CN115903912A - Multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception - Google Patents
Multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception Download PDFInfo
- Publication number
- CN115903912A CN115903912A CN202211700654.6A CN202211700654A CN115903912A CN 115903912 A CN115903912 A CN 115903912A CN 202211700654 A CN202211700654 A CN 202211700654A CN 115903912 A CN115903912 A CN 115903912A
- Authority
- CN
- China
- Prior art keywords
- unmanned aerial
- aerial vehicle
- deployment
- task
- task allocation
- 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
Images
Classifications
-
- 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
- Traffic Control Systems (AREA)
Abstract
The invention relates to a multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception, which comprises the following steps: and (1) acquiring information of the unmanned aerial vehicle control system. (2) The unmanned aerial vehicle control system models the unmanned aerial vehicle collaborative deployment and task allocation optimization problem. (3) the unmanned aerial vehicle control system optimizes the unmanned aerial vehicle deployment scheme: when task assignment scheme A opt And designing a differential evolution algorithm during giving, and optimizing and generating an unmanned aerial vehicle deployment scheme. And (4) optimizing a task allocation scheme by the unmanned aerial vehicle control system. (5) And the unmanned aerial vehicle control system generates an unmanned aerial vehicle collaborative deployment and task allocation formula. According to the invention, the capabilities of sensing, communication and calculation in three dimensions and the mobility of the unmanned aerial vehicle are reasonably utilized, the reasonable deployment of the position of the unmanned aerial vehicle and the effective distribution of tasks are realized, and the intelligent sensing performance of the urban environment is greatly improved.
Description
Technical Field
The invention relates to the field of unmanned aerial vehicle control, in particular to a multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception.
Background
In recent years, with the development of intelligent technology, the construction of "smart cities" gradually becomes the development trend and target of modern cities, and the premise of intelligent control of cities is effective environmental perception of the cities. The existing urban environment perception information mainly comes from ground sensors deployed in cities, such as cameras and environment monitoring stations, and is difficult to meet the requirements of urban intelligent perception, and the main problems of the urban environment perception information are summarized as follows: (1) The urban environment is complex and has more shelters, the existing sensor has fixed position, is difficult to sense the environment in non-line of sight, and can not realize the global coverage of the city; (2) Most of urban perception sensors only have a perception detection function and do not have computing capability, so perceived data needs to be transmitted to a cloud for computing, the amount of transmitted data is large, the time delay is high, and the real-time performance of a perception environment is difficult to guarantee; (3) Environmental perception is independently carried out between the multisensor, can't carry out effective cooperation, and the perception efficiency is lower.
In order to overcome the limitation of current urban environment intelligent perception, can carry on the sensor unit that possesses communication, perception, computing power through unmanned aerial vehicle, carry out intelligent perception to the urban environment, the concrete expression is: by means of the mobility of the unmanned aerial vehicle, the sensing unit can acquire sensing information of wide-area full coverage to the urban environment. Meanwhile, the unmanned aerial vehicle can utilize the communication unit to transmit data with the ground sensor deployed in the city, receive perception information from the ground sensor, and utilize the calculation unit to calculate and fuse the perception information, so that high-quality environment perception information is obtained.
In practice, the performance of the unmanned aerial vehicle for sensing the environment can be characterized as the completion condition of a sensing task and a computing task, the unmanned aerial vehicle senses a given area and defines the sensing task, receives the sensing information of the ground sensor, performs computing processing and fuses, and defines the sensing task and the computing task as the computing task. When the unmanned aerial vehicle completes more sensing and computing tasks, better environment sensing performance can be obtained.
Because the sensor unit capacity that unmanned aerial vehicle carried on is limited, consequently perception and communication have limited coverage, and different unmanned aerial vehicles are different according to the position difference of oneself, and the perception of performable and calculation task set is different. Meanwhile, the resources of the unmanned aerial vehicle are limited, and the number of executable sensing and computing tasks is also limited. Therefore, it is necessary to perform joint optimization design on cooperative deployment and task allocation of multiple unmanned aerial vehicles, so as to maximize the perception performance of the unmanned aerial vehicles on the urban environment under the constraint of satisfying the deployable areas and the resource capacity of the unmanned aerial vehicles. The position deployment of the unmanned aerial vehicle is optimized, and the position of the unmanned aerial vehicle is determined, so that the unmanned aerial vehicle can cover more tasks in the condition of meeting the constraint of the deployable area. The task allocation of the unmanned aerial vehicles is optimized, the situation that different unmanned aerial vehicles repeatedly execute the same task can be avoided, more tasks are completed under limited resources, and accordingly perception performance is improved.
At present, no design research exists for a multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception.
Disclosure of Invention
In order to solve the technical problem in the prior art, the invention provides a multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception.
The invention specifically comprises the following contents: a multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception comprises the following steps:
step (10), acquiring information of the unmanned aerial vehicle control system: the cloud end generates a corresponding perception task instruction through intelligent analysis and prediction of situation information perceived by the urban environment, and sends task information to the unmanned aerial vehicle control system; the ground sensor sends the calculation tasks to be unloaded to the unmanned aerial vehicle control system according to the sensing information; the unmanned aerial vehicle feeds back information such as self residual resources, communication range, sensing range and the like to an unmanned aerial vehicle control system in real time;
step (20), modeling an unmanned aerial vehicle collaborative deployment and task allocation optimization problem by an unmanned aerial vehicle control system: the unmanned aerial vehicle control system models the unmanned aerial vehicle collaborative deployment and task allocation optimization problem according to the perceived performance index and unmanned aerial vehicle deployment and resource constraint;
step (30), the unmanned aerial vehicle control system optimizes the unmanned aerial vehicle deployment scheme: when task assignment scheme A opt When the unmanned aerial vehicle deployment plan is given, designing a differential evolution algorithm, and optimizing and generating the unmanned aerial vehicle deployment plan;
step (40), the unmanned aerial vehicle control system optimizes the task allocation scheme: when unmanned plane scheme deploys D opt Designing a greedy reconstruction-based discrete particle swarm algorithm at the given time, and optimizing to generate a task allocation scheme;
step (50), the unmanned aerial vehicle control system generates an unmanned aerial vehicle collaborative deployment and task allocation scheme: alternately optimizing unmanned aerial vehicle deployment and task allocation by using an iteration mechanism, generating a final unmanned aerial vehicle collaborative deployment and task allocation scheme, and issuing the scheme to an unmanned aerial vehicle for execution; unmanned aerial vehicle will acquire perception information and send to the high in the clouds, and the high in the clouds realizes the acquisition to city environment overall situation information through the perception information of many unmanned aerial vehicles of analysis integration.
Further, the step (10) of acquiring the information of the unmanned aerial vehicle control system comprises the following steps:
step (11), the cloud generates a perception task instruction set through intelligent analysis and prediction of situation information perceived by the urban environment
Meanwhile, the ground sensor generates a calculation task instruction based on the perception information>
Sending the two task instruction sets to an unmanned aerial vehicle control system; for a perception task->
Which contains a region to be detected>
For a calculation task->
It includes the bit of the task source sensorPut->
Total set of tasks is recorded as +>
Step (12), the unmanned aerial vehicles in the system are integrated into
Unmanned aerial vehicle transmits the residual resource Q back in real time resource To the unmanned aerial vehicle control system; meanwhile, drone u has a limited sensing range S u And communication range C u And transmits it back to the unmanned aerial vehicle control system.
Further, the step (20) of modeling unmanned aerial vehicle collaborative deployment and task allocation optimization problem by the unmanned aerial vehicle control system comprises the following steps:
step (21), an optimized variable unmanned aerial vehicle deployment matrix and a task allocation matrix are defined, wherein the unmanned aerial vehicle deployment matrix is recorded as
d u Representing the three-dimensional position coordinates of the unmanned aerial vehicle u, and D representing the position deployment schemes of all unmanned aerial vehicles; the task assignment matrix is denoted A U×M Wherein the element a u,m E {0,1} represents whether task m is allocated to unmanned plane u for execution;
step (22), defining a connection matrix C between the unmanned aerial vehicle and the task U×M Wherein the element c u,m E {0,1} indicates whether task m can be allocated to unmanned plane u for execution; the connection matrix is determined by the deployment matrix D of the drone: for perceptual tasks
If the area to be detected is->
Perception range S at unmanned plane u u In, then there is c u,m =1, denormal is 0; for computing tasks
If it comes from the sensor position d m Communication range C at unmanned aerial vehicle u u In, then there is c u,m =1, otherwise 0;
step (23), under the unmanned aerial vehicle deployment scheme D and the task allocation scheme A, defining urban environment perception performance F (D, A) as a weighted sum of completed tasks:
wherein p is m Representing the importance of the task m;
(24) Under the condition of meeting the position deployment and task allocation resource constraints of the unmanned aerial vehicle, the urban environment perception performance is maximized by performing combined optimization on the unmanned aerial vehicle deployment and task allocation scheme; the optimization problem can be modeled as:
P1:max {D,A} F(D,A)
s.t.
a u,m ∈{0,1},
the objective function is urban environment perception performance, the first constraint is the position constraint of unmanned aerial vehicle deployment, namely that the unmanned aerial vehicles can only be distributed in legal areas
The second constraint is the resource constraint of the unmanned aerial vehicle, and certain resources r are consumed for executing tasks m The resource consumed by the unmanned aerial vehicle to execute the task cannot exceed the capacity Q of the unmanned aerial vehicle resource 。
Further, the step (30) of optimizing the unmanned aerial vehicle deployment plan by the unmanned aerial vehicle control system comprises the following steps:
step (31), initializing the population scale to be N, and the total evolutionary algebra to be G;
and (32) under the condition that the unmanned aerial vehicle deployment position constraint is met, randomly generating N unmanned aerial vehicle deployment matrixes D as individuals of the first generation, wherein D n (1) A drone deployment matrix representing a first generation of individuals n;
step (33), when evolving the generation g, generating variant individuals for each individual n according to the following formula:
wherein n is 1 ,n 2 ,n 3 Selecting three individuals randomly from the population;
step (34), in the evolution algebra, the original individuals and the variant individuals are subjected to cross operation to obtain new individuals K n (g) The elements are as follows:
step (35), in the given task allocation scheme A opt Next, original individuals D were compared n (g) And a novel individual K n (g) City perception performance of (1), selecting the better performing individual into the next generation, i.e.
Step (36), repeating the steps (33) - (35) until the evolution algebra reaches G; comparing city perception performances of the last generation of N individuals, and selecting the individual with the optimal performance as an output unmanned aerial vehicle deployment scheme D opt 。
Further, the step (40) of optimizing the task allocation scheme by the unmanned aerial vehicle control system comprises the steps of:
step (41), initializing the particle swarm size to be L, and setting the total search period to be T;
and (42) randomly generating L task allocation schemes serving as initial positions of the particles under the condition that the task allocation resource constraint is met, wherein A l (1) Indicating the position of the ith particle in the first search period; in addition, each particle/initialization generates a velocity matrix V l (1) The element is in-v max And v max Are subjected to uniform distribution;
step (43), in the search period t, calculating the deployment scheme D at the given unmanned aerial vehicle opt City perception performance F (D) under task allocation scheme corresponding to different particles opt ,A l (t))；
Step (44), each particle updates the local optimal position lbest from initialization to the current search period t l (t) that is
Comparing the local optimal positions of all the particles in the particle swarm, and selecting the local optimal position with the optimal city perception performance as the global optimal gbest (t):
and (45) updating the speed matrix of the particles in the search period t:
V l (t)＝V l (t-1)+a 1 b 1 (lbest l (t-1)-A l (t-1))+a 2 b 2 (gbest(t-1)-A l (t-1))，
wherein the first term represents the kinetic inertia of the particle, the second and third terms represent the effect of the individual and population cognition of the particle, respectively, on it, a 1 And a 2 Denotes the acceleration factor, b 1 And b 2 Obey [0,1]Are uniformly distributed;
and (46) updating the position matrix of the particles according to the speed matrix:
wherein s is a random variable subject to a standard normal distribution;
in step (47), since the discrete particle swarm algorithm can only ensure that the position matrix of the particle satisfies the constraint of {0,1} and cannot satisfy the resource capacity constraint, the particle needs to be reconstructed. And judging whether the resources consumed by the task allocation scheme corresponding to the particle position matrix exceed the resource constraint, and if so, performing greedy reconstruction on the particles. Namely, for all the distributed tasks, the reduction of the city perception performance when one task is eliminated is calculated, and the distributed task with the minimum performance reduction and resource consumption ratio is selected to be deleted. Repeating the above steps until the resource capacity constraint is satisfied;
and (48) repeating the steps (43) to (47) until the number of search cycles reaches T, and taking the global optimal gbest (T) as the task allocation scheme A of the output opt 。
Further, the step (50) of generating the unmanned aerial vehicle collaborative deployment and task allocation scheme by the unmanned aerial vehicle control system comprises the following steps:
step (51), randomly generating unmanned aerial vehicle cooperative deployment and task allocation scheme meeting constraint as initialization scheme D of system (1) And A (1) ；
Step (52), in the iteration period i, fixing the task allocation scheme A (i-1) Optimizing and generating unmanned aerial vehicle deployment distribution scheme D according to the differential evolution algorithm in the step (30) (i) ；
Step (53), in the iteration period i, fixing the unmanned aerial vehicle deployment scheme D (i) Optimizing and generating a task allocation scheme A according to the discrete particle swarm optimization algorithm based on greedy reconstruction in the step (40) (i) ；
Step (54), repeating the steps (52) to (53), and performing iterative optimization on the unmanned aerial vehicle deployment scheme and the task allocation scheme until the iteration times reach the maximum value or the algorithm performance is not improved, so as to generate a final unmanned aerial vehicle cooperative deployment and task allocation scheme;
step (55), the unmanned aerial vehicle moves and executes tasks according to the generated scheme, and sends the acquired sensing information to the cloud; the cloud end integrates the received perception information through analysis, and the whole situation information of the urban environment is acquired.
The invention has the beneficial effects that: aiming at the intelligent perception scene of the urban environment, the invention designs a multi-unmanned aerial vehicle collaborative deployment and task allocation method, reasonably utilizes the capabilities of perception, communication and calculation in three dimensions and the mobility of the unmanned aerial vehicle, realizes reasonable deployment of the positions of the unmanned aerial vehicles and effective allocation of tasks, and greatly improves the intelligent perception performance of the urban environment.
Drawings
The following further explains embodiments of the present invention with reference to the drawings.
Fig. 1 is a flow chart of the operation of the unmanned aerial vehicle control system;
FIG. 2 is a model of an unmanned aerial vehicle control system;
FIG. 3 is a flow chart of an unmanned aerial vehicle collaborative deployment and task allocation algorithm;
FIG. 4 is a graph comparing the performance of the method of the present invention with other algorithms at different resource capacities.
Detailed Description
With reference to fig. 1 to 4, the method for cooperative deployment and task allocation of multiple unmanned aerial vehicles under urban intelligent perception of the invention comprises the following steps:
(10) Acquiring information of the unmanned aerial vehicle control system: the cloud end generates a corresponding perception task instruction through intelligent analysis and prediction of situation information perceived by the urban environment, and sends task information to the unmanned aerial vehicle control system; the ground sensor sends the calculation tasks to be unloaded to the unmanned aerial vehicle to an unmanned aerial vehicle control system according to the sensing information; the unmanned aerial vehicle feeds back information such as self residual resources, communication range, perception range and the like to the unmanned aerial vehicle control system in real time. A network system model diagram is shown in fig. 2.
(11) Cloud through situational information perceived by urban environmentIntelligent analysis and prediction, generating a set of perceptual task instructions
Meanwhile, the ground sensor generates a calculation task instruction based on the perception information>
And sending the two types of task instruction sets to the unmanned aerial vehicle control system. For a perceptual task>
Which contains a region to be detected>
For computing tasks
Which contains the position of the task origin sensor->
The total set of tasks is noted as
(12) The unmanned aerial vehicles in the system are integrated into
Unmanned aerial vehicle transmits the residual resource Q back in real time resource To the drone control system. Meanwhile, drone u has a limited sensing range S u And communication range C u And transmits it back to the unmanned aerial vehicle control system.
(20) The unmanned aerial vehicle control system models the unmanned aerial vehicle collaborative deployment and task allocation optimization problem: and the unmanned aerial vehicle control system models the unmanned aerial vehicle collaborative deployment and task allocation optimization problem according to the perceived performance index and unmanned aerial vehicle deployment and resource constraint.
(21) Defining an optimized variable unmanned aerial vehicle deployment matrix and a task allocation matrix, whereinHuman deployment matrix notation
d u Representing the three-dimensional position coordinates of the unmanned aerial vehicle u, and D representing the position deployment schemes of all unmanned aerial vehicles; the task allocation matrix is denoted A U×M Wherein the element a u,m E {0,1} indicates whether task m is allocated for drone u execution.
(22) Defining a connection matrix C between a drone and a task U×M Wherein the element c u,m E {0,1} indicates whether task m can be allocated for drone u execution. The connection matrix is determined by the unmanned aerial vehicle deployment matrix D: for perceptual tasks
If the area to be detected>
Perception range S at unmanned plane u u In, then there is c u,m =1, denormal is 0; for a calculation task->
If it comes from the sensor position d m Communication range C at drone u u In, then there is c u,m And =1, otherwise 0.
(23) Under the unmanned plane deployment scheme D and the task allocation scheme A, the urban environment perception performance F (D, A) is defined as the weighted sum of the completed tasks:
wherein p is m Indicating the importance of task m.
(24) Under the condition of meeting the position deployment and task allocation resource constraints of the unmanned aerial vehicle, the urban environment perception performance is maximized by performing joint optimization on the unmanned aerial vehicle deployment and task allocation scheme. The optimization problem can be modeled as:
P1:max {D,A} F(D,A)
a u,m ∈{0,1},
the target function is urban environment perception performance, the first constraint is the position constraint of unmanned aerial vehicle deployment, namely that the unmanned aerial vehicles can only be distributed in legal areas
The second constraint is the resource constraint of the unmanned aerial vehicle, and certain resources r are consumed for executing tasks m The resource consumed by the unmanned aerial vehicle to execute the task cannot exceed the capacity Q of the unmanned aerial vehicle resource 。
(30) The unmanned aerial vehicle control system optimizes the unmanned aerial vehicle deployment scheme: when task assignment scheme A opt And designing a differential evolution algorithm during giving, and optimizing and generating an unmanned aerial vehicle deployment scheme.
(31) The initialization population size is N, and the total evolutionary algebra is G.
(32) Randomly generating N unmanned aerial vehicle deployment matrixes D as individuals of a first generation under the condition of satisfying unmanned aerial vehicle deployment position constraint, wherein D n (1) A drone deployment matrix representing a first generation of individuals n.
(33) In evolving algebra g, each individual n generates variant individuals according to the following formula:
wherein n is 1 ,n 2 ,n 3 Three individuals randomly selected in the population.
(34) In advance ofWhen algebra is changed, the original individual and the variant individual are subjected to cross operation to obtain a new individual K n (g) The elements are as follows:
(35) At a given task allocation scheme A opt Next, original individuals D were compared n (g) And a novel individual K n (g) City perception performance of (1) selecting the more excellent individual to enter the next generation, i.e.
(36) And (5) repeating the steps (33) - (35) until the evolution passage number reaches G. Comparing city perception performances of the last generation of N individuals, and selecting the individual with the optimal performance as an output unmanned aerial vehicle deployment scheme D opt 。
(40) The unmanned aerial vehicle control system optimizes a task allocation scheme: when unmanned plane scheme deploys D opt And designing a greedy reconstruction-based discrete particle swarm algorithm at the given time, and optimizing and generating a task allocation scheme.
(41) Initializing the particle swarm to be L, and setting the total search period to be T.
(42) Under the condition of meeting the task allocation resource constraint, randomly generating L task allocation schemes as the initial positions of the particles, wherein A l (1) Indicating the position of the ith particle in the first search cycle. In addition, each particle/initialization generates a velocity matrix V l (1) The element of which is in-v max And v max Subject to uniform distribution.
(43) In the search period t, the deployment scenario D at a given drone is calculated opt City perception performance F (D) under task allocation scheme corresponding to different particles opt ,A l (t))。
(44) Local optimal position lbest from initialization to current search period t for each particle update l (t) that is
Comparing the local optimal positions of all the particles in the particle swarm, and selecting the local optimal position with the optimal city perception performance as the global optimal gbest (t):
(45) In the search period t, the velocity matrix of the particles is updated:
V l (t)＝V l (t-1)+a 1 b 1 (lbest l (t-1)-A l (t-1))+a 2 b 2 (gbest(t-1)-A l (t-1))，
wherein the first term represents the kinetic inertia of the particle, the second and third terms represent the effect of the individual and population cognition of the particle, respectively, on it, a 1 And a 2 Denotes the acceleration factor, b 1 And b 2 Compliance [0,1]Are evenly distributed in between.
(46) The position matrix of the particles is updated according to the speed matrix:
where s is a random variable that follows a standard normal distribution.
(47) Since the discrete particle swarm algorithm can only ensure that the position matrix of the particle meets the constraint of the element {0,1} and cannot meet the resource capacity constraint, the particle needs to be reconstructed. And judging whether the resources consumed by the task allocation scheme corresponding to the particle position matrix exceed the resource constraint, and if so, performing greedy reconstruction on the particles. Namely, for all the distributed tasks, the reduction of the city perception performance when one task is eliminated is calculated, and the distributed task with the minimum performance reduction and resource consumption ratio is selected to be deleted. And repeating the reconstruction steps until the resource capacity constraint is met.
(48) Repeating the steps (43) to (47) until the number of search cycles reaches T, and taking the global optimal gbest (T) as the task allocation scheme A of the output opt 。
(50) The unmanned aerial vehicle control system generates an unmanned aerial vehicle collaborative deployment and task allocation scheme: and alternately optimizing unmanned aerial vehicle deployment and task allocation by using an iteration mechanism, generating a final unmanned aerial vehicle collaborative deployment and task allocation scheme, and issuing the scheme to the unmanned aerial vehicle for execution. Unmanned aerial vehicle will acquire perception information and send to the high in the clouds, and the high in the clouds realizes the acquisition to city environment overall situation information through the perception information of many unmanned aerial vehicles of analysis integration.
(51) Randomly generating unmanned aerial vehicle collaborative deployment and task allocation scheme meeting constraint as initialization scheme D of system (1) And A (1) 。
(52) In iteration cycle i, the task assignment scheme A is fixed (i-1) Optimizing and generating the unmanned aerial vehicle deployment distribution scheme D according to the differential evolution algorithm in the step (30) (i) 。
(53) In iteration cycle i, fixed unmanned aerial vehicle deployment scenario D (i) And (4) optimizing and generating a task distribution scheme A according to the discrete particle swarm optimization algorithm based on the greedy reconstruction in the step (40) (i) 。
(54) And (5) repeating the steps (52) to (53), and performing iterative optimization on the unmanned aerial vehicle deployment scheme and the task allocation scheme until the iteration times reach the maximum value or the algorithm performance is not improved, so as to generate a final unmanned aerial vehicle cooperative deployment and task allocation scheme.
(55) And the unmanned aerial vehicle moves and executes tasks according to the generated scheme, and sends the acquired sensing information to the cloud. The cloud end integrates the received perception information through analysis, and the whole situation information of the urban environment is acquired.
Fig. 4 compares the performance of the crowd-sourcing iterative algorithm proposed by the present invention with other algorithms at different drone resource capacities. It can be seen that the performance of the algorithm proposed by the present invention is always better than the other two algorithms. For the unmanned aerial vehicle deployment scheme, the differential evolution algorithm provided by the invention is superior to a heuristic random deployment method; for task allocation, the discrete particle swarm optimization algorithm based on the greedy reconstruction is superior to the traditional greedy algorithm. Meanwhile, the alternate iteration optimization mechanism can well converge the algorithm, and the performance of cooperative intelligent sensing of the multiple unmanned aerial vehicles is greatly improved.
Compared with the prior art, the invention has the following remarkable advantages: carry on the sensor unit that possesses perception, communication, computing power through unmanned aerial vehicle, propose the many unmanned aerial vehicle urban environment intelligence perception framework that combines perception, communication, calculation three kinds of abilities, through carrying out perception and calculation task, acquire perception information, promote urban environment perception performance. Meanwhile, through collaborative optimization design of deployment and task allocation schemes of the multiple unmanned aerial vehicles, effective intelligent perception of the multiple unmanned aerial vehicles on urban environments in a limited resource scene is achieved, and a foundation is laid for construction of a smart city.
In the above description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The foregoing description is only a preferred embodiment of the invention, which can be embodied in many different forms than described herein, and therefore the invention is not limited to the specific embodiments disclosed above. And that those skilled in the art may, using the methods and techniques disclosed above, make numerous possible variations and modifications to the disclosed embodiments, or modify equivalents thereof, without departing from the scope of the claimed embodiments. Any simple modification, equivalent change and modification of the above embodiments according to the technical essence of the present invention are within the scope of the technical solution of the present invention.
Claims (6)
1. A multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception is characterized in that: the method comprises the following steps:
step (10), acquiring information of the unmanned aerial vehicle control system: the cloud end generates a corresponding perception task instruction through intelligent analysis and prediction of situation information perceived by the urban environment, and sends task information to the unmanned aerial vehicle control system; the ground sensor sends the calculation tasks to be unloaded to the unmanned aerial vehicle control system according to the sensing information; the unmanned aerial vehicle feeds back information such as self residual resources, communication range, sensing range and the like to an unmanned aerial vehicle control system in real time;
step (20), modeling an unmanned aerial vehicle collaborative deployment and task allocation optimization problem by an unmanned aerial vehicle control system: the unmanned aerial vehicle control system models the unmanned aerial vehicle collaborative deployment and task allocation optimization problem according to the perceived performance index and unmanned aerial vehicle deployment and resource constraint;
step (30), the unmanned aerial vehicle control system optimizes the unmanned aerial vehicle deployment scheme: when task assignment scheme A opt When the unmanned aerial vehicle deployment plan is given, designing a differential evolution algorithm, and optimizing and generating the unmanned aerial vehicle deployment plan;
step (40), the unmanned aerial vehicle control system optimizes the task allocation scheme: when unmanned plane scheme deploys D opt Designing a greedy reconstruction-based discrete particle swarm algorithm at the given time, and optimizing to generate a task allocation scheme;
and (50) generating an unmanned aerial vehicle collaborative deployment and task allocation scheme by the unmanned aerial vehicle control system: alternately optimizing unmanned aerial vehicle deployment and task allocation by using an iteration mechanism, generating a final unmanned aerial vehicle collaborative deployment and task allocation scheme, and issuing the scheme to an unmanned aerial vehicle for execution; unmanned aerial vehicle will acquire perception information and send to the high in the clouds, and the high in the clouds realizes the acquisition to city environment overall situation information through the perception information of many unmanned aerial vehicles of analysis integration.
2. The urban intelligent perception multi-unmanned aerial vehicle collaborative deployment and task allocation method according to claim 1, characterized in that: the information acquisition step of the unmanned aerial vehicle control system comprises the following steps:
step (11), the cloud generates a perception task instruction set through intelligent analysis and prediction of situation information of urban environment perception
Meanwhile, the ground sensor generates a calculation task instruction based on the perception information>
Sending the two types of task instruction sets to an unmanned aerial vehicle control system; for a perceptual task>
Which comprises a region to be detected>
For a calculation task->
Which contains the position of the task origin sensor->
The total set of tasks is noted as
Step (12), the unmanned aerial vehicles in the system are integrated into
Unmanned aerial vehicle transmits the residual resource Q back in real time resource To the unmanned aerial vehicle control system; meanwhile, drone u has a limited sensing range S u And communication range C u And transmits it back to the unmanned aerial vehicle control system.
3. The urban intelligent perception multi-unmanned aerial vehicle collaborative deployment and task allocation method according to claim 1, characterized in that: step (20), the unmanned aerial vehicle collaborative deployment and task allocation optimization problem modeling step of the unmanned aerial vehicle control system comprises the following steps:
step (21), an optimized variable unmanned aerial vehicle deployment matrix and a task allocation matrix are defined, wherein the unmanned aerial vehicle deployment matrix is recorded as
d u Three-dimensional position coordinates representing unmanned plane uD represents the location deployment scenario for all drones; the task allocation matrix is denoted A U×M Wherein the element a u,m E {0,1} represents whether task m is allocated to unmanned plane u for execution;
step (22), defining a connection matrix C between the unmanned aerial vehicle and the task U×M Wherein the element c u,m E {0,1} represents whether task m can be allocated to unmanned plane u for execution; the connection matrix is determined by the unmanned aerial vehicle deployment matrix D: for perceptual tasks
If the area to be detected>
Perception range S at unmanned plane u u In, then there is c u,m =1, denormal is 0; for computing tasks
If it comes from the sensor position d m Communication range C at drone u u In, then there is c u,m =1, otherwise 0; />
Step (23), under the unmanned plane deployment scheme D and the task allocation scheme A, defining urban environment perception performance F (D, A) as a weighted sum of completed tasks:
wherein p is m Representing the importance of the task m;
(24) Under the condition of meeting the position deployment and task allocation resource constraints of the unmanned aerial vehicle, the urban environment perception performance is maximized by performing combined optimization on the unmanned aerial vehicle deployment and task allocation scheme; the optimization problem can be modeled as:
the objective function is urban environment perception performance, the first constraint is the position constraint of unmanned aerial vehicle deployment, namely that the unmanned aerial vehicles can only be distributed in legal areas
The second constraint is the resource constraint of the unmanned aerial vehicle, and certain resources r are consumed for executing tasks m The resource consumed by the unmanned aerial vehicle to execute the task cannot exceed the capacity Q of the unmanned aerial vehicle resource 。
4. The urban intelligent perception multi-unmanned aerial vehicle collaborative deployment and task allocation method according to claim 1, wherein the step (30) of the unmanned aerial vehicle control system optimizing the unmanned aerial vehicle deployment scenario comprises the steps of:
step (31), initializing the population scale to be N, and the total evolutionary algebra to be G;
and (32) randomly generating N unmanned aerial vehicle deployment matrixes D as individuals of a first generation under the condition of meeting the unmanned aerial vehicle deployment position constraint, wherein D n (1) A drone deployment matrix representing a first generation of individuals n;
step (33), when the algebra g is evolved, generating variant individuals by each individual n according to the following formula:
wherein n is 1 ,n 2 ,n 3 Selecting three individuals randomly in the population;
step (34), in the evolution algebra, the original individuals and the variant individuals are subjected to cross operation to obtain new individuals K n (g) The elements are as follows:
step (35), at the given task allocation plan A opt Next, original individuals D were compared n (g) And a novel individual K n (g) City perception performance of (1) selecting the more excellent individual to enter the next generation, i.e.
Step (36), repeating steps (33) - (35) until the evolution algebra reaches G; comparing city perception performances of the last generation of N individuals, and selecting the individual with the optimal performance as an output unmanned aerial vehicle deployment scheme D opt 。
5. The urban intelligent perception multi-unmanned aerial vehicle collaborative deployment and task allocation method according to claim 1, wherein the step (40) of the unmanned aerial vehicle control system optimizing the task allocation scheme comprises the steps of:
step (41), initializing the particle swarm size to be L, and setting the total search period to be T;
and (42) randomly generating L task allocation schemes as initial positions of the particles under the condition of meeting the task allocation resource constraint, wherein A l (1) Indicating the position of the ith particle in the first search period; in addition, each particle/initialization generates a velocity matrix V l (1) The element is in-v max And v max Are subjected to uniform distribution;
step (43), in the search period t, calculating the deployment scheme D at the given unmanned aerial vehicle opt City perception performance F (D) under task allocation scheme corresponding to different particles opt ,A l (t))；
Step (44), each particle updates the local optimal position lbest from initialization to the current search period t l (t) that is
Comparing the local optimal positions of all the particles in the particle swarm, and selecting the local optimal position with the optimal city perception performance as the global optimal gbest (t):
and (45) updating the speed matrix of the particles in the search period t:
V l (t)＝V l (t-1)+a 1 b 1 (lbest l (t-1)-A l (t-1))+a 2 b 2 (gbest(t-1)-A l (t-1))，
wherein the first term represents the kinetic inertia of the particle, the second and third terms represent the effect of the individual and population cognition of the particle, respectively, on it, a 1 And a 2 Denotes the acceleration factor, b 1 And b 2 Compliance [0,1]Are uniformly distributed;
and (46) updating the position matrix of the particles according to the speed matrix:
wherein s is a random variable that follows a standard normal distribution;
in step (47), since the discrete particle swarm algorithm can only ensure that the position matrix of the particle satisfies the constraint of {0,1} and cannot satisfy the resource capacity constraint, the particle needs to be reconstructed. And judging whether the resources consumed by the task allocation scheme corresponding to the particle position matrix exceed the resource constraint, and if so, performing greedy reconstruction on the particles. Namely, for all the distributed tasks, the reduction of the city perception performance when one task is eliminated is calculated, and the distributed task with the minimum performance reduction and resource consumption ratio is selected to be deleted. Repeating the above steps until the resource capacity constraint is satisfied;
and (48) repeating the steps (43) to (47) until the number of search cycles reaches T, and taking the global optimal gbest (T) as the task allocation scheme A of the output opt 。
6. The urban intelligent perception multi-unmanned aerial vehicle collaborative deployment and task allocation method according to claim 1, wherein the step (50) of the unmanned aerial vehicle control system generating the unmanned aerial vehicle collaborative deployment and task allocation scheme comprises the steps of:
step (51), randomly generating unmanned aerial vehicle cooperative deployment and task allocation scheme meeting constraint as initialization scheme D of system (1) And A (1) ；
Step (52), in the iteration period i, fixing the task allocation scheme A (i-1) Optimizing and generating the unmanned aerial vehicle deployment distribution scheme D according to the differential evolution algorithm in the step (30) (i) ；
Step (53), in the iteration period i, fixing the unmanned aerial vehicle deployment scheme D (i) And (4) optimizing and generating a task distribution scheme A according to the discrete particle swarm optimization algorithm based on the greedy reconstruction in the step (40) (i) ；
Step (54), repeating the steps (52) to (53), and performing iterative optimization on the unmanned aerial vehicle deployment scheme and the task allocation scheme until the iteration times reach the maximum value or the algorithm performance is not improved, so as to generate a final unmanned aerial vehicle cooperative deployment and task allocation scheme;
step (55), the unmanned aerial vehicle moves and executes tasks according to the generated scheme, and sends the acquired sensing information to the cloud; the cloud side integrates the received perception information through analysis, and the urban environment overall situation information is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211700654.6A CN115903912A (en) | 2022-12-29 | 2022-12-29 | Multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211700654.6A CN115903912A (en) | 2022-12-29 | 2022-12-29 | Multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115903912A true CN115903912A (en) | 2023-04-04 |
Family
ID=86489738
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211700654.6A Pending CN115903912A (en) | 2022-12-29 | 2022-12-29 | Multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115903912A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117742973A (en) * | 2024-02-19 | 2024-03-22 | 中国电子科技集团公司第十五研究所 | Unmanned intelligent perception system and implementation method thereof |
-
2022
- 2022-12-29 CN CN202211700654.6A patent/CN115903912A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117742973A (en) * | 2024-02-19 | 2024-03-22 | 中国电子科技集团公司第十五研究所 | Unmanned intelligent perception system and implementation method thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liu et al. | Distributed and energy-efficient mobile crowdsensing with charging stations by deep reinforcement learning | |
| CN111862579B (en) | Taxi scheduling method and system based on deep reinforcement learning | |
| CN112016812B (en) | Multi-unmanned aerial vehicle task scheduling method, system and storage medium | |
| CN109583665B (en) | Unmanned aerial vehicle charging task scheduling method in wireless sensor network | |
| CN113282368B (en) | Edge computing resource scheduling method for substation inspection | |
| CN111158401B (en) | Distributed unmanned aerial vehicle path planning system and method for encouraging space-time data exploration | |
| Boubin et al. | Managing edge resources for fully autonomous aerial systems | |
| CN110411464B (en) | Three-dimensional point cloud map generation method, device, equipment and storage medium | |
| Rahman et al. | A cloud robotics framework of optimal task offloading for smart city applications | |
| Wang et al. | Wireless sensor network deployment of 3D surface based on enhanced grey wolf optimizer | |
| CN115903912A (en) | Multi-unmanned aerial vehicle collaborative deployment and task allocation method under urban intelligent perception | |
| CN112911705A (en) | Bayesian iteration improved particle swarm optimization algorithm-based indoor positioning method | |
| CN115373424B (en) | Time window-oriented multi-unmanned aerial vehicle airport site selection and random task scheduling method | |
| CN115951587B (en) | Automatic driving control method, device, equipment, medium and automatic driving vehicle | |
| CN113163365A (en) | Unmanned aerial vehicle support networking resource optimization method based on alternating direction multiplier algorithm | |
| CN113485409A (en) | Unmanned aerial vehicle path planning and distribution method and system for geographic fairness | |
| CN114943456A (en) | Resource scheduling method and device, electronic equipment and storage medium | |
| CN115632956A (en) | Chargeable Internet of things node optimization deployment method based on trusted information coverage | |
| CN107748502A (en) | The passive spatial perception exchange method of entity in operation emulation based on discrete event | |
| Zhang et al. | Offloading demand prediction-driven latency-aware resource reservation in edge networks | |
| CN116302401A (en) | Airborne intelligent computing device and method for optimizing multi-computing task scheduling | |
| CN115720226A (en) | Time delay perception resource reservation method based on space-time task demand prediction | |
| CN110146102B (en) | Path planning method, device, equipment and storage medium | |
| Cao et al. | Research On Regional Traffic Flow Prediction Based On MGCN-WOALSTM | |
| CN111833088A (en) | Supply and demand prediction method and device |
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 |