CN115310792A - Task cooperation method, device and equipment for multi-target unmanned swarm - Google Patents
Task cooperation method, device and equipment for multi-target unmanned swarm Download PDFInfo
- Publication number
- CN115310792A CN115310792A CN202210891617.1A CN202210891617A CN115310792A CN 115310792 A CN115310792 A CN 115310792A CN 202210891617 A CN202210891617 A CN 202210891617A CN 115310792 A CN115310792 A CN 115310792A
- Authority
- CN
- China
- Prior art keywords
- task
- unmanned aerial
- cost
- aerial vehicle
- queue
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 78
- 238000013439 planning Methods 0.000 claims abstract description 64
- 239000011159 matrix material Substances 0.000 claims abstract description 44
- 238000004064 recycling Methods 0.000 claims abstract description 3
- 238000011084 recovery Methods 0.000 claims description 40
- 239000013598 vector Substances 0.000 claims description 35
- 238000005457 optimization Methods 0.000 claims description 16
- 230000011218 segmentation Effects 0.000 claims description 12
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000013468 resource allocation Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 4
- 230000009191 jumping Effects 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 2
- 238000000605 extraction Methods 0.000 claims 1
- 230000008569 process Effects 0.000 description 45
- 241000257303 Hymenoptera Species 0.000 description 15
- 238000010586 diagram Methods 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 7
- 238000013473 artificial intelligence Methods 0.000 description 5
- 230000006399 behavior Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000005192 partition Methods 0.000 description 5
- 230000015572 biosynthetic process Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 3
- 230000033001 locomotion Effects 0.000 description 3
- 230000008520 organization Effects 0.000 description 3
- 238000003786 synthesis reaction Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 230000007123 defense Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000002085 persistent effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- FFBHFFJDDLITSX-UHFFFAOYSA-N benzyl N-[2-hydroxy-4-(3-oxomorpholin-4-yl)phenyl]carbamate Chemical compound OC1=C(NC(=O)OCC2=CC=CC=C2)C=CC(=C1)N1CCOCC1=O FFBHFFJDDLITSX-UHFFFAOYSA-N 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000009417 prefabrication Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Images
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
- G06Q10/06—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
-
- 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]
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Development Economics (AREA)
- Biodiversity & Conservation Biology (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Marketing (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a task coordination method, device and equipment for a multi-target unmanned swarm, and relates to the technical field of unmanned aerial vehicles. The method comprises the following steps: acquiring distribution information of a target area and capability expression of each unmanned aerial vehicle in a current unmanned aerial vehicle swarm queue, wherein the capability expression comprises the following steps: standard job unit execution cost, current state, capability reservation, cost constraints and basic parameters; aiming at the distribution information, acquiring available unmanned aerial vehicles from the current unmanned aerial vehicle swarm queue according to the capacity expression; based on the distribution information and the capability expression, constructing a task cost matrix of the available unmanned aerial vehicle and target area operation, and allocating task resources of the available unmanned aerial vehicle according to the task cost matrix; and planning a task positioning path and a task recycling path for each unmanned aerial vehicle for distributing task resources. The invention realizes the optimal solution of the task planning through multi-machine overall planning.
Description
Technical Field
The invention belongs to the field of big data and artificial intelligence, and particularly relates to a task collaboration method, device and equipment for a multi-target unmanned swarm.
Background
The city is composed of a plurality of tall buildings, when disasters such as earthquake, flood, debris flow and the like occur, the buildings are damaged to different degrees, and a large number of people in the buildings are sealed and buried in the buildings. The building has certain shielding performance, and personnel buried in the building can not be rapidly identified through an intuitive means, so that when a city disaster occurs, rapid finding of the conditions of building buried points and buried personnel becomes one of the main problems in the current emergency rescue field.
Compare other detection equipment, unmanned aerial vehicle has obvious advantage in the aspect of action throughput capacity, flexibility etc. is the effective means of developing disaster area building press-buried point and pinpointing fast. At present, unmanned aerial vehicles are widely applied to the field of emergency and are limited by factors such as timely conditions and management systems, and at present, the application of unmanned aerial vehicles is developed in modes such as single-point cruise and the like. And analyzing the disaster area through aerial image data. When the area of the disaster area is larger, the efficiency of the mode is lower. The method is mainly characterized in that:
1. the existing unmanned aerial vehicle mainly collects large-range image data, has higher flying height and lacks the application capability of accurately finding out the current situation of the ground;
2. the existing unmanned aerial vehicle mainly adopts a ground control means, flight operation is issued by a task system, and flight operators manually control flight routes. A large number of targets in the disaster area are dispersed in the area, and the work efficiency is low due to the influence of human-computer cooperation, communication and other factors in the task execution process;
3. the lack of cooperation among different unmanned aerial vehicles can not form effective coverage to the region in the task organization process.
The problems have certain influence on the deep application of the unmanned aerial vehicle, and from the development point of view, the air search operation is urgently organized in a multi-machine cooperation and autonomous intelligent mode, so that the requirement of city disaster rescue in the future is met.
Disclosure of Invention
Aiming at the problems, the invention discloses a task coordination method, a device and equipment for a multi-target unmanned bee colony. The method provides a hierarchical planning architecture. In the architecture, an unmanned aerial vehicle resource capacity model is established, and unmanned aerial vehicle operation (time \ energy consumption) resources are used as a basis for planning and distribution; the planning of tasks and the execution of command scheduling need to be reasonably balanced in cost constraint and resource allowance. On the basis, the optimal solution of the task planning is realized through multi-machine overall planning. The technology has positive application value in the aspects of national defense and military industry, intelligent logistics, intelligent cities and the like.
The technical content of the invention comprises:
a task coordination method for a multi-target unmanned bee colony, the method comprising:
acquiring distribution information of a target area and capability expression of each unmanned aerial vehicle in a current unmanned aerial vehicle swarm queue, wherein the capability expression comprises the following steps: standard job unit execution cost, current state, capability reservation, cost constraints and basic parameters;
acquiring available unmanned aerial vehicles from the current unmanned aerial vehicle swarm queue according to the capability expression aiming at the distribution information;
based on the distribution information and the capability expression, constructing a task cost matrix of the available unmanned aerial vehicle and target area operation, and distributing task resources of the available unmanned aerial vehicle according to the task cost matrix;
and planning a task positioning path and a task recycling path for each unmanned aerial vehicle distributed with the task resources.
Optionally, the acquiring, for the distribution information, available drones from the current drone swarm queue according to the capability expression includes:
selecting an available unmanned aerial vehicle in a current unmanned aerial vehicle swarm queue;
for the unmanned aerial vehicle with the available state, calculating the position relation between the unmanned aerial vehicle and the target based on the current state and the distribution information;
calculating the cost of different tasks to the current unmanned aerial vehicle according to the position relation and the cost constraint;
and if the cost of any task to the current unmanned aerial vehicle is less than the capacity reservation, taking the unmanned aerial vehicle as an available unmanned aerial vehicle.
Optionally, the constructing a task cost matrix of the available drones and the target area operation based on the distribution information and the capability expression includes:
based on the distribution information and the capability expression, calculating the in-place cost, the task execution cost evaluation and the recovery cost between each available unmanned aerial vehicle and each target area operation to obtain corresponding cost T t ；
Setting a cost tolerance value of path consumption between each available unmanned aerial vehicle and each target area operation;
comparing the capability reservation with the cost tolerance value for any available drone and corresponding target area operation, and
when the difference between the capacity reservation and the cost tolerance value is greater than the cost T t In time, the cost between the available unmanned aerial vehicle and the corresponding target area operation is set as the cost T t ；
When the difference between the capacity reservation and the cost tolerance value is not greater than the cost T t Setting the cost between the available unmanned aerial vehicle and the corresponding target area operation to be infinite;
and taking the cost between the available unmanned aerial vehicle and the corresponding operation as an element in the task cost matrix.
Optionally, the allocating task resources of the available drones according to the task cost matrix includes:
clipping the task cost matrix according to the fact that the unmanned aerial vehicle cannot complete all target area operations and all unmanned aerial vehicles cannot complete operations in the target area;
processing the clipped task cost matrix by utilizing an exhaustive traversal method or an ant algorithm to obtain a plurality of scheme queues plan p Wherein p represents the row number of the clipped task cost matrix, and the row vector H in the clipped task cost matrix p Representing all cost vectors for the p-th target area job;
statistical plan queue plan p Sum of target area ranges S p And when S is p If the value is larger than the threshold value, plan queue is played p The result is regarded as a valid result;
acquiring an optimal scheme queue based on the cost of each effective result;
and allocating the task resources of the available unmanned aerial vehicle according to the optimal scheme queue plan.
Optionally, the clipped task cost matrix is processed by using an exhaustive traversal method or an ant algorithm to obtain a plurality of scheme queues plan p The method comprises the following steps:
construction scheme queue plan p And is initially empty;
from row vector H p Medium extracting element C with minimum cost pk The element C is added pk Placing in the plan queue p And deleting the row vector H p With corresponding column vector V k To obtain a matrix MC 0 Wherein k represents the column number of the clipped task cost matrix;
slave matrix MC t Row vector H of e Medium extracting element C with minimum cost ek Said element C is ek Placing in the plan queue plan p And deleting the row vector H e With corresponding column vector V k To obtain a matrix MC e+1 Wherein t representsExtraction of element C ek E denotes the matrix MC t The row number in (1);
after the clipped task cost matrix is processed, a scheme queue plan is obtained p 。
Optionally, the allocating task resources of the available drones according to the optimal solution queue plan includes:
comparing the number of the available unmanned aerial vehicles with the dimensionality of the optimal scheme queue plan;
if the number of the available unmanned aerial vehicles is not larger than the dimensionality of the optimal scheme queue plan, distributing task resources of the available unmanned aerial vehicles according to the optimal scheme queue plan;
if the number of the available unmanned aerial vehicles is not larger than the dimension of the optimal solution queue plan, optimizing the optimal solution queue plan, wherein the optimizing comprises:
setting a segmentation threshold limit;
according to the operation cost, performing sequence processing on the operation of the unallocated target area to obtain a queue reJob;
judging whether the operation cost of each target area operation in the queue reJob is smaller than the limit of a segmentation threshold value:
if yes, the optimal scheme queue plan is used as an optimization result to distribute task resources;
if not, then:
segmenting the operation of the unallocated target area with the operation cost larger than the limitation of a segmentation threshold value, returning to the distribution information, and acquiring the available unmanned aerial vehicle from the current unmanned aerial vehicle swarm queue according to the capacity expression to obtain an optimal scheme queue plan';
comparing the task coverage rate and the cost of the optimal scheme queue plan and the optimal scheme queue plan' to obtain an optimal scheme queue plan 1 ；
Queuing plan by optimization scheme t Corresponding queue reJob t Whether the operation cost of each target area operation is less than the segmentation threshold valueLimiting, judging whether to continue optimizing, wherein t is the optimizing frequency;
plan the optimal scheme queue T And performing task resource allocation as an optimization result, wherein T is the total number of optimization times.
Optionally, the planning a task in-place path for each drone that allocates task resources includes:
constructing a situation map in the form of a bitmap aiming at the whole area, wherein the bitmap consists of a plurality of basic units;
setting a task in-place path cost constraint of the unmanned aerial vehicle and a gray value threshold of the basic unit;
constructing vectors based on basic units corresponding to starting positions and target positions of unmanned aerial vehicles
Extracting the vector
All the passed basic units form a sequence according to the vector relation and are placed in a line result R in place;
comparing the cost of the in-place line result R to the task in-place path cost constraint:
if the cost of the in-position line result R is not less than the cost constraint of the task in-position path, jumping to an output unmanned aerial vehicle planning task in-position path;
if the cost of the in-place line result R is not less than the task in-place path cost constraint, then:
extracting a basic unit with the maximum gray value from the in-place line result R;
comparing the gray value of the basic unit with the maximum gray value with the gray value threshold:
if the gray value of the basic unit with the maximum gray value is smaller than the gray value threshold, skipping to judge whether the basic units in the in-place line result R are all processed;
if the gray value of the basic unit with the maximum gray value is larger than the gray value threshold value and at least one basic unit adjacent to the boundary exists, selecting the basic unit adjacent to the boundary as the basic unit with the maximum gray value, and returning to the comparison of the gray value of the basic unit with the maximum gray value with the gray value threshold value;
if the gray value of the basic unit with the maximum gray value is larger than the gray value threshold value and no basic unit adjacent to the boundary exists, selecting the basic unit with the minimum gray value to replace and place the basic unit in the in-place line result R, and skipping to judge whether the basic unit in the in-place line result R is completely processed or not;
judging whether all the basic units in the in-place line result R are processed:
if not, returning to the comparison of the cost of the in-place line result R and the task in-place path cost constraint;
if yes, jumping to an output unmanned aerial vehicle planning task in-position path;
and outputting the unmanned plane planning task in-place path.
Optionally, the planning a task recovery path for each unmanned aerial vehicle that allocates task resources includes:
optionally, the planning a task recovery path for each unmanned aerial vehicle that allocates task resources includes:
acquiring the current state and the current position of the unmanned aerial vehicle;
acquiring the task in-place cost of the unmanned aerial vehicle for executing the task in-place path, and comparing the task in-place cost with a set threshold S;
under the condition that the task in-place cost is greater than the set threshold S, selecting and judging whether the nearest recovery point meets the requirement; when the requirement is met, planning a task recovery path between the current position and the nearest recovery point; when the requirement is not met, planning a task recovery path between the current position and the original recovery point;
and planning a task recovery path between the current position and an original recovery point under the condition that the task in-place cost is not greater than the set threshold S.
A task orchestration device for a multi-target drone swarm, the device comprising:
the information acquisition module is used for acquiring the distribution information of a target area and the capability expression of each unmanned aerial vehicle in the current unmanned aerial vehicle swarm queue, wherein the capability expression comprises the following steps: standard job unit execution cost, current state, capability reservation, cost constraints and basic parameters;
the unmanned aerial vehicle screening module is used for acquiring available unmanned aerial vehicles from the current unmanned aerial vehicle swarm queue according to the capability expression aiming at the distribution information;
the resource allocation module is used for constructing a task cost matrix of the available unmanned aerial vehicle and the operation of a target area based on the distribution information and the capability expression, and allocating the task resources of the available unmanned aerial vehicle according to the task cost matrix;
and the path planning module is used for planning a task in-place path and a task recovery path for each unmanned aerial vehicle which is allocated with the task resources.
A computer device comprising a memory and a processor, the memory having stored therein a computer program, the computer program being loaded and executed by the processor to implement the above-described task collaborative method of a multi-objective unmanned bee colony.
Compared with the prior art, the invention has the following advantages and effects
According to the invention, a task command and resource scheduling framework with a closed loop is constructed on the basis of an unmanned aerial vehicle cluster, and all-region and all-period coverage is realized for targets in urban disaster areas; realizing automatic task adaptation of an unmanned aerial vehicle (bee) on the basis of target capability adaptation; and a complete task guidance scheme is formed by combining the actual environment of the disaster area. The search working efficiency is improved in a multi-level \ multi-batch \ multi-subject cooperative mode.
Drawings
FIG. 1 is a block diagram of the overall organization of tasks.
Fig. 2 task phase partition diagram.
Fig. 3 is a diagram of unmanned bee colony task execution target allocation.
FIG. 4 is a flow diagram of an object allocation process.
Fig. 5 is a schematic diagram of unmanned aerial vehicle (bee) task in-place guidance.
Fig. 6 unmanned aerial vehicle (bee) task in-place route planning diagram.
Fig. 7 is a diagram of unmanned aerial vehicle (bee) task recovery routing.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is to be understood that the described embodiments are merely specific embodiments of the present invention, rather than all embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments of the present invention, belong to the protection scope of the present invention.
The multi-target unmanned bee colony task cooperation method is suitable for various use scenes such as urban disaster rescue, national defense and military industry, smart logistics, smart cities and the like. The technical contents of the present invention will be described below by taking city disaster relief as an example.
The invention can develop unmanned bee colony task planning and collaborative key technology research work aiming at the characteristics of building objects in urban disaster rescue. Combining emergency rescue professional knowledge, constructing a task scheduling framework, and realizing unified command scheduling of multi-machine cooperative tasks; constructing a colony capacity unified expression model, realizing unified measurement of colony capacity states, and supporting quick synthesis of search power; a task demand analysis and decision evaluation model is established by combining methods such as artificial intelligence and the like, and an optimal solution of task decision is realized on the basis of multi-point cooperation through a multi-target reconnaissance task discretization processing algorithm and a resource matching optimization algorithm which are oriented to a complex space body; by a path planning algorithm based on complex indoor and outdoor environments, under the constraint of cost, task paths of a single machine (bee) are planned according to factors such as environment, risk and the like, and the requirement of multi-machine cooperation is met.
After destructive disasters such as earthquake, flood, debris flow and the like occur, the on-site exploration of the disaster area is a continuous operation implementation process. The method mainly comprises the steps of task aggregation, task planning, task instruction transmission, task in-place, operation implementation, task recovery and the like. From the perspective of the entire task hierarchy. The process includes three basic levels, as shown in fig. 1, the whole technical system is composed of three basic levels, namely a planning level, a task level and an operation level, wherein:
planning a layer: on the basis of basic information, the task state information of the unmanned bee colony is gathered by relying on a disaster area field communication network to form a state and capability portrait. Meanwhile, various summary search data and third-party information data are gathered on the basis of a disaster area unmanned aerial vehicle search and detection system, and the cognition of the disaster area summary situation is formed. On the basis, according to the command instruction, the targets in the disaster area environment are identified and the task planning is carried out. In the planning process, matching of target distribution and swarm resources is carried out according to target distribution, swarm capacity and states, and an unmanned aerial vehicle searching force synthesis scheme is formed. And combining an artificial intelligence algorithm, under the condition of cost constraint, realizing the optimal solution of the force synthesis scheme through overall planning, and meeting the requirements on the efficiency and the efficiency of implementing the search task in the disaster area.
And (3) task layer: and in the process of arranging the search task scheme, path planning is carried out on unmanned aerial vehicles (bees) participating in the task. In the process, taking the task time cost as a total constraint, enumerating paths in a task in-place link, forming optimal path selection according to task requirements, and providing a basis for the task execution scheduling of the unmanned aerial vehicle; meanwhile, in the task recovery stage, according to condition constraints such as communication and organism states, the withdrawing route of the unmanned aerial vehicle is rapidly planned, and the requirement of unmanned aerial vehicle (bee) task command and scheduling is met.
An operation layer: and the unmanned aerial vehicle carries out reconnaissance operation after reaching the designated operation area according to the task planning. In the process, the unmanned aerial vehicle completes data acquisition of the building space by adopting a mode of combining autonomy, man-machine cooperation and multi-machine cooperation according to the working environment and executes operation recovery according to task planning.
The following points are followed in the planning:
1. on the basis of a given task target, the constraint of the number of required completions of one wave is met;
2. on the basis of meeting 1, the overall cost of the formed scheme is minimum;
3. on the basis of meeting 1,2, the overall risk of the formed scheme unmanned aerial vehicle (bee) group is minimum.
The reconnaissance task of the building space is formed by combining a plurality of layers. Due to the complexity of the interior of a building environment and the surrounding areas of the building, the unmanned aerial vehicle (bee) needs to have strong scene adaptability in the task execution process. The planning and the execution of command scheduling of tasks need to be reasonably balanced in cost constraint and resource allowance. On the basis, the optimal solution of the task planning is realized through multi-machine overall planning. Aiming at the technical requirement, a planning layer and a task layer are taken as main parts, and on the basis of multi-target resource optimization and matching, the invention provides a multi-target unmanned swarm task collaborative key technology for urban disaster rescue, which mainly comprises the following steps:
1. unmanned bee colony task execution framework and task model construction
As previously mentioned, reconnaissance for a building space is a continuous, progressive process, consisting of multiple waves. As shown in fig. 2.
The following constraints are provided for the task wave times:
1. the scene comprises m target areas and n unmanned aerial vehicles;
2. in each wave, one unmanned aerial vehicle can only carry out operation on one target area;
3. in one wave, one target area can only be operated by one unmanned aerial vehicle in real time;
4. if the target area is large, the target area may be broken down into several partitions, each of which forms an independent target area, the target area being constrained as before.
As shown in fig. 2, from the perspective of the process, a task wave may be composed of the following links:
1. the task is in place: after the task is determined, the unmanned aerial vehicle (bee) receives a task instruction and moves from a starting point to an approach point of operation implementation. In this process, the drone passes through the airThe motorized mode completes the route movement. In the task commanding process at this stage, the unmanned aerial vehicle maneuvering time cost, the environmental risk cost and the like are mainly considered, and the path is reasonably planned on the basis of the total task total cost constraint. In this stage, two main reference variables are formed: t is a unit of 1 And D 1 Wherein: t is a unit of 1 The total time consumption in the process of positioning (cruising) of the unmanned aerial vehicle task; d 1 The total distance length of the unmanned aerial vehicle in the task positioning process is obtained.
2. Field operation: after the unmanned aerial vehicle (bee) flexibly reaches a designated operation position through the air, the load is utilized to carry out field operation, and the building space is detected and data is acquired. In the process, task basic time cost calculation is formed according to the area of the operation area and the operation capacity of the unmanned aerial vehicle unit:
wherein:
a: collecting each target/partition of the current operation area;
ai: is an independent target \ partition;
dt: the time cost to complete a unit target \ partition operation for the drone (bee);
T A : the total time cost for unmanned aerial vehicles (bees) to complete a designated work area;
d: unmanned aerial vehicle (bee) basic capability subject;
s: the operation name, which is used to calculate the time for a single drone to complete a zone operation, may be simply considered as S = area/d.
3. Task recovery: after the operation of the unmanned aerial vehicle (bee) is finished, the unmanned aerial vehicle (bee) moves from the off-site point to the recovery point according to the mission planning. In the process, the unmanned aerial vehicle completes the route movement in an air maneuvering mode. In the task command process at this stage, a recovery path is prefabricated according to task planning at the early stage; after the unmanned aerial vehicle finishes the operation, the prefabricated paths are matched according to the states of the unmanned aerial vehicle. If the current state satisfies the cost constraint of the prefabricated reclamation pathCompleting recovery according to the prefabrication plan; and if the current state does not meet the cost constraint, automatically planning a new recovery scheme and a new path according to the conditions, and moving to a corresponding position. The unmanned aerial vehicle maneuvering time cost, the environmental risk cost and the like are mainly considered, and the path is reasonably planned on the basis of total task total cost constraint. In this stage, two main reference variables are formed: t is 2 And T 2 ', wherein: t is 2 The total time consumption in the unmanned aerial vehicle task recovery process is achieved; t is 2 ' resource reservation for unmanned aerial vehicle in task recovery process. The following procedure is formed according to the foregoing:
wherein:
retrieve is the name of a reclaim operation, which can be understood as a call function;
cap is the name of the capability obtaining operation, and examples are as follows: after the operation is finished, the d' residual capacity of the unmanned aerial vehicle is obtained
d' is the expression of unmanned aerial vehicle d after finishing the operation.
According to the organization situation of the above task process, a basic task model is formed, which is as follows:
Task＝{job i |i＝1,2,....p}
job＝{ID,location,jobScale,{risk j |j＝1,2,....q|}}
risk＝{ID,prop,risklevel}
wherein:
task, which is the overall composition of the Task and is composed of a group of job jobs;
job: is basically defined for the job. The detailed definition is as follows:
ID: a globally unique identification for the job;
and (3) location: is the location of the job;
jobScale: estimating the total cost for the range of the operation area according to the comparison of the jobScale and the operation cost of the unmanned aerial vehicle (bee) unit;
risk: the risks screened in the working area are composed of identifications, attribute sets and risk levels. The whole risk of one operation area is obtained by identifying and converging all risks;
risklevel: is a risk classification.
2. Unified expression model for unmanned bee colony capacity
Unmanned aerial vehicles (bees) are the subject of task execution. In the process of task planning, unmanned aerial vehicles (bees) are allocated according to the characteristics of the target area, and resource matching and force composition of tasks are achieved. During the task execution, the drone (bee) implements maneuvers and operations according to the planned action plan. Meanwhile, the behavior is corrected on the basis of autonomous decision making by combining with the actual influence of the scene. It can be seen from the above links that the capability expression and measurement of the unmanned aerial vehicle (bee) become the basis of the whole task command and scheduling.
According to the situation, the unmanned aerial vehicle capability unified expression model is constructed. The basic definition of this model is as follows:
d = { ID, stdCap, curState, resKpt, capRuleSet, belongTo, prop }, where:
ID: globally unique identification for the unmanned aerial vehicle;
stdCap: a cost corresponding to a standard unit of work to perform a process;
curState: the current state of the unmanned aerial vehicle; it is defined as follows:
curState = { preState, nowState, location }, where:
preState: expressed for the previous state;
nowState: is expressed for the current state;
and (3) location: the current spatial position is obtained;
resKpt: the task execution capacity is reserved for the unmanned aerial vehicle at the current moment;
capRuleSet: a capability cost consumption rule set for the drone to execute tasks under different conditions;
capRuleSet＝{(cID,condDef,cost) i i =1,2, · n }, wherein:
cID: identifying a current rule;
condDef: constraint conditions corresponding to the current rules;
cost: cost consumption corresponding to the standard behavior unit under the current constraint condition;
belongTo: information of the belonged bee (cluster);
prop: basic parameters for unmanned aerial vehicles (bees), including cruise speed normal speed, etc
The unified expression of the basic state information and the capability of the unmanned aerial vehicle is realized through the content. In the task execution process, after the unmanned aerial vehicle (bee) is allocated with a task operation, the state of the unmanned aerial vehicle (bee) changes to a certain extent, namely:
jobSet(job，D)＝{job,D'},D'＝stateChange(D,s')
wherein,
jobSet: and (4) unmanned aerial vehicle task allocation operation, namely binding the unmanned aerial vehicle with a task after the task state is changed with the task.
stateChange: and (5) unmanned aerial vehicle task state changing operation.
3. Regional multi-objective task matching and resource optimization
In the task planning process, firstly, targets and unmanned aerial vehicles (bees) are distributed according to the distribution condition of building objects and the distribution condition of unmanned aerial vehicles in the regional range. In the process, according to the factors such as the size of a working area, the reserved task execution capacity (resKpt) of the unmanned aerial vehicle (bee), the influence of the surrounding environment and the like, the cost consumption of the unmanned aerial vehicle (bee) for completing the task in different distribution schemes is calculated, the overall optimal solution is realized by an ant algorithm, and finally a target distribution and unmanned aerial vehicle task arrangement scheme is formed. As shown in fig. 3:
step1: constructing a task set:
in the process, firstly, all target objects in the area range are obtained according to the fusion of different source data to form a task set:
Task＝{job i i =1,2,.. N }, and the Task set Task is composed of all the targets to be observed. Its detailed definition please see the above.
Step2: constructing an unmanned aerial vehicle (bee) resource candidate set:
on the basis of the task set, all available unmanned aerial vehicles (bees) are acquired according to state and cost constraints, and a candidate resource set Ds is formed:
Ds＝{D j |j＝1,2,.....m},D j ∈AllD
curState∈D j ,nowState∈curState
wherein:
AllD is a set of all unmanned aerial vehicle objects;
wp is a weight value;
c' is the actual cost consumption of the current task;
nowState is a component in curState;
check is an unmanned aerial vehicle value selection calculation function in the cost matrix construction process;
easy is a specific value of nowState.
Namely: first selecting different drone (bee), state is available (nowState = easy);
and then, calculating the cost of different tasks to the current unmanned aerial vehicle (bee) according to the position relation and the cost constraint between different unmanned aerial vehicles (bee) and the target, and if the cost of the current unmanned aerial vehicle (bee) and the task is less than the self capacity reservation, regarding the current unmanned aerial vehicle (bee) and the task as the object to be selected and placing the object in the Ds set.
Step3: calculating the cost of the target task:
after the selection of the candidate set is completed, according to the unmanned aerial vehicle (bee) (D) i ，D i E.g. Ds) state, with different task objectives (job) j ，job j Epsilon to Task) to form a Task cost matrix MC:
cost is a Cost calculation function, and the calculation method is as follows:
T t ＝T 1 '+T A +T 2 '，
T A ＝jobCostAss(job,D)
T 1 '＝T 2 '＝shortcut(job,D)/normalSpeed，normalSpeed∈prop，prop∈D
wherein, T 1 ' reserving resources for the unmanned aerial vehicle in the positioning process, wherein short is a cost evaluation function of the linear distance between the unmanned aerial vehicle and a target two points, jobcostas is a task execution cost evaluation function, and the cost (such as time/energy consumption) for completing the operation of a specified task area by the current unmanned aerial vehicle (bee) is comprehensively evaluated and calculated according to the operation time of the unmanned aerial vehicle (bee) unit, the range of the task area, the environmental influence and the like.
In the above calculation, T offset Is a cost tolerant value. When the value is 0, the cost tolerance of path consumption is 0 when the unmanned aerial vehicle (bee) is in place and in the task recovery process. The unmanned aerial vehicle can only maneuver in real time in a mode that an endpoint flies directly. In urban environment, due to the existence of a large number of factors such as high-rise building shielding, microenvironment meteorological influence and the like, the value is obviously limited; when this value is scaled up, it means that the path cost tolerance is increased, and a tolerance that is too scaled up may result in poor resource availability. In practical application, a training model is constructed in combination with artificial intelligence, and optimal setting of the value is achieved.
Step4: and (3) task resource allocation:
and the target task cost matrix MC expresses the time distribution condition for all available unmanned aerial vehicles to complete the tasks in the current area in the current state.
In which matrix the column vector V k Representing all operation cost vectors of target areas of the kth unmanned aerial vehicle;
V k ＝[cost(job 1 ,D k ),.........,cost(job m ,D k )] T
cost(job j ,D k )∈MC,j≤m
in the matrix the row vector H p Representing all cost vectors of the p-th target task job;
H p ＝[cost(job p ,D 1 ),........,cost(job p ,D n )] T
cost(job p ,D j )∈MC,j≤n
in the MC, the operation cannot be completed due to the reason that the operation cost is too high in some target areas, and the like, and the target cost unit is ∞ in some unmanned aerial vehicles (bees). Therefore, the MC needs to be tailored for subsequent processing. This process is based on the following rules:
1. when an unmanned plane (bee) D k When all targets cannot be completed, deleting the column vector V corresponding to the unmanned aerial vehicle (bee) from the MC k ；
2. A target region joba p No unmanned aerial vehicle (bee) can complete the operation aiming at the unmanned aerial vehicle (bee) in the current wave number, and the row vector H corresponding to the target is deleted from the MC P ；
ValidateMc(MC)＝MC”,MC'∈MC,V k ∈MC',H p ∈MC'
Where ValidateMc is an MC matrix invalid vector processing operation.
The goal of task allocation is to achieve maximum target coverage at minimum cost from an overall perspective. Meanwhile, due to the reasons of position relation and the like, the respective costs of the participated unmanned aerial vehicles (bees) are different in the process of executing one task batch. On the basis, different task schemes are formed through the cost scheme traversal of the target. These mission schemes combine mission objectives with specific drones (bees). And in the process:
1. setting i =1, let MC "= MC';
2. construct plan queue plan i And is initially empty;
3. starting from MC "line i, get H i And (5) vector quantity. From H i Extracting C with minimum cost ik ；
getMinCost(H i )＝C ik ,C ik ＝cost(job i ,D k )
C ik ＜C jk ,i≠j,C ik ∈H i ,C jk ∈H i
4. C is to be ik Put in the current plan queue, remove the ith row vector H from MC ″ i And the kth column vector V k ；
5. When the current MC' exceeds 1 dimension, i =2, executing step 2;
6. placing the rest cost (jobD) of the current MC' as a unit in the plan i Performing the following steps;
7. let i = i +1, if the current i exceeds the row vector number of MC', then execute step 8, otherwise execute step 2;
8. queue formation scheme plan i Counting the sum S of all target area ranges in the current plan i If S is i Greater than threshold
Put, then the current plan is put i If the result is a valid result, otherwise, the result is taken as an invalid result;
plan＝{C ij，k |k＝1,2,....q},C ik ∈MC'，
the above is a solution for traversing MC' from a target area as a starting point with the least cost.
When the target number m and the number n of unmanned aerial vehicles (bees) in the disaster area environment are smaller, the dimensionality of MC' is not high, all the planes can be solved in an exhaustive traversal mode, and the planes with the lowest cost are selected as results to be output under the condition of preferential income;
when the number m of targets and the number n of unmanned aerial vehicles (bees) in the disaster area environment are larger, the dimensionality of the MC' is higher, and the performance of the method of exhaustive traversal is limited. The MC' can be processed by adopting an ant algorithm to form a quick optimal solution, and a result is finally output.
The whole process is shown in fig. 4.
Step5: re-optimization of the task scheme:
during the aforementioned process, some target areas may be abandoned due to excessive cost (exceeding the drone (bee) capability reservation boundary) or too far in the way. For this situation, on the basis of the task resource allocation, a larger area or a farther area is divided into smaller target areas by an area re-division method, resource capability adaptation is performed, and iteration of a specification scheme is formed. This process is as follows:
1. let plan T ＝plan；
2. Whether the current scheme needs to be optimized continuously or not is judged,
number n of all available drones if in the current plan<Plan output plan T The vector dimension indicates that all unmanned aerial vehicles have task allocation, allocation adjustment is not needed, and the current operation is exited to output the optimal result plan T 。
When n is greater than plan T Indicates that there are available drones to arm. Further allocation optimization is performed.
This process is as follows:
3. performing sequence processing on the unallocated target object jobk, and forming a queue reJob according to the size of the operation cost;
jobScale k ＞jobScale p ,k＜p＜l,jobScale k ∈job k ,jobScale p ∈job p
4. setting k =1, and letting T = Task ', where Task' represents an unallocated target object set;
5. extracting the kth job from reJob k Object, if jobScale k <The current jobs is exited after the limitation of the segmentation threshold value k Processing and outputting plan T ；
6. For joba k Performing segmentation to form a new target region object, and extracting the jobs in the T k Replacing the target object with a new target object to form a new task set T';
7. taking T 'as a new task set, executing step1 to obtain a corresponding output result plan';
8. and comparing the task coverage rates of the plan and the plan' and selecting the optimal one. If the benefits are consistent, the plan' with the least cost of plan is selected as the resulting plan T ；
9. Judging whether to stop continuing optimization at present, if not, executing 1, otherwise, stopping outputting the final result;
in the above process, the segmentation of the target region in step 6 has a certain influence on the calculation result. The specific segmentation method can realize operation on the basis of artificial intelligence, and improves the segmentation efficiency and the use efficiency of the unmanned aerial vehicle (bee). The specific method is not described herein.
4. Unmanned aerial vehicle (bee) task path planning
As mentioned above, after the task planning of the unmanned aerial vehicle (bee) is completed, the related information is sent to the command unit. And according to the relation of the command service link, the task scheme is issued to the corresponding unmanned aerial vehicle (bee) to start to implement. In the whole unmanned aerial vehicle (bee) operation execution process, the operation execution process is divided into three links according to the action purpose: task in-place, job execution and task recovery. In combination with the above, the project mainly aims at path planning and task matching of the task in-place part and the task recovery part to carry out research work.
1. Task in-place path planning
Task in-place refers to the process of maneuvering of unmanned aerial vehicles (bees) from a departure point to an approach point. Typically, drones (bees) cost the least in an end-to-end linear motion. Meanwhile, the command and dispatching of the unmanned aerial vehicle (bee) are facilitated. However, the architectural structure and space conditions in urban environments are complex. On one hand, tall buildings form a shelter for a moving route; on the other hand, certain microclimate environment can be formed among specific buildings, and great influence is caused on the flight safety of the unmanned aerial vehicle (bee) which passes through the buildings. Both of these conditions pose a risk to the flight manoeuvres of the drone (bee). Therefore, in the maneuvering process, reasonable avoidance needs to be performed to ensure smooth implementation of the task.
The evasive behavior of the drone (bee) also causes additional cost to be paid while improving the safety of the drone itself. In the foregoing, the cost T of the task-in-place process 1 To the total cost T t There is a direct impact. If the problem is ignored in the maneuvering process, the unmanned plane (bee) cannot normally execute tasks and work, and planning fails. Therefore, when the drone (bee) is in place, the path needs to be reasonably selected under the constraint of the overall mission planning cost planning, as shown in fig. 5.
As mentioned above, an drone (bee) mission path plan consists of drone (bee) behavior cost constraints and objectives:
routePlanReq＝{D i ,p1,p2，T 1i },D i ∈Ds,D i ∈plan T wherein:
D i : unmanned aerial vehicle (bee) to be path planned;
p1: drone (bee) home location;
p2: unmanned aerial vehicle (bee) target position (operation entry point);
T 1i : and (4) setting the cost constraint of the in-place path planning behavior of the unmanned aerial vehicle (bee).
The disaster area environment has high dynamics. Therefore, unmanned aerial vehicles (bees) gather various risks in disaster area environment according to multi-source disaster information, and situation perception is formed on the basis of unified labeling. By means of the situation map, the unmanned aerial vehicle (bee) carries out task path planning. For the convenience of computational processing, this situation diagram is reduced to the form of a bitmap, which is basically defined as follows:
instEnvMap＝{px ij |i＝1,2,...n,j＝1,2,....,m}
px ij ＝{center,riskGreyTag}
px ij : the whole bitmap is rasterized according to a fixed side length, which is one basic unit (pixel) of the bitmap. One px ij A corresponding one of the basic grids.
center: the central position of the grid is;
riskGrayTag: and (5) marking risks. Px in the bitmap corresponds to a grid in the map. And in the map, mapping the risk objects in the area corresponding to the grid according to the levels. The risks of different levels are mapped to a corresponding gray value. And assigning the gray value corresponding to px in the bitmap. Here, the risk refers to a tall building affecting the flight safety of the unmanned aerial vehicle (bee), a slightly meteorological environment badly caused by a complex space, and the like. These risks are distributed in the map according to the actual position and mapped in a grey scale map.
And (4) rapidly planning the route of the unmanned aerial vehicle (bee) based on the map. This process is as follows:
1. setting t =1,p = p 1 Route result R;
2. calculating p 1 And p 2 The corresponding px:
locate({p 1 ,p 2 })＝{px 1 ,px 2 }，p 1 ∈px 1 ,p 2 ∈px 2
3. constructing a vector r, namely the current p and p2:
4. starting from p, will
The entire px that passed through was extracted and the resulting sequence was deposited in R.
5. Calculating the total cost C of the current R R ，
If C is present R Not less than T 1i If yes, outputting the current R;
if C is present R Less than T 1i If yes, executing step 6;
6. extracting px with the maximum gray level from R:
getMaxFromR(R)＝px k ,px k ∈R,px j ∈R,
riskGrayTag k ＞riskGrayTag j ,riskGrayTag k ∈px k ,riskGrayTag j ∈px j
wherein getmaxfrom: operation of obtaining the maximum unit of the gray value from R, riskgrytag: cell gray value.
7. Let p' = px k ；
8. Judging the feasibility of p':
if the gray value of the current p' is less than the threshold setting, then the current px k If yes, executing step 9;
if the peripheral px where the current p' is located is already calculated, selecting the px with the minimum gray level to replace the px and placing the px in the R queue, and executing the step 9;
px if the gray value of the current p' is greater than the threshold setting k For impassability, select px with adjacent boundary pq Let us order
p’＝px pq Executing 8;
9. whether all the basic units stored in R are processed is judged, if the step5 is not finished, the step 10 is executed;
10. and completing the calculation of the current route, extracting all px sequences from R, forming a route Rline and outputting.
The basic process is shown in fig. 6:
when a route is planned, the cost of the route is larger than T under certain conditions 1i Passing the cost tolerance value T in the foregoing offset A certain reservation is achieved.
Meanwhile, according to the gray levels of all px in R, the risk distribution and degree on the route can be extracted and calculated. Providing a basis for task evaluation:
riskMax(R)＝riskGrayTag k ,riskGrayTag k ∈px k ,px k ∈R
and according to the two indexes, an evaluation analysis model is constructed, the evaluation of the task risk is realized through simulation training, and support is provided for task decision.
2. Task recovery path planning
After the unmanned aerial vehicle (bee) finishes the operation, the unmanned aerial vehicle (bee) sets to reach the departure point according to the task and starts to execute task recovery. In the actual task execution process, the unmanned aerial vehicle (bee) does not necessarily have the cruising ability of returning to the initial starting point due to reasons such as the excess of the early-stage task and the like. Therefore, in this process, it is necessary to select a recovery scheme according to the self-state and the field situation, as shown in fig. 7. Detecting and acquiring the resources reserved by the unmanned aerial vehicle (bee). According to persisting the resource and choosing the recovery point, guarantee that unmanned aerial vehicle (bee)'s persisting resource can support its recovery point of flying back. On the basis, a flight path is planned according to the current position and the recovery point, and a guiding scheme is formed.
The above embodiments are only intended to illustrate the technical solution of the present invention, but not to limit the same, and a person skilled in the art may modify the technical solution of the present invention or substitute the same, and the protection scope of the present invention shall be subject to the claims.
Claims (10)
1. A task coordination method for a multi-target unmanned swarm is characterized by comprising the following steps:
acquiring distribution information of a target area and capability expression of each unmanned aerial vehicle in a current unmanned aerial vehicle swarm queue, wherein the capability expression comprises the following steps: standard job unit execution cost, current state, capability reservation, cost constraints and basic parameters;
acquiring available unmanned aerial vehicles from the current unmanned aerial vehicle swarm queue according to the capability expression aiming at the distribution information;
based on the distribution information and the capability expression, constructing a task cost matrix of the available unmanned aerial vehicle and target area operation, and distributing task resources of the available unmanned aerial vehicle according to the task cost matrix;
and planning a task positioning path and a task recycling path for each unmanned aerial vehicle for distributing task resources.
2. The method of claim 1, wherein said obtaining available drones from a current drone swarm queue according to the capability expression for the distribution information comprises:
selecting an available unmanned aerial vehicle in a current unmanned aerial vehicle swarm queue;
for the unmanned aerial vehicle with the available state, calculating the position relation between the unmanned aerial vehicle and the target based on the current state and the distribution information;
calculating the cost of different tasks to the current unmanned aerial vehicle according to the position relation and the cost constraint;
and if the cost of any task to the current unmanned aerial vehicle is less than the capacity reservation, taking the unmanned aerial vehicle as an available unmanned aerial vehicle.
3. The method of claim 1, wherein said constructing a mission cost matrix of said available drones for operation with a target area based on said distribution information and said capability representation comprises:
based on the distribution information and the capability expression, calculating an in-place cost, a task execution cost evaluation and a recovery cost between each available unmanned aerial vehicle and each target area operationTo obtain a corresponding cost T t ；
Setting a cost tolerance value of path consumption between each available unmanned aerial vehicle and each target area operation;
comparing the capability reservation with the cost tolerance value for any available drone and corresponding target area operation, and
when the difference between the capacity reservation and the cost tolerance value is greater than the cost T t In time, the cost between the available unmanned aerial vehicle and the corresponding target area operation is set as the cost T t ；
When the difference between the capacity reservation and the cost tolerance value is not greater than the cost T t Setting the cost between the available unmanned aerial vehicle and the corresponding target area operation to be infinite;
and taking the cost between the available unmanned aerial vehicle and the corresponding operation as an element in the task cost matrix.
4. The method of claim 1, wherein said allocating task resources of the available drones according to the task cost matrix comprises:
cutting the task cost matrix according to the condition that the unmanned aerial vehicles cannot complete all target area operations and all unmanned aerial vehicles cannot complete operations in the target area;
processing the clipped task cost matrix by utilizing an exhaustive traversal method or an ant algorithm to obtain a plurality of scheme queues plan p Wherein p represents the row number of the clipped task cost matrix, and the row vector H in the clipped task cost matrix p Representing all cost vectors for the p-th target area job;
statistical plan queue plan p Sum of target area ranges S p And when S is p If the value is larger than the threshold value, plan queue is played p The result is regarded as a valid result;
acquiring an optimal scheme queue based on the cost of each effective result;
and allocating the task resources of the available unmanned aerial vehicle according to the optimal scheme queue plan.
5. The method as claimed in claim 4, wherein the clipped task cost matrix is processed using exhaustive traversal or ant algorithm to obtain a plurality of solution queues plan p The method comprises the following steps:
construction scheme queue plan p And is initially empty;
slave row vector H p Medium extracting element C with minimum cost pk The element C is added pk Placing in the plan queue p And deleting the row vector H p With corresponding column vector V k To obtain a matrix MC 0 Wherein k represents a column number of the clipped task cost matrix;
secondary matrix MC t Row vector H of e Medium extracting element C with minimum cost ek The element C is added ek Placing in the plan queue p And deleting the row vector H e With corresponding column vector V k To obtain a matrix MC e+1 Wherein t represents an extraction element C ek E denotes the matrix MC t The row number in (1);
after the clipped task cost matrix is processed, a scheme queue plan is obtained p 。
6. The method of claim 4, wherein the allocating task resources of the available drones according to the optimal solution queue plan comprises:
comparing the number of the available unmanned aerial vehicles with the dimensionality of the optimal scheme queue plan;
if the number of the available unmanned aerial vehicles is not larger than the dimensionality of the optimal scheme queue plan, distributing task resources of the available unmanned aerial vehicles according to the optimal scheme queue plan;
if the number of the available unmanned aerial vehicles is not larger than the dimension of the optimal solution queue plan, optimizing the optimal solution queue plan, wherein the optimizing comprises:
setting a segmentation threshold limit;
according to the operation cost, performing sequence processing on the operation of the unallocated target area to obtain a queue reJob;
judging whether the operation cost of each target area operation in the queue reJob is smaller than the limit of a segmentation threshold value:
if yes, taking the optimal scheme queue plan as an optimization result to perform task resource allocation;
if not, then:
segmenting the operation of the unallocated target area with the operation cost larger than the limitation of a segmentation threshold value, returning to the distribution information, and acquiring the available unmanned aerial vehicle from the current unmanned aerial vehicle swarm queue according to the capacity expression to obtain an optimal scheme queue plan';
comparing the task coverage rate and the cost of the optimal scheme queue plan and the optimal scheme queue plan' to obtain an optimal scheme queue plan 1 ；
Queuing plan by optimization scheme t Corresponding queue reJob t Judging whether the operation cost of each target area operation is smaller than the division threshold limit or not, and judging whether the optimization is continued or not, wherein t is the optimization times;
plan the optimal scheme queue T And performing task resource allocation as an optimization result, wherein T is the total number of optimization times.
7. The method of claim 1, wherein planning a task-in-place path for each drone that allocates task resources comprises:
constructing a situation map in the form of a bitmap aiming at the whole area, wherein the bitmap consists of a plurality of basic units;
setting a task in-place path cost constraint of the unmanned aerial vehicle and a gray value threshold of the basic unit;
constructing vectors based on basic units corresponding to starting positions and target positions of unmanned aerial vehicles
Extracting the vector
All the passed basic units form a sequence according to the vector relation and are placed in a line result R in place;
comparing the cost of the in-place route result R with the task in-place path cost constraint:
if the cost of the in-position line result R is not less than the cost constraint of the task in-position path, jumping to an output unmanned aerial vehicle planning task in-position path;
if the cost of the in-place line result R is not less than the task in-place path cost constraint, then:
extracting a basic unit with the maximum gray value from the in-place line result R;
comparing the gray value of the basic unit with the maximum gray value with the gray value threshold:
if the gray value of the basic unit with the maximum gray value is smaller than the gray value threshold, skipping to judge whether the basic units in the in-place line result R are all processed;
if the gray value of the basic unit with the maximum gray value is larger than the gray value threshold value and at least one basic unit adjacent to the boundary exists, selecting the basic unit adjacent to the boundary as the basic unit with the maximum gray value, and returning to the comparison of the gray value of the basic unit with the maximum gray value with the gray value threshold value;
if the gray value of the basic unit with the maximum gray value is larger than the gray value threshold value and the basic unit with the adjacent boundary does not exist, selecting the basic unit with the minimum gray value to replace and place the basic unit in the in-position line result R, and skipping to judge whether the basic unit in the in-position line result R is completely processed or not;
judging whether all the basic units in the in-place line result R are processed:
if not, returning to the comparison of the cost of the in-place line result R and the task in-place path cost constraint;
if yes, jumping to an output unmanned aerial vehicle planning task in-place path;
and outputting the unmanned plane planning task in-place path.
8. The method of claim 1, wherein planning a task reclamation path for each drone that allocates task resources comprises:
acquiring the current state and the current position of the unmanned aerial vehicle;
acquiring the task in-position cost of the unmanned aerial vehicle for executing the task in-position path, and comparing the task in-position cost with a set threshold S;
under the condition that the task in-place cost is greater than the set threshold S, selecting and judging whether the nearest recovery point meets the requirement; when the requirement is met, planning a task recovery path between the current position and the nearest recovery point; when the requirement is not met, planning a task recovery path between the current position and the original recovery point;
and planning a task recovery path between the current position and an original recovery point under the condition that the task in-place cost is not more than the set threshold S.
9. A multi-objective drone swarm task orchestration device, the device comprising:
the information acquisition module is used for acquiring the distribution information of a target area and the capability expression of each unmanned aerial vehicle in the current unmanned aerial vehicle swarm queue, wherein the capability expression comprises the following steps: standard job unit execution cost, current state, capability reservation, cost constraints and basic parameters;
the unmanned aerial vehicle screening module is used for acquiring available unmanned aerial vehicles from the current unmanned aerial vehicle swarm queue according to the capability expression aiming at the distribution information;
the resource allocation module is used for constructing a task cost matrix of the available unmanned aerial vehicle and the operation of a target area based on the distribution information and the capability expression, and allocating the task resources of the available unmanned aerial vehicle according to the task cost matrix;
and the path planning module is used for planning a task in-place path and a task recovery path for each unmanned aerial vehicle which is allocated with the task resources.
10. A computer device comprising a memory and a processor, the memory having stored therein a computer program, the computer program being loaded and executed by the processor to implement the method of any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210891617.1A CN115310792A (en) | 2022-07-27 | 2022-07-27 | Task cooperation method, device and equipment for multi-target unmanned swarm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210891617.1A CN115310792A (en) | 2022-07-27 | 2022-07-27 | Task cooperation method, device and equipment for multi-target unmanned swarm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115310792A true CN115310792A (en) | 2022-11-08 |
Family
ID=83859535
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210891617.1A Pending CN115310792A (en) | 2022-07-27 | 2022-07-27 | Task cooperation method, device and equipment for multi-target unmanned swarm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115310792A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115456487A (en) * | 2022-11-10 | 2022-12-09 | 深圳市道通智能航空技术股份有限公司 | Task planning method and device of cluster system and electronic equipment thereof |
| CN116596287A (en) * | 2023-07-18 | 2023-08-15 | 中国电子科技集团公司第二十九研究所 | Task driving decision-making method and system |
-
2022
- 2022-07-27 CN CN202210891617.1A patent/CN115310792A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115456487A (en) * | 2022-11-10 | 2022-12-09 | 深圳市道通智能航空技术股份有限公司 | Task planning method and device of cluster system and electronic equipment thereof |
| CN116596287A (en) * | 2023-07-18 | 2023-08-15 | 中国电子科技集团公司第二十九研究所 | Task driving decision-making method and system |
| CN116596287B (en) * | 2023-07-18 | 2023-10-03 | 中国电子科技集团公司第二十九研究所 | Task driving decision-making method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115310792A (en) | Task cooperation method, device and equipment for multi-target unmanned swarm | |
| Hussein et al. | Crane operations and planning in modular integrated construction: Mixed review of literature | |
| CN111582697B (en) | Power distribution network fault assessment and scheduling method and system | |
| CN110703802A (en) | Automatic bridge detection method and system based on multi-unmanned aerial vehicle cooperative operation | |
| US20190235488A1 (en) | Autonomous long range aerial vehicles and fleet management system | |
| CN114840022B (en) | MAS-based multi-unmanned system supervision control method | |
| CN111158401A (en) | Distributed unmanned aerial vehicle path planning system and method for encouraging space-time data exploration | |
| CN115187005B (en) | Scheduling method, device, equipment and storage medium | |
| CN116301055A (en) | Unmanned aerial vehicle inspection method and system based on building construction | |
| CN112230675A (en) | Unmanned aerial vehicle task allocation method considering operation environment and performance in collaborative search and rescue | |
| Li et al. | An UAV scheduling and planning method for post-disaster survey | |
| CN114740899B (en) | Gridding airspace distribution and collaborative search planning method | |
| CN117806355A (en) | Control method and system for electric power line inspection unmanned aerial vehicle | |
| CN118280168B (en) | Low-altitude airspace management method and system based on general sense integration | |
| CN117150757A (en) | Simulation deduction system based on digital twin | |
| CN116110025A (en) | Method and system for constructing environment semantic occupation and velocity field by grid detection tracking framework | |
| CN117519944A (en) | Unmanned vehicle based on cooperation of computing power perception and edge cloud computing and cooperation method thereof | |
| CN117472083B (en) | Multi-unmanned aerial vehicle collaborative marine search path planning method | |
| CN117521932A (en) | Unmanned aerial vehicle inspection management system based on meshing division | |
| Aydin et al. | Heuristics based optimization for multidepot drone location and routing problem to detect post-earthquake damages | |
| CN109947130A (en) | A kind of transmission of electricity multiple aircraft formation tour management method and system based on grid-based management | |
| Vongsantivanich et al. | Mission planning for non-homogeneous Earth observation satellites constellation for disaster response | |
| CN116578120A (en) | Unmanned aerial vehicle scheduling method and device, unmanned aerial vehicle system and computer equipment | |
| CN115562334B (en) | Unmanned plane cluster task cooperation method and system oriented to multi-source disaster situation data fusion | |
| CN115629600B (en) | Multi-machine collaborative trapping method based on buffer Wino diagram in complex dynamic security environment |
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 |