehicle path are represented.
The embodiment of the invention provides electronic equipment, which comprises:
the system comprises a memory and a processor, wherein the memory is used for storing a computer program, and the processor runs the computer program to enable the electronic equipment to execute the unmanned cluster autonomous exploration method.
The embodiment of the invention provides a computer readable storage medium which stores a computer program, wherein the computer program realizes the unmanned cluster autonomous exploration method when being executed by a processor.
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to assist in understanding the methods of the present invention and the core ideas thereof; also, it is within the scope of the present invention to be modified by those of ordinary skill in the art in light of the present teachings. In view of the foregoing, this description should not be construed as limiting the invention.
Claims (4)
1. An unmanned cluster autonomous exploration method is characterized by comprising the following steps:
inputting an occupied grid map of a current environment into a trained U-net model to obtain position coordinates of each front edge point corresponding to the occupied grid map; the front edge points occupy points on junctions of unexplored areas and explored areas in the grid map; the trained U-net model has 3 layers, 3 times of downsampling and 3 times of upsampling are respectively carried out on the occupied grid map needing to be subjected to front edge point searching, the map in each layer is subjected to two times of 3X 3 convolution operations, downsampling pooling or upsampling convolution is carried out on the map between the layers, so that the size of the map feature of each layer is half or twice of that of the original map feature, the left map feature and the right map feature obtained through upsampling are combined in the same layer in a copying and cutting mode, and the combined features are subjected to convolution operation to obtain a final prediction map;
clustering the leading edge points according to the position coordinates of the leading edge points to obtain the center position coordinates corresponding to the categories of the leading edge points;
randomly selecting a set number of points in the unexplored region as initial cells in a cell growth method;
Taking the initial cells as the center, and growing the unexplored region by adopting a cell growth method to obtain a plurality of segmented regions;
taking the center corresponding to the front edge point category as a starting point of the unmanned aerial vehicle path, taking the minimum distance of unmanned aerial vehicle cluster to finish autonomous exploration as a target, and obtaining the optimal path of each unmanned aerial vehicle according to the position coordinates of each unmanned aerial vehicle, the center position coordinates corresponding to each front edge point category and the position coordinates of the center of each division area; the centers corresponding to the front edge point categories correspond to the unmanned aerial vehicle paths one by one;
clustering the leading edge points according to the position coordinates of the leading edge points to obtain the center position coordinates corresponding to the categories of the leading edge points, specifically comprising:
calculating the distance between the front edge points according to the positions of the front edge points;
clustering the front edge points according to the distance between the front edge points to obtain front edge point sets corresponding to the front edge point categories; the number of the leading edge points in each leading edge point set is smaller than or equal to the preset number;
obtaining a center position coordinate corresponding to each front edge point category according to the front edge point set corresponding to each front edge point category;
the method for growing the unexplored region by using the initial cells as a center to obtain a plurality of segmented regions specifically includes:
Under the current iteration times, respectively taking all the initial cells under the current iteration times as centers, and radiating a grid length to the periphery to obtain growth cells corresponding to all the initial cells under the current iteration times;
judging whether any two growing cells meet a fusion condition or not to obtain a first judgment result;
if the first judgment result is yes, two growing cells are communicated to obtain a communicated cell under the current iteration number, and then whether an iteration stopping condition is reached or not is judged to obtain a second judgment result;
if the first judgment result is negative, judging whether an iteration stop condition is reached or not to obtain a second judgment result;
if the second judgment result is yes, determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times form a plurality of segmentation areas;
if the second judgment result is negative, determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times are initial cells under the next iteration times, and entering into the next iteration;
the center corresponding to the previous point type is used as a starting point of the unmanned aerial vehicle path, the unmanned aerial vehicle cluster is used as a target with minimum autonomous exploration path length, and the optimal path of each unmanned aerial vehicle is obtained according to the position coordinates of each unmanned aerial vehicle, the center position coordinates corresponding to each previous point type and the position coordinates of the center of each divided area, specifically:
According to the formula
Obtaining an optimal path of the unmanned aerial vehicle, wherein Plan (X) represents the length of the autonomous exploration path of the unmanned aerial vehicle cluster, min represents the minimum, i represents the center of the ith divided area, V represents the total number of the divided areas, j represents the center of the jth divided area, K represents the total number of the unmanned aerial vehicle, K represents the total number of the unmanned aerial vehicle, A i,j Representing from the center of the ith divided area to the jthThe flight distance cost of the center of each split area,representing the result of the kth unmanned plane from the center of the ith divided area to the center of the jth divided area,/th>Is a zero-one matrix, if the kth unmanned aerial vehicle reaches the center of the jth divided area from the center of the ith divided area,the value is 1, otherwise 0, u k Representing the position coordinates of the kth unmanned aerial vehicle, < ->And the center position coordinates corresponding to the leading edge point category corresponding to the kth unmanned aerial vehicle path are represented.
2. An unmanned cluster autonomous exploration system, comprising:
the front edge point determining module is used for inputting an occupied grid map of the current environment into a trained U-net model to obtain position coordinates of each front edge point corresponding to the occupied grid map; the front edge points occupy points on junctions of unexplored areas and explored areas in the grid map; the trained U-net model has 3 layers, 3 times of downsampling and 3 times of upsampling are respectively carried out on the occupied grid map needing to be subjected to front edge point searching, the map in each layer is subjected to two times of 3X 3 convolution operations, downsampling pooling or upsampling convolution is carried out on the map between the layers, so that the size of the map feature of each layer is half or twice of that of the original map feature, the left map feature and the right map feature obtained through upsampling are combined in the same layer in a copying and cutting mode, and the combined features are subjected to convolution operation to obtain a final prediction map;
The clustering module is used for clustering the leading edge points according to the position coordinates of the leading edge points to obtain the central position coordinates corresponding to the categories of the leading edge points;
an initial cell center set determining module, configured to randomly select a set number of points in the unexplored region as initial cells in a cell growth method;
the cell growth module is used for growing the unexplored area by taking the initial cells as the center and adopting a cell growth method to obtain a plurality of divided areas;
the optimal path determining module is used for taking the center corresponding to the type of the leading edge point as the starting point of the unmanned aerial vehicle path, taking the minimum length of the unmanned aerial vehicle path for completing autonomous exploration by the unmanned cluster as a target, and obtaining the optimal path of each unmanned aerial vehicle according to the position coordinates of each unmanned aerial vehicle, the center position coordinates corresponding to the type of each leading edge point and the position coordinates of the center of each divided area; the centers corresponding to the front edge point categories correspond to the unmanned aerial vehicle paths one by one;
the clustering module specifically comprises:
a distance calculating unit for calculating the distance between the front edge points according to the positions of the front edge points;
the clustering unit is used for clustering the leading edge points according to the distance between the leading edge points to obtain leading edge point sets corresponding to the categories of the leading edge points; the number of the leading edge points in each leading edge point set is smaller than or equal to the preset number;
The central position coordinate calculation unit is used for obtaining the central position coordinate corresponding to each front edge point category according to the front edge point set corresponding to each front edge point category;
the cell growth module specifically comprises:
the growing unit is used for radiating a grid length to the periphery by taking all the initial cells under the current iteration number as the centers under the current iteration number to obtain growing cells corresponding to all the initial cells under the current iteration number;
the fusion judging unit is used for judging whether any two growing cells meet fusion conditions or not to obtain a first judging result;
the fusion unit is used for communicating the two growing cells to obtain a communicated cell under the current iteration number if the first judgment result is yes, and judging whether an iteration stop condition is reached to obtain a second judgment result;
the iteration stop judging unit is used for judging whether an iteration stop condition is reached or not to obtain a second judging result if the first judging result is negative;
the dividing region determining unit is used for determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times form a plurality of dividing regions if the second judging result is yes;
The iteration updating unit is used for determining that the connected cells under the current iteration times and the unfused growing cells under the current iteration times are initial cells under the next iteration times if the second judgment result is negative, and entering the next iteration;
the optimal path determining module includes:
an optimal path determining unit for determining the optimal path according to the formula
Obtaining an optimal path of the unmanned aerial vehicle, wherein Plan (X) represents the length of the autonomous exploration path of the unmanned aerial vehicle cluster, min represents the minimum, i represents the center of the ith divided area, V represents the total number of the divided areas, j represents the center of the jth divided area, K represents the total number of the unmanned aerial vehicle, K represents the total number of the unmanned aerial vehicle, A i,j Representing the cost of the flight distance from the center of the i-th divided area to the center of the j-th divided area,representing the kth unmanned plane from the center of the ith divided area to the jth divided areaResult of dividing the center of the region, < >>Is a zero-one matrix, if the kth unmanned aerial vehicle reaches the center of the jth divided area from the center of the ith divided area,the value is 1, otherwise 0, u k Representing the position coordinates of the kth unmanned aerial vehicle, < ->And the center position coordinates corresponding to the leading edge point category corresponding to the kth unmanned aerial vehicle path are represented.
3. An electronic device, comprising:
a memory for storing a computer program, and a processor that runs the computer program to cause the electronic device to perform the unmanned cluster autonomous exploration method of claim 1.
4. A computer readable storage medium, characterized in that it stores a computer program which, when executed by a processor, implements the unmanned cluster autonomous exploration method of claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310163841.3A CN116088577B (en) | 2023-02-24 | 2023-02-24 | Unmanned cluster autonomous exploration method, unmanned cluster autonomous exploration system, electronic equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310163841.3A CN116088577B (en) | 2023-02-24 | 2023-02-24 | Unmanned cluster autonomous exploration method, unmanned cluster autonomous exploration system, electronic equipment and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116088577A CN116088577A (en) | 2023-05-09 |
| CN116088577B true CN116088577B (en) | 2023-07-18 |
Family
ID=86210267
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310163841.3A Active CN116088577B (en) | 2023-02-24 | 2023-02-24 | Unmanned cluster autonomous exploration method, unmanned cluster autonomous exploration system, electronic equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116088577B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109976164A (en) * | 2019-04-25 | 2019-07-05 | 南开大学 | A kind of energy-optimised vision covering method for planning track of multi-rotor unmanned aerial vehicle |
| CN110221614A (en) * | 2019-06-14 | 2019-09-10 | 福州大学 | A kind of multirobot map heuristic approach based on rapid discovery random tree |
| CN114384911A (en) * | 2022-01-11 | 2022-04-22 | 云南民族大学 | Multi-unmanned system collaborative autonomous exploration method and device based on boundary guide points |
| CN114859375A (en) * | 2022-04-13 | 2022-08-05 | 杭州电子科技大学 | Autonomous exploration mapping system and method based on multi-robot cooperation |
| CN115248592A (en) * | 2022-01-10 | 2022-10-28 | 齐齐哈尔大学 | Multi-robot autonomous exploration method and system based on improved rapid exploration random tree |
-
2023
- 2023-02-24 CN CN202310163841.3A patent/CN116088577B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109976164A (en) * | 2019-04-25 | 2019-07-05 | 南开大学 | A kind of energy-optimised vision covering method for planning track of multi-rotor unmanned aerial vehicle |
| CN110221614A (en) * | 2019-06-14 | 2019-09-10 | 福州大学 | A kind of multirobot map heuristic approach based on rapid discovery random tree |
| CN115248592A (en) * | 2022-01-10 | 2022-10-28 | 齐齐哈尔大学 | Multi-robot autonomous exploration method and system based on improved rapid exploration random tree |
| CN114384911A (en) * | 2022-01-11 | 2022-04-22 | 云南民族大学 | Multi-unmanned system collaborative autonomous exploration method and device based on boundary guide points |
| CN114859375A (en) * | 2022-04-13 | 2022-08-05 | 杭州电子科技大学 | Autonomous exploration mapping system and method based on multi-robot cooperation |
Non-Patent Citations (8)
| Title |
|---|
| An autonomous exploration algorithm using environment-robot interacted traversability analysis;Tang, Yujie 等;2019 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS;全文 * |
| Autonomous robotic exploration based on frontier point optimization and multistep path planning;Baofu Fan 等;IEEE Access;全文 * |
| Cooperative Search Strategy for UAVs with Frontal Image Sensor;Minqiang Zhang 等;2019 Chinese Control And Decision Conference;全文 * |
| Swarm splitting and multiple targets seeking in multi-agent dynamic systems;Zhifu Chen 等;49th IEEE Conference on Decision and Control;全文 * |
| 一种基于高效边界探索的机器人自主建图方法;刘瑞雪;曾碧;汪明慧;卢智亮;;广东工业大学学报(05);全文 * |
| 基于聚类的无人机狼群任务分配算法;赵煜民 等;人工智能;全文 * |
| 多机器人视觉同时定位与建图技术研究综述;阴贺生 等;机械工程学报;全文 * |
| 旋翼无人机环境覆盖与探索规划方法综述;张世勇 等;控制与决策;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116088577A (en) | 2023-05-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111240319B (en) | Outdoor multi-robot cooperative operation system and method thereof | |
| Chen et al. | An adaptive clustering-based algorithm for automatic path planning of heterogeneous UAVs | |
| Batinovic et al. | A multi-resolution frontier-based planner for autonomous 3D exploration | |
| CN107807665B (en) | Unmanned aerial vehicle formation detection task cooperative allocation method and device | |
| D’Amato et al. | Bi-level flight path planning of UAV formations with collision avoidance | |
| Respall et al. | Fast sampling-based next-best-view exploration algorithm for a MAV | |
| CN113342008B (en) | Path planning system and method for sea-air cooperative underwater target tracking | |
| Ni et al. | An improved real-time path planning method based on dragonfly algorithm for heterogeneous multi-robot system | |
| Ghambari et al. | An enhanced NSGA-II for multiobjective UAV path planning in urban environments | |
| Liu | A progressive motion-planning algorithm and traffic flow analysis for high-density 2D traffic | |
| Huang et al. | A novel hybrid discrete grey wolf optimizer algorithm for multi-UAV path planning | |
| CN114967744A (en) | Planning method for multi-unmanned aerial vehicle cooperative obstacle avoidance | |
| Xin et al. | Coordinated motion planning of multiple robots in multi-point dynamic aggregation task | |
| Lian et al. | Improved coding landmark-based visual sensor position measurement and planning strategy for multiwarehouse automated guided vehicle | |
| Belavadi et al. | Frontier exploration technique for 3d autonomous slam using k-means based divisive clustering | |
| CN116088577B (en) | Unmanned cluster autonomous exploration method, unmanned cluster autonomous exploration system, electronic equipment and medium | |
| CN113625733A (en) | DDPG (distributed data processing) -based multi-target three-dimensional unmanned aerial vehicle path planning method | |
| CN111427368A (en) | Improved multi-target collision-prevention driving method for unmanned intelligent vehicle | |
| Politi et al. | Path planning and landing for unmanned aerial vehicles using ai | |
| CN115755975A (en) | Multi-unmanned aerial vehicle cooperative distributed space searching and trajectory planning method and device | |
| Sabattini et al. | Hierarchical coordination strategy for multi-AGV systems based on dynamic geodesic environment partitioning | |
| Babić et al. | Heuristics pool for hyper-heuristic selection during task allocation in a heterogeneous swarm of marine robots | |
| CN114967694A (en) | Mobile robot collaborative environment exploration method | |
| Han et al. | Hybrid Algorithm-Based Full Coverage Search Approach With Multiple AUVs to Unknown Environments in Internet of Underwater Things | |
| Wang et al. | Cooperative Task Assignment of Multi-UAV in Road-network Reconnaissance Using Customized Genetic Algorithm |
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 |