KR20210066696A - Uav arrangement device and method based on machine learning for maximizing communication throughput - Google Patents

Abstract

The present invention relates to a clustered unmanned aerial vehicle (UAV) arrangement device based on machine learning for maximizing communication throughput. The clustered UAV arrangement device comprises: a communication module unit which receives location information and communication demand information of a user and transmits UAV arrangement information; a user information unit which stores and analyzes the location information and the communication demand information of the user; a UAV information unit which stores basic UAV information including UAV communication throughput; and a clustering algorithm modeling unit which calculates a location at which communication throughput is maximized, with a clustering algorithm based on the location information and the communication demand information of the user. The present invention relates to a UAV arrangement method based on machine learning for maximizing communication throughput, which comprises the steps of: checking the location information of the user; checking communication demand of the user; calculating central nodes of respective clusters at which the communication throughput is maximized, by using the clustering algorithm; and arranging UAVs of the central nodes of the respective clusters. The present invention can respond to user communication requirements which are not constant.

Description

Machine learning-based swarm drone deployment device and method for maximizing communication throughput {UAV ARRANGEMENT DEVICE AND METHOD BASED ON MACHINE LEARNING FOR MAXIMIZING COMMUNICATION THROUGHPUT}

The present invention relates to swarm unmanned aerial vehicle deployment based on machine learning for maximizing communication throughput, and more particularly, to an optimal unmanned aerial vehicle deployment according to a change in a user's communication demand based on reinforcement learning.

5G, the next-generation mobile communication technology, is a wireless network technology using millimeter waves. Although commercialization of 5G is currently underway, more base stations are required for more stable commercialization. However, additionally installing a new base station causes economic inefficiency, and causes disadvantages in being unable to flexibly cope with changes in the communication environment. Therefore, instead of installing an additional terrestrial base station, research on a method using an unmanned aerial vehicle is being actively conducted.

An unmanned aerial vehicle or drone refers to a vehicle that the pilot does not directly board. Unmanned aerial vehicles were first developed for military use, but are now being used in a variety of fields, including the private and public sectors. In addition, the unmanned aerial vehicle has great potential in the communication field due to its cost aspect and free movement, and is evaluated as an alternative that can perform the function of a terrestrial base station in the future. However, due to the limited battery capacity and communication range, there is a limit regarding the performance of the function as the base-Station. Furthermore, when a sudden change in the user's communication demand occurs, the problem of the ability of the unmanned aerial vehicle to respond flexibly remains a problem to be solved.

An object of the present invention is to provide an apparatus and method for disposing a swarm unmanned aerial vehicle so that the communication throughput can be maximized according to the communication requirements of a plurality of users.

An object of the present invention is to provide an apparatus and method for relocating an unmanned aerial vehicle so that communication throughput can be maximized when a communication overload occurs after the unmanned aerial vehicle is deployed.

An object of the present invention is to provide an apparatus and method for disposing an unmanned aerial vehicle so that communication throughput can be maximized according to a change in a plurality of communication demands through a reinforcement learning algorithm.

The present invention relates to a machine learning-based swarm unmanned aerial vehicle deployment device for maximizing communication throughput, a communication module unit that receives user's location information and communication demand information and transmits unmanned aerial vehicle deployment information, and a user's location information and communication demand The user information unit that stores and analyzes information, the unmanned aerial vehicle information unit that stores basic drone information including the drone communication throughput, and the location where communication throughput is maximized through a clustering algorithm based on the user's location information and communication demand information It includes a clustering algorithm model unit.

The present invention relates to a method for deploying a swarm unmanned aerial vehicle based on machine learning for maximizing communication throughput. calculating the central node of each cluster, and deploying the unmanned aerial vehicle of the central node of each cluster.

According to an embodiment of the present invention, it is possible to provide a communication network service capable of responding to an inconsistent user's communication demand by using an unmanned aerial vehicle.

According to an embodiment of the present invention, when a user's communication demand increases instantaneously, the unmanned aerial vehicle can be rearranged to quickly cope with a communication overload environment.

According to an embodiment of the present invention, an unmanned aerial vehicle may be placed in an optimal position for maximizing communication throughput without initial training data through a reinforcement learning algorithm.

According to an embodiment of the present invention, it is possible to provide a new communication network environment that can reduce the installation cost of a terrestrial base station and quickly cope with a changing environment.

1 is an exemplary diagram of swarm unmanned aerial vehicle deployment for maximizing communication throughput according to an embodiment of the present invention.
2 is a block diagram of a swarm unmanned aerial vehicle deployment apparatus for maximizing communication throughput according to an embodiment of the present invention.
3 is a diagram for explaining a modified K-means algorithm according to an embodiment of the present invention.
4 is a diagram for explaining an avoidance algorithm according to an embodiment of the present invention.
5 is a flowchart illustrating a method for deploying a swarm unmanned aerial vehicle for maximizing communication throughput according to an embodiment of the present invention.
6 is a flowchart illustrating an avoidance algorithm according to an embodiment of the present invention.
7 is a view showing the experimental results regarding the performance of the swarm unmanned aerial vehicle deployment device for maximizing communication throughput according to an embodiment of the present invention.
8 is a view of an experimental result for confirming the performance of a swarm unmanned aerial vehicle deployment device for maximizing communication throughput according to an embodiment of the present invention.

The above-described objects, features, and advantages will be described below in detail with reference to the accompanying drawings, and accordingly, those skilled in the art to which the present invention pertains will be able to easily implement the technical idea of the present invention. In describing the present invention, if it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description will be omitted.

In the drawings, the same reference numerals are used to refer to the same or similar elements, and all combinations described in the specification and claims may be combined in any manner. And unless otherwise specified, it is to be understood that references to the singular may include one or more, and references to the singular may also include the plural.

The terminology used herein is for the purpose of describing specific exemplary embodiments only and is not intended to be limiting. As used herein, singular expressions may also be intended to include plural meanings unless the sentence clearly indicates otherwise. The term “and/or,” “and/or” includes any and all combinations of the items listed therewith. The terms "comprises", "comprising", "comprising", "comprising", "having", "having" and the like have an implied meaning, so that these terms refer to their described features, integers, It specifies steps, operations, elements, and/or components and does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. The steps, processes, and acts of the methods described herein should not be construed as necessarily performing their performance in such a specific order as discussed or exemplified, unless the order of performance thereof is physically determined. . It should also be understood that additional or alternative steps may be used.

In addition, each component may be implemented as a hardware processor, respectively, the above components may be integrated into one hardware processor, or the above components may be combined with each other and implemented as a plurality of hardware processors.

Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

1 is an exemplary diagram for swarm unmanned aerial vehicle deployment for maximizing communication throughput according to an embodiment of the present invention.

The unmanned aerial vehicle of the present invention is a device capable of flying by itself. For example, it may be named as an unmanned aerial vehicle, a drone, or the like. In the present invention, the term 'unmanned aerial vehicle' will be uniformly described. The unmanned aerial vehicle may serve as a communication node or relay node in a wireless communication network. A wireless communication network refers to a network that performs communication wirelessly. Wireless communication networks exist as various networks according to a communication method used. For example, various technologies exist, such as a mobile communication network such as 3G/LTE, a wireless communication network according to IEEE 802.11, and a network utilizing 5G technology. The unmanned aerial vehicle of the present invention may have a communication module capable of performing a communication function.

Referring to FIG. 1 , it can be seen that the swarm unmanned aerial vehicle 100 is disposed. The unmanned aerial vehicle may perform the communication function of a ground base station. The unmanned aerial vehicle may provide a communication service to the user by checking the communication demand of the user and utilizing the communication module of each unmanned aerial vehicle accordingly. A user's communication requirements may vary depending on the user's communication usage. Accordingly, the swarm unmanned aerial vehicle 100 may be deployed according to the users 200 having different communication requirements. The swarm unmanned aerial vehicle 100 may be deployed to maximize the user's communication requirements.

According to an embodiment of the present invention, a phenomenon in which communication traffic converges may occur because users are concentrated in one physical area 201 . Accordingly, it is possible to handle the communication throughput of the area 201 where communication traffic is concentrated by changing the arrangement of the swarm aircraft 101 in order to handle the communication demand. Since the communication coverage and communication throughput of one unmanned aerial vehicle of the present invention are constant, when communication traffic is congested, the number of unmanned aerial vehicle deployments in the area may be increased or an additional unmanned aerial vehicle may be introduced to handle the communication demand.

Hereinafter, with reference to FIG. 2, a detailed description will be given of an apparatus for optimally disposing an unmanned aerial vehicle for maximizing communication throughput.

2 is a block diagram of a swarm unmanned aerial vehicle deployment apparatus for maximizing communication throughput according to an embodiment of the present invention.

In the present invention, the communication throughput refers to a communication service throughput provided to a user by deploying a cluster of unmanned aerial vehicles for providing a communication service. More specifically, the description of communication throughput and the meaning of maximizing communication throughput will be explained using The Air-To-Ground Path Loss Model and Max-Min Airtime Fairness Model.

The Air-To-Ground Path Loss Model is a model for calculating the SNR (Signal-to-Noise Ratio) due to the signal attenuation effect according to the distance (d) in the communication process between the unmanned aerial vehicle and the user according to Equation (1).

는 추가적인 감쇄 수치에 해당된다. 사용자 i의 데이터 처리량인 C_i 신호 대 잡음비인 SNR을 수학식 (2)에 대입하여 구할 수 있다.In Equation (1), P(Los) is the probability of a line-of-sight having no obstacles between communications, and P(NLos) is the probability that there is an obstacle between communications. Therefore, P(Los)+P(NLos)=1.

corresponds to the additional attenuation value. _{The SNR, which is the signal-to-noise ratio, C i} , which is the data throughput of user i, can be obtained by substituting it into Equation (2).

The Max-Min Airtime Fairness model is _{a model for evenly distributing the channel usage time (t j} ), ie, airtime, according to different communication requirements of users. Referring to Equation (3), the sum of the channel usage times for different communication requirements is set to be less than or equal to 1, so that the bandwidth that the unmanned aerial vehicle can provide can be divided according to the communication requirements.

는 사용자 j의 통신 처리량을 뜻한다. 따라서 수학식 (4)와 같이 모든 사용자에 대한 통신 처리량을 최대화할 수 있는 군집 무인 항공기 배치 위치를 구하는 것이 본 발명의 목적이 된다.From Equation (3)

is the communication throughput of user j. Therefore, as shown in Equation (4), it is an object of the present invention to obtain a swarm unmanned aerial vehicle deployment location that can maximize the communication throughput for all users.

The cluster unmanned aerial vehicle deployment apparatus 400 for maximizing communication throughput of the present invention includes a communication module unit 401, a user information unit 402, an unmanned aerial vehicle information unit 403, a clustering algorithm model unit 404, and an overload prediction algorithm model unit. 405 , a batch learning model unit 406 may be included. Hereinafter, specific functions of each component will be described.

The communication module unit 401 may receive information on a communication request amount of a user and a communication throughput related to a communication network service performed by the unmanned aerial vehicle. The communication module unit 401 may transmit information related to swarm unmanned aerial vehicle deployment for maximizing communication throughput to the swarm unmanned aerial vehicle. According to the transmitted deployment information, the swarm unmanned aerial vehicle can be deployed to an optimal position.

The user information unit 402 may store and analyze information about the communication request amount of the user received through the communication module unit 401 . The user information unit 402 may store and analyze the location information of the user. For example, the user's location information may include GPS-based location information. The user information unit 402 may periodically store the user's communication demand, and analyze the period and change trend of the change in the user's communication demand.

For example, the user information unit 402 may check the average change amount of the user communication demand for a certain period, and define users having a change amount higher than the average change amount as intensive analysis users.

The user information unit 402 may collect and analyze information about an area where communication traffic is instantaneously concentrated. For example, the user information unit 402 may check the average of the change amount of communication traffic in an arbitrary area unit for a certain period, and define a region having a change amount higher than the average change amount as the centralized analysis area.

The unmanned aerial vehicle information unit 403 may receive basic information of the unmanned aerial vehicle. The basic information of the unmanned aerial vehicle may include the navigable distance of the unmanned aerial vehicle, time, battery information, unmanned aerial vehicle model, manufacturing date, and the like. In addition, it may include information on maximum communication throughput that one unmanned aerial vehicle can provide, communication service provision coverage of one unmanned aerial vehicle, and the like.

In addition, based on the user communication requirement of the user information unit 402, it may include information on the unmanned aerial vehicle that can be used when the unmanned aerial vehicle is deployed in a communication area. The unmanned aerial vehicle information unit 403 may include all information that the unmanned aerial vehicle may have and is not limitedly interpreted.

The clustering algorithm model unit 404 performs a function of determining the optimal arrangement of the unmanned aerial vehicle cluster for maximizing the communication throughput. The clustering algorithm model unit 404 may utilize a clustering algorithm for grouping unlabeled data as a representative technique of unsupervised learning. Such a clustering algorithm includes a mean shift algorithm, a Gaussian mixture model (GMM) using a Gaussian distribution model, and a density-based spatial clustering of applications with noise (DBSCAN).

In the present invention, based on the K-means clustering algorithm, a modified K-means clustering algorithm that considers the user's communication requirements as an additional parameter is utilized.

The cost function used in the K-means clustering algorithm divides the data into several groups, identifies the central node of each group, and returns the sum of the squares of the distance between the central node of each group and the data to which each group belongs as a function. do. The function of cost can be expressed as Equation (5) above.

The K-means algorithm corresponds to an algorithm for grouping data to minimize a function of cost and finding the central node of each group. In the conventional K-means algorithm, when calculating the sum of the squares of the distances between data at the central node of a group, it is a model that does not consider the weights between each data. However, the clustering algorithm model unit 404 of the present invention utilizes the modified K-means algorithm model by further developing the K-means algorithm.

The modified K-means algorithm utilized in the clustering algorithm model unit 404 of the present invention is an algorithm that considers the user's communication demand as an additional parameter. The modified K-means algorithm will be described in detail using FIG. 3 .

3 is a diagram for explaining a modified K-means algorithm according to an embodiment of the present invention.

Referring to FIG. 3 , the modified K-means algorithm in consideration of the user's communication requirements according to the present invention can be confirmed. In an actual communication environment, a user's communication demand is not constant and may have a different value for each user. Therefore, when calculating the central node in the K-means algorithm, a method for calculating the central node of each cluster is needed in consideration of different communication requirements. In the present invention, Equation (6) can be used to consider the communication requirements of different users.

는 각 군집의 중앙 노드에서 각 데이터 간의 거리를 나타낸다. 따라서 수정된 거리값인

는 각 군집의 중앙 노드에서부터 각 데이터 간의 거리에 일정상수

와 사용자 요구량 d_i를 곱하여 계산할 수 있다. 따라서, 본 발명에서, 각 군집의 중앙 노드에서 각 군집의 데이터들 간의 거리값의 합을 최소로 하는 각 군집의 중앙 노드를 계산할 때, 각 데이터들 간의 일정 가중치를 부여하여, 각 군집의 중앙 노드를 계산할 수 있다. 본 발명에서 일정 가중치는 서로 다른 사용자의 통신 요구량이 될 수 있다. In Equation (6), d _i represents the communication demand of user i.

represents the distance between each data at the central node of each cluster. Therefore, the corrected distance value is

is a constant constant for the distance between each data from the central node of each cluster.

It can be calculated by multiplying by the user demand d _{i .} Therefore, in the present invention, when calculating the central node of each cluster that minimizes the sum of distance values between the data of each cluster at the central node of each cluster, a certain weight is given between each data, and the central node of each cluster is calculated. can be calculated. In the present invention, the predetermined weight may be communication requirements of different users.

는 각각 사용자 i의 위치, 사용자 i의 통신 요구량, 무인 항공기 j의 위치에 해당된다. 출력 데이터로,

는 각 각 무인 항공기 j의 속해 있는 사용자 집합, 무인 항공기 j의 통신 처리량 최대화를 위한 위치에 해당된다. 도 3을 참조하면

는 임의의 무인 항공기 j에 속해 있는 임시 사용자 군집에 해당된다. 수정된 K-means 알고리즘은

를 변경하면서

가 최적의 조건을 만족할 때,

로 출력한다. Referring to FIG. 3 , input data and output data of the modified K-means algorithm can be confirmed. as input data

are corresponding to the location of user i, the communication demand of user i, and the location of the unmanned aerial vehicle j, respectively. as output data,

Each corresponds to a set of users to which UAV j belongs, and a location for maximizing communication throughput of UAV j. 3 ,

corresponds to the temporary user cluster belonging to a random drone j. The modified K-means algorithm is

while changing

When the optimal condition is satisfied,

output as

As described above in Equation (2), the optimal condition is to obtain a corrected distance value based on the location and communication requirements of all users in the assigned cluster of any unmanned aerial vehicle j, and to obtain the corrected distance value for all unmanned aerial vehicles When is the time when the sum of At this time, the location of the central node of each cluster to obtain the corrected distance value of each cluster is set as the optimal placement location of the unmanned aerial vehicle for maximizing communication throughput, and this location is the optimal location value for maximizing the processing of users' different communication requirements. becomes this

The clustering algorithm model unit 404 may calculate a position value at which the communication throughput is maximized by using the modified K-means algorithm. The clustering algorithm model unit 404 may transmit the position value of each unmanned aerial vehicle to the communication module unit 401 and the communication module unit 401 may transmit communication information for the deployment of the unmanned aerial vehicle.

However, in the actual communication environment, these position values calculated by the clustering algorithm model unit do not take into account the communication processing limit of the unmanned aerial vehicle. Specifically, these position values are obtained under the assumption that the communication throughput of one unmanned aerial vehicle is infinite. However, in reality, there is a communication processing limit that one unmanned aerial vehicle can handle, and if the communication demand increases the communication processing limit, communication overload of one unmanned aerial vehicle may occur. Therefore, a method for preventing such communication overload will be described.

Returning to FIG. 2 , the swarm unmanned aerial vehicle deployment device 400 may further include an overload prediction algorithm model unit 405 . The overload prediction algorithm model unit 405 performs a function of relocating the unmanned aerial vehicle so that a communication overload state does not occur in any unmanned aerial vehicle as described above.

According to an embodiment of the present invention, the overload prediction algorithm model unit 405 utilizes an avoidance algorithm to prevent communication overload from occurring in any unmanned aerial vehicle based on the location value of the unmanned aerial vehicle extracted from the clustering algorithm model unit. Aircraft relocation results can be extracted. A detailed description of the avoidance algorithm is provided with reference to FIG. 4 .

4 is a diagram for explaining an avoidance algorithm according to an embodiment of the present invention.

Referring to FIG. 4 , after the unmanned aerial vehicle is disposed with respect to the position value extracted from the clustering algorithm model unit, a communication overload situation may occur. The communication overload situation refers to a state in which the communication throughput C _j of the arbitrary unmanned aerial vehicle j is lower than the communication demand of the user of the group to which the arbitrary unmanned aerial vehicle j belongs. The avoidance algorithm of the present invention may identify k, which is the closest unmanned aerial vehicle to any user i, for any user i among the clusters to which the arbitrary unmanned aerial vehicle j belongs to which communication overload has occurred.

Next, the avoidance algorithm can include any user i as a cluster to which the drone k belongs. Then, it can be repeated until the unmanned aerial vehicle j, which is overloaded with communication, can handle all of the communication demand. When any unmanned aerial vehicle j is able to handle the communication demand, the cluster of unmanned aerial vehicles is determined and the angle at which the sum of the corrected distance values is minimized by using the modified K-means algorithm of the clustering algorithm model part in the cluster. We can compute the central node of the cluster. The calculated central node of each cluster may be the location where the UAV is relocated.

은 회피 알고리즘의 출력값으로 임의의 무인 항공기 j의 통신 요구량을 처리할 수 있게 된 경우의, 통신 처리량이 최대가 되는 무인 항공기의 군집 및 최적의 배치에 관한 것이다. 이러한 출력값은 무인 항공기의 통신 처리 한계량까지 고려한 결과값이 된다. 4 of

It relates to the cluster and optimal arrangement of the unmanned aerial vehicle in which the communication throughput is maximized when it is possible to handle the communication demand of an arbitrary unmanned aerial vehicle j as the output value of the avoidance algorithm. These output values become the result values considering the communication processing limit of the unmanned aerial vehicle.

Therefore, the unmanned aerial vehicle deployment device of the present invention becomes a model in consideration of the communication overload situation of the unmanned aerial vehicle with respect to the unmanned aerial vehicle deployment for maximizing communication throughput. However, in the actual communication environment, the user's location and communication requirements are not fixed values, but constantly changing values. Therefore, a method for deploying swarm unmanned aerial vehicles to better cope with such an actual communication environment will be described.

Returning to FIG. 2 , the swarm unmanned aerial vehicle deployment device 400 may further include a deployment learning model unit 406 . The batch learning model unit 406 refers to the reinforcement learning algorithm modifying parameters in a direction to receive a large amount of reward while repeating actions set in a given environment. Therefore, reinforcement learning algorithms generally include three types of rewards for a given environment, behavior, and behavior.

There are many types of reinforcement learning algorithms, and they can be broadly classified into value-based and policy-based. The most representative policy-based ones are Trust Region Policy Optimization (TRPO) and Proximal Policy Optimization (PPO). The batch learning model unit 406 of the present invention may utilize the PPO.

In the batch learning model unit 406 of the present invention, a plurality of users with different communication requirements exist in the environment of the reinforcement learning algorithm, the action is the movement of the unmanned aerial vehicle, and the reward for the action is an increase in communication throughput after movement. The batch learning model unit 406 learns in a direction in which the communication throughput after moving is increased compared to the communication throughput before moving. Through this, it is possible to manage the deployment of the unmanned aerial vehicle so that the communication throughput can be increased through the reinforcement learning algorithm without an additional unmanned aerial vehicle deployment management manpower.

5 is a flowchart illustrating a method for deploying a swarm unmanned aerial vehicle for maximizing communication throughput according to an embodiment of the present invention.

Referring to FIG. 5 , the swarm unmanned aerial vehicle deployment method for maximizing communication throughput includes the steps of checking the user's location information (S101), the user's communication demand checking step of storing and analyzing the user's communication demand (S102), the swarm algorithm clustering users and calculating a central node for maximizing the communication throughput of each cluster (S103), determining whether communication overload occurs (S104), depending on whether communication overload occurs, if the overload does not occur It may include the step of disposing the unmanned aerial vehicle in the central node of each cluster (S105), and the step of relocating the unmanned aerial vehicle by using an avoidance algorithm when communication overload occurs (S106). In addition, it may further include a step (S107) of learning to deploy the unmanned aerial vehicle by using a reinforcement learning algorithm to maximize the communication throughput. The description of the swarm unmanned aerial vehicle deployment method for maximizing communication throughput overlaps with the description of the swarm unmanned aerial vehicle deployment device for maximizing the communication throughput described above in FIGS. 2 to 4 , a description of a specific embodiment is to be omitted.

6 is a flowchart illustrating an avoidance algorithm according to an embodiment of the present invention.

Referring to FIG. 6 , the avoidance algorithm identifies the overloaded drone and checks the location information of the unmanned aerial vehicle (S201), selects a random user belonging to the overloaded drone, and simulates the distance with the random user It may include a step of checking the unmanned aerial vehicle in a nearby location (S202), including a random user into the cluster of the unmanned aerial vehicle (S203), and determining whether the overloaded unmanned aerial vehicle is overloaded (S205). . In addition, if the unmanned aerial vehicle is still overloaded, steps S201 and S203 are repeated, and if not overloaded, a step S206 of calculating the central node of each cluster using a cluster algorithm proceeds. and relocating the unmanned aerial vehicle to the central node of each cluster (S207). A specific embodiment of the avoidance algorithm of the present invention overlaps with the description of the swarm unmanned aerial vehicle deployment apparatus described above with reference to FIGS. 2 to 4 , and a description of the specific embodiment will be omitted. Hereinafter, with reference to FIGS. 7 to 8, the experimental results on the performance of the unmanned aerial vehicle deployment device of the present invention will be described.

7 is a view showing the experimental results regarding the performance of the swarm unmanned aerial vehicle deployment device for maximizing communication throughput according to an embodiment of the present invention.

The experimental design for confirming the performance of the batch device of the present invention is as follows. The experiment sets up an environment in which 30 users with different communication requirements and 3 drones are deployed accordingly, and the positions of the initial three drones are placed at random locations in the [250m X 250m] area.

The altitude of the unmanned aerial vehicle is set to be at a position of 200 m, the optimal position to facilitate communication. The experiment confirms the result while changing the clustering algorithm used in the clustering algorithm model part to maximize the communication throughput of 30 users with different communication requirements.

Accordingly, FIG. 7(A) shows the result of deploying the unmanned aerial vehicle by using the general K-means algorithm that does not consider the user communication requirement in the clustering algorithm model unit. In (A) of FIG. 7 , it can be confirmed that User B and User C are gathered in an arbitrary area, and thus a situation in which the communication demand is very high in the corresponding area occurs. If a general K-means algorithm that does not consider the user's communication requirements is used, it can be confirmed that only one Drone B drone is deployed in an area with a very high communication demand.

On the other hand, the case of FIG. 7B shows the result of deploying the unmanned aerial vehicle using a modified K-means algorithm that considers users' different communication requirements. In the case of (B) of Figure 7, it can be seen that the two unmanned aerial vehicles Drone B and C are deployed in the area where the user's communication demand is high. Furthermore, referring to FIG. 7C , the communication throughput of Drone B can be checked. The communication throughput of Drone B when the user's communication requirements are not taken into account corresponds to about 111Mbps, but it can be confirmed that the communication throughput of Drone B is about 153Mbps when the user's communication requirements of the present invention are taken into consideration. This is an experimental result that can confirm that it is more effective in terms of communication throughput to distribute and process the communication throughput by distributing a plurality of unmanned aerial vehicles when the user communication demand increases rapidly in an arbitrary area.

8 is a diagram of an experimental result for confirming the performance of a swarm unmanned aerial vehicle deployment apparatus for maximizing communication throughput according to an embodiment of the present invention.

Referring to FIG. 8 , a case in which users having different communication requirements are arranged in arbitrary positions are set as Cases A to J. Accordingly, the communication throughput in the case of using the modified K-means algorithm of the present invention and the general K-means algorithm is compared. Cases C, D, F, and G correspond to cases in which users are more evenly distributed than in other cases, and in this case, both algorithms record similar communication throughput. On the other hand, in Cases A, B, E, H, I, and J, it can be seen that the result of using the modified K-means algorithm has higher communication throughput than the case where it is not. Furthermore, when the modified K-means algorithm of the present invention was used, the communication throughput from Cases A to J was uniformly confirmed without a relatively large change. It can be confirmed that the unmanned aerial vehicle deployment device of the present invention can flexibly cope with changes in various communication environments.

Embodiments of the present invention disclosed in the present specification and drawings are merely provided for specific examples to easily explain the technical content of the present invention and help the understanding of the present invention, and are not intended to limit the scope of the present invention. It will be apparent to those of ordinary skill in the art to which the present invention pertains that other modifications based on the technical spirit of the present invention can be implemented in addition to the embodiments disclosed herein.

Claims (8)

a communication module unit for receiving the user's location information and communication demand information and transmitting the unmanned aerial vehicle deployment information;
a user information unit for storing and analyzing the user's location information and communication demand information;
an unmanned aerial vehicle information unit that stores basic unmanned aerial vehicle information, including unmanned aerial vehicle communication throughput;
a clustering algorithm model unit for calculating a location where communication throughput is maximized through a clustering algorithm based on the user's location information and communication demand information; containing
Unmanned aerial vehicle deployment device for maximizing communication throughput.
The method of claim 1,
The clustering algorithm model unit
A position at which communication throughput is maximized is calculated by assigning a weight according to the size of the communication demand of the user.
Unmanned aerial vehicle deployment device for maximizing communication throughput.
The method of claim 1,
Further comprising an overload prediction algorithm model unit for relocating the unmanned aerial vehicle except for the communication overloaded arbitrary unmanned aerial vehicle through an avoidance algorithm when any unmanned aerial vehicle overloaded with communication occurs
Unmanned aerial vehicle deployment device for maximizing communication throughput.
4. The method of claim 3,
Using a reinforcement learning algorithm, further comprising a batch learning model unit that learns the placement position of the unmanned aerial vehicle that maximizes communication throughput
Unmanned aerial vehicle deployment device for maximizing communication throughput.
checking the user's location information;
checking the user's communication requirements;
calculating a central node of each cluster for which communication throughput is maximized by using a clustering algorithm;
deploying the unmanned aerial vehicle of the central node of each cluster; containing
How to deploy drones to maximize communication throughput.
6. The method of claim 5,
The step of calculating the central node is
A position at which communication throughput is maximized is calculated by assigning a weight according to the size of the communication demand of the user.
How to deploy drones to maximize communication throughput.
6. The method of claim 5,
When any unmanned aerial vehicle overloaded with communication occurs, further comprising the step of relocating the unmanned aerial vehicle except for the communication overloaded arbitrary unmanned aerial vehicle through an avoidance algorithm
How to deploy drones to maximize communication throughput.
6. The method of claim 5,
When any unmanned aerial vehicle overloaded with communication occurs, further comprising the step of relocating the unmanned aerial vehicle except for the communication overloaded arbitrary unmanned aerial vehicle through an avoidance algorithm
How to deploy drones to maximize communication throughput.

Images