KR20220140098A - Autonomous flight control method for UAVs using Swarm Intelligence - Google Patents

Abstract

The present invention relates to a method for controlling autonomous flight of a plurality of unmanned aerial vehicles using swarm intelligence, and more particularly, the manpower and repulsion of unmanned aerial vehicles in swarm flight by receiving position and speed information of other unmanned aerial vehicles, mission information and control constants, etc. , by making it possible to control in the direction of narrowing the distance between UAVs and the direction away from them, thereby enabling autonomous flight without central control of individual UAVs or separate leading UAVs, thereby reducing environmental risks such as earthquakes and fires. It relates to a method for controlling autonomous flight of multiple UAVs using swarm intelligence that enables more efficient performance of tasks such as searching for survivors in a major disaster situation.

Description

Autonomous flight control method for UAVs using Swarm Intelligence

The present invention relates to a method for controlling autonomous flight of a plurality of unmanned aerial vehicles using swarm intelligence, and more particularly, the manpower and repulsion of unmanned aerial vehicles in swarm flight by receiving position and speed information of other unmanned aerial vehicles, mission information and control constants, etc. , by making it possible to control in the direction of narrowing the distance between UAVs and the direction away from them, thereby enabling autonomous flight without central control of individual UAVs or separate leading UAVs, thereby reducing environmental risks such as earthquakes and fires. It relates to a method for controlling autonomous flight of multiple UAVs using swarm intelligence that enables more efficient performance of tasks such as searching for survivors in a major disaster situation.

In recent years, the occurrence of disasters such as fires, earthquakes, and floods is increasing due to industrial development and environmental factors.

In the event of such a disaster, a search operation for saving lives is performed. In the existing search operation, a military unmanned aerial vehicle is mainly used.

However, in the case of military UAVs, not only the operation cost is high, but also when it is necessary to perform a wide area search mission, using one or two military UAVs is inefficient in terms of time and accuracy in terms of time and accuracy. research is being actively conducted.

A typical system for operating a large number of small UAVs is a method using swarm intelligence, which is inspired by the intelligent movement that appears in groups of insects or animals. When multiple UAVs are operated at the same time, it is possible to perform various missions that cannot be performed with one or two UAVs, and since the effective mission area for each UAV is reduced, the mission execution time is shortened as well as some UAVs are damaged due to malfunctions, etc. Even if flight is impossible, the mission can be performed by the remaining UAVs, so the success rate of the mission increases, and UAVs equipped with different mission equipment such as electro-optical (EO)/infrared (IR) equipment, image radar (SAR) equipment, and communication repeaters Because they fly in groups, it has the advantage of being able to acquire various information.

As an example of such a system using a plurality of unmanned aerial vehicles, Republic of Korea Patent Publication No. 10-2010568 discloses a real 3D space search system and method using a plurality of unmanned aerial vehicle systems and swarm intelligence.

According to the principle of a swarm-based optimization algorithm for drones that are a plurality of unmanned aerial vehicles (UAV), the prior art provides the maximum acceleration, maximum speed, and location information of each drone of the swarm drone in real time. It is collected and shared by the swarm control server, and intelligently searches drone swarms under the control of a swarm control system equipped with a control computer and a swarm control server to which the swarm search algorithm is applied. It determines the next movement position of each drone to prevent collision by controlling it and has a technical feature in that it is configured to obtain search results quickly and accurately. As the number of drones used increases, there is a problem in that operational performance may deteriorate.

That is, in the prior art, the maximum acceleration, maximum speed, and location information of each drone are collected and shared with a swarm control server in real time, the swarm control server searches the drone swarm, and a 1:N method is performed in advance by remote control. By reflecting the maximum speed and maximum acceleration of each drone in the to control the flight of the drone by transmitting a control command including the next movement position, speed, and acceleration to The amount of computation that increases as the number of drones increases as the number of drones increases, because not only the risk of collision of drones must be calculated individually, but also the next positions of all drones must be individually controlled by calculating the positions of the drones at each point in time. As the number of drones increases rapidly, it is difficult to control drones that must be performed in real time, and the dependence on communication increases, so it cannot be used for missions in areas where communication is impossible or poorly performed, and autonomous flight of drones is also impossible. There is a problem.

1. Republic of Korea Patent Publication No. 10-2010568 (Registered on Aug. 07, 2019)

The present invention has been devised to solve the problems of the prior art as described above, and an object of the present invention is to convert the acceleration calculated through a swarm algorithm into path point information and calculate it as a path point command to enable autonomous flight of multiple UAVs. Accordingly, to provide an autonomous flight control method of multiple UAVs using swarm intelligence that enables efficient search for survivors by operating a plurality of UAVs in disaster situations with high environmental risks such as earthquakes and fires.

In addition, the present invention allows each UAV to form a formation with an autonomous behavioral pattern using information from nearby UAVs, so that the increase in the amount of computation due to the increase in the number of UAVs input to the mission can be minimized, and some UAVs fail the mission Yet another object is to provide an autonomous flight control method of multiple UAVs using swarm intelligence that can improve the robustness of mission performance since missions can be performed by the remaining UAVs.

In addition, the present invention can minimize central control, does not require a lead UAV, and does not require the location of each UAV within the formation to be determined, so that a formation capable of efficiently searching according to the search terrain can be freely formed. Another object is to provide an autonomous flight control method of multiple UAVs using intelligence.

The present invention for achieving the above objects,

In the flight control method of a plurality of UAVs using swarm intelligence, a data input step of receiving position and speed data of the UAV and inputting the data into a swarm algorithm; A calculation step, a waypoint derivation step of converting the acceleration of the UAV calculated by the swarming algorithm into waypoint information using a waypoint guide module, and a waypoint induction module converted by the waypoint guide module is used. and a path point command step of calculating a path point command for each UAV.

At this time, the position and speed data of the UAV that has been moved to the path point transmitted in the path point command step are input again into the swarm algorithm and are used to generate a new path point.

In addition, in the data input step, it is characterized in that the mission information of the UAV and the control constant and damping coefficient according to the mission information are additionally input to the swarm algorithm.

And, it characterized in that it further comprises a formation and maintenance step of forming and maintaining the formation of a plurality of UAVs themselves through the exchange of location and speed information by short-range communication between the UAVs.

Here, the formation and maintenance of the formation includes: a collision avoidance step of avoiding collisions with each other using relative distance and relative speed information between the UAVs; It is characterized in that it includes a mission boundary maintenance step to prevent out of the boundary.

At this time, in the collision avoidance step,

에 의해 무인기들이 자체적으로 대형을 형성하는 것을 특징으로 한다. (여기서,

는 i번째 무인기에 작용하는 다른 무인기와의 제어명령 합벡터, N은 운용되는 무인기의 개수, k는 제어상수, c는 댐핑계수,

는 i번째 무인기와 j번째 무인기 사이의 상대거리 및

는 무인기 사이의 대형 유지를 위한 기준거리를 의미하는 것임.)

It is characterized in that the UAVs form a formation by themselves. (here,

is the sum vector of control commands with other UAVs acting on the i-th UAV, N is the number of UAVs operated, k is a control constant, c is a damping factor,

is the relative distance between the i-th UAV and the j-th UAV and

is the reference distance for maintaining the formation between UAVs.)

In addition, the control constant (k) and the damping coefficient (c) are characterized in that the larger the formation speed of the UAVs and the higher the density is set.

In addition, in the mission boundary maintenance step,

에 의해 무인기들이 자체적으로 임무경계를 벗어나지 않는 것을 특징으로 한다. (여기서,

는 i번째 무인기와 임무경계 사이에 작용하는 제어명령 합벡터, N은 운용되는 무인기의 개수, k는 제어상수,

는 i번째 무인기와 임무경계 사이의 상대거리 및

는 무인기와 임무경계 사이의 충돌 회피를 위한 기준거리를 의미하는 것임.)

It is characterized in that the UAVs do not deviate from the mission boundary by themselves. (here,

is the control command sum vector acting between the i-th UAV and the mission boundary, N is the number of UAVs operated, k is the control constant,

is the relative distance between the i-th UAV and the mission boundary and

is the reference distance for collision avoidance between the UAV and the mission boundary.)

In addition, the collision avoidance step and the mission boundary maintenance step may be utilized for the acceleration calculation in the acceleration calculation step.

According to the present invention, the acceleration calculated through a swarm algorithm that can be mounted on individual UAVs is converted into waypoint information and calculated as a waypoint command to enable autonomous flight of multiple UAVs, and accordingly, environments such as earthquakes and fires It has an excellent effect of efficiently searching for survivors by operating multiple UAVs in a disaster situation with high enemy risk.

In addition, according to the present invention, by allowing each UAV to form a formation with an autonomous behavioral pattern using information from nearby UAVs, the increase in the amount of computation due to the increase in the number of UAVs used for a mission can be minimized, and some UAVs can perform missions. Since it is possible to perform the mission by the remaining UAVs even if it fails, it has the additional effect of improving the robustness of the mission.

In addition, according to the present invention, it is possible to minimize central control, and accordingly, it is possible to minimize the dependence on communication, and also has the effect of reducing the fatigue of the controller.

In addition, according to the present invention, it is possible to optimize the fluidity of the large-scale change process of the UAV according to the type of the mission to which a plurality of UAVs are input through the change of design variables including the control constant and the damping factor.

1 is a flowchart sequentially illustrating a method for controlling autonomous flight of a plurality of UAVs using swarm intelligence according to the present invention.
Figure 2 is a diagram conceptually showing the flight control integration algorithm used in the present invention.
3 is a view conceptually illustrating a method of generating a path point in the path point derivation step of the present invention shown in FIG. 1. FIG.
4 is a diagram conceptually illustrating Pursuit Guidance that can be used in the present invention.
5 is a view conceptually showing how the UAV maintains its own formation by the autonomous flight control method of a plurality of UAVs using swarm intelligence according to the present invention.
6 is a view showing initial conditions of a simulation of a relative distance change according to a control constant and a damping coefficient applied to the present invention.
7A and 7B are graphs showing the simulation results shown in FIG. 6 .

Hereinafter, preferred embodiments of the autonomous flight control method of a plurality of unmanned aerial vehicles using swarm intelligence according to the present invention will be described in detail with reference to the accompanying drawings.

Prior to explaining the present invention, data of the unmanned aerial vehicle used in the control method described below is transmitted to the main server of the computer by known wireless communication and analyzed and calculated by a known analysis program or operation program, or by itself in the unmanned aerial vehicle. It can be analyzed and calculated by the analysis and operation program provided, and the control command generated by the analysis and operation program is similarly transmitted to the UAV by well-known wireless communication or can be transmitted by wireless communication between UAVs. , the hardware or software system used in the present invention is not intended to be claimed as a right in the present invention, and detailed descriptions thereof will be omitted.

In addition, the present invention described below can be universally applied to the flight control method of various small unmanned aerial vehicles used in various industrial fields. The method will be described with reference to it.

In addition, in the present invention below, for convenience of explanation, it is assumed that the UAV is a particle model, and the movement of the UAV for forming and maintaining a group formation is assumed to be performed on a two-dimensional plane, assuming that the altitude is maintained.

1 is a flowchart sequentially illustrating an autonomous flight control method of a plurality of UAVs using swarm intelligence according to the present invention, FIG. 2 is a diagram conceptually illustrating an integrated flight control algorithm used in the present invention, and FIG. It is a diagram conceptually showing a method of generating a pathpoint in the pathpoint derivation step of the present invention shown, Fig. 4 is a diagram conceptually showing the Pursuit Guidance that can be used in the present invention, and Fig. 5 is a view showing the cluster intelligence according to the present invention. It is a diagram conceptually showing how the UAV maintains its own formation by the autonomous flight control method of a plurality of UAVs used, and FIG. 6 shows the initial conditions of the simulation of relative distance change according to the control constant and damping coefficient applied to the present invention. It is a drawing, and (a), (b) of FIG. 7 is a graph which shows the result of the simulation shown in FIG.

The present invention receives the position and speed information of other unmanned aerial vehicles, mission information, and control constants, and the attractive force and repulsive force of the unmanned aerial vehicles in group flight, that is, the direction of narrowing the distance between the unmanned aerial vehicles and the away direction. By enabling autonomous flight without central control of individual UAVs or a separate leading UAV, missions such as searching for survivors in disaster situations with high environmental risks such as earthquakes and fires can be performed more efficiently. It relates to a method for autonomous flight control of multiple UAVs using swarm intelligence (hereinafter referred to as 'flight control method').

That is, as shown in FIG. 2, the present invention calculates the manpower and repulsion force that allows the UAVs to perform missions in groups by receiving location information, mission information, and control constants of other UAVs as inputs using a clustering algorithm. In this way, the input and repulsive force used to control the UAV are output as an acceleration command, and the output acceleration command is converted into a path point and can be used for flight control of the UAVs.

The present invention, which will be described in detail below, may be implemented by the configuration of a main server (not shown) of a control center for controlling the UAVs, but is performed by a communication module and a calculation module mounted on the UAV itself so that the UAVs are autonomous. It is also possible to perform flight control such as aggregation and scattering.

More specifically, the flight control method according to the present invention includes a data input step (S10), an acceleration calculation step (S20), a waypoint induction step (S30) and a waypoint command step (S40) as shown in FIG. First, the data input step (S10) is a step of receiving location data and speed data of a plurality of UAVs to be put into a mission and inputting them into a swarm algorithm. It can be transmitted to nearby UAVs or transmitted to the main server of the control center by wireless communication to be used as input data for a swarm algorithm to be described later.

In this case, the information input to the swarm algorithm in the data input step ( S10 ) may further include mission information of the UAV and a control constant (k) and damping coefficient (c) according to the mission information, which will be described later. do it with

)를 산출하는 단계로, 상기 군집 알고리즘은 각 무인기로부터 전송된 위치 데이터를 통해 다른 무인기들과의 상대거리를 연산하고, 연산된 상대거리와 무인기의 속도 데이터를 이용하여 무인기들의 가속도(

)를 출력으로 산출할 수 있다.Next, in the step of calculating the acceleration (S20), the acceleration (

), the swarm algorithm calculates the relative distance with other UAVs through the location data transmitted from each UAV, and uses the calculated relative distance and the speed data of the UAVs to accelerate the UAVs (

) can be calculated as an output.

At this time, in the acceleration calculation step (S20), the acceleration of the UAVs can be calculated by reflecting the calculation results in the collision avoidance step (S52) and the mission boundary maintenance step (S54) of the formation and maintenance step S50 to be described later. More details on this will be described later.

)를 경로점 유도 모듈(Waypoint Guidance)을 이용하여 경로점(Waypoint)으로 전환시키는 역할을 하는 것으로, 상기 경로점은 무인기의 다음 목표지점, 즉 무인기가 이동해야 할 지점이 될 수 있다.Next, in the path point induction step (S30), the acceleration of the UAV calculated by the cluster algorithm in the acceleration calculation step (S20)

) to a waypoint using a waypoint guidance module (Waypoint Guidance), and the waypoint may be the next target point of the UAV, that is, a point to which the UAV should move.

In more detail, the acceleration of the UAV output by the swarm algorithm can be converted to a path point by a command module as shown in FIG. The acceleration direction is calculated by dividing the acceleration of the UAV output by the x and y direction components.

... (1)

... (One)

는 가속도의 방향을 의미하는 것으로, 상기 (1)식에 의해 연산된 가속도의 방향을 무인기가 이동해야 할 방향으로 정한다.Here, a _x and a _y mean the x and y direction components of the acceleration, respectively,

denotes the direction of acceleration, and the direction of acceleration calculated by Equation (1) above is determined as the direction in which the UAV should move.

As shown in Equation (2) below, it is possible to generate a path point separated by K in the direction of acceleration from the current position by multiplying the direction calculated by Equation (1) by the gain K.

... (2)

... (2)

Here, x and y are x and y coordinates corresponding to the current location of the UAV, respectively, and x _wp and y _wp are x and y coordinates of the calculated path point, respectively.

At this time, when the acceleration is 0, the x,y coordinates of the path point become (0,0), so a phenomenon in which the UAV continues to move to the origin of the coordinate system may occur. A path point separated by K in the direction of acceleration may be generated.

Next, the waypoint command step (S40) is a step of allowing the UAVs to move toward the waypoint by calculating a waypoint command (Waypoint CMD) to each UAV using the waypoint generated in the waypoint derivation step (S30). , such a route point command may be transmitted through short-range wireless communication between the UAVs, or may be transmitted to each UAV through wireless communication through the main server of the control center.

More specifically, in the path point command step S40, a path point command can be calculated by a Pursuit Guidance method, which is a guide method mainly used in spherical guided missiles, and is a velocity vector of a guided missile. It is a method of controlling the missile so that it always has the direction of the target, that is, the line of sight.

를 갖고, 목표 경로점 방향으로의 목표 방위각(desired heading angle)

를 갖는다.In the present invention, the tracking guidance method is used for the purpose of allowing the UAV to follow the path point command.

, and the desired heading angle in the direction of the target path point.

has

The equation of motion for applying the tracking induction method in the two-dimensional Cartesian coordinate system is as Equation (3) below.

... (3)

... (3)

)는 아래의 (4)식에 나타낸 바와 같이, 속도 오차에 대하여 게인 K_T를 곱해주는 방식으로 구할 수 있다.And, the magnitude of the acceleration (

) can be obtained by multiplying the speed error by the gain K _T as shown in Equation (4) below.

)은 무인기의 속도 V로 나눈 방위각 오차에 대해 게인 K_P를 곱해주는 방식으로 구할 수 있다.In addition, if V _D is the maximum speed of the UAV, the direction of acceleration (

) can be obtained by multiplying the azimuth error divided by the speed V of the drone by the gain K _P .

... (4)

... (4)

In this case, K _T and K _P have units of 1/time and acceleration, respectively, and may affect the reactivity of the UAV moving along the path point by the tracking induction method.

Accordingly, in the path point command step (S40), the path point command calculated by the tracking induction method using the path point information switched in the path point induction step (S30), that is, the acceleration in the path point direction, is transmitted to each UAV. This allows the UAVs to fly towards the route point.

The position and velocity data of the UAV whose movement to the path point has been completed by the path point command as described above is input again into the clustering algorithm, and can be used to calculate a new acceleration and accordingly generate a path point and calculate the path point command.

On the other hand, the flight control method according to the present invention may further include a formation and maintenance step (S50) of the formation, the formation and maintenance step (S50) of the large number of unmanned aerial vehicles in the mission position information by short-range wireless communication with each other. And by exchanging speed information to form and maintain a formation of a group state by itself, it may include a collision avoidance step (S52) and a mission boundary maintenance step (S54).

In other words, when performing a search mission using swarm intelligence, a large number of UAVs form a formation to perform the mission. is accompanied

Therefore, in the present invention, a virtual force based on local information shared by short-distance wireless communication between UAVs, that is, location and speed information of UAVs, is applied to allow UAVs to form or maintain a formation. A more efficient formation can be formed according to the search terrain and types of missions because the leading UAV is unnecessary and the positions of individual UAVs within the formation are not determined.

This large formation and maintenance step (S50) is performed by the communication module and operation module provided in each UAV, so that the UAVs can be configured to control for the formation and maintenance of the large size by themselves without separate central control. A control command acting on these UAVs may be made by an acceleration command of the UAV having the size and direction as described above.

)을 생성할 수 있다.In the formation and maintenance of the formation (S50), the behavior of the UAVs is simplified in order to minimize the computational burden of individual UAVs. and calculating the relative distance between the unmanned _aerial vehicle and the mission boundary, and as shown in FIG. 5, a control command (

) can be created.

More specifically, the collision avoidance step (S52) is a step of generating a control command to avoid collision with each other using the relative distance and relative speed information between the UAVs. When the relative distance between the UAVs calculated using the information and speed information is greater than or equal to the preset reference distance (d _ref ) between the UAVs, the relative distance between UAVs approaches, that is, in the direction in which the UAVs pull each other. A control command may be generated, and when the relative distance between the UAVs is less than the reference distance d _ref , the control command may be generated in a direction to push the UAVs to each other.

Also, in the mission boundary maintenance step (S54), the relative distance information between the UAV and the mission boundary calculated using the location information and speed information of the UAVs shared by short-range wireless communication and the location information of the preset mission boundary is similarly performed. This is a step of generating a control command to prevent the _UAV from escaping the mission boundary using the Control commands can be generated in the direction away from the boundary or in the direction in which the mission boundary pushes the UAV.

In more detail, assuming the UAV as a particle model and the motion of the UAV as motion on a two-dimensional plane where altitude is maintained, the equation of motion of the UAV can be expressed as Equation (5) below.

... (5)

... (5)

는 각각 다수의 무인기 중 i번째 무인기의 위치벡터, 속도벡터 및 제어명령을 의미한다.here,

Each of the plurality of UAVs means the position vector, the speed vector and the control command of the i-th UAV.

In addition, the relative distance vector and the relative velocity vector between the UAVs can be expressed as Equation (6) below.

... (6)

... (6)

와

는 각각 i번째 무인기와 j번째 무인기 사이의 상대거리벡터와 상대속도벡터를 나타낸다.here,

Wow

denotes the relative distance vector and the relative velocity vector between the i-th UAV and the j-th UAV, respectively.

)의 생성에 사용되는데, 상기 제어명령(

)은 아래의 (7)식에 의해 정의될 수 있다.The relative distance and relative speed calculated in Equation (6) above are the control commands (

) is used to generate the control command (

) can be defined by Equation (7) below.

... (7)

... (7)

는 i번째 무인기에 작용하는 다른 무인기와의 제어명령 합벡터이고, N은 운용되는 무인기의 개수, k는 제어상수, c는 댐핑계수,

는 i번째 무인기와 j번째 무인기 사이의 상대거리 및

는 무인기 사이의 대형 유지를 위한 기준거리, 즉 대형이 형성될 때 무인기 사이에 유지되어야 하는 거리를 의미하는 것이다.here,

is the control command sum vector with other UAVs acting on the i-th UAV, N is the number of UAVs operated, k is a control constant, c is a damping coefficient,

is the relative distance between the i-th UAV and the j-th UAV and

is the reference distance for formation maintenance between UAVs, that is, the distance to be maintained between UAVs when formations are formed.

)는 운용하고자 하는 무인기의 시야각(FOV; Field Of View)과, 무인기 사이에 충돌하지 않기 위한 여유거리(margin) 및 정보공유를 위해 공유하는 다른 무인기 위치정보의 불확실성 등을 고려하여 결정될 수 있다.At this time, the reference distance (

) can be determined in consideration of the field of view (FOV) of the unmanned aerial vehicle to be operated, the margin to avoid collision between the unmanned aerial vehicles, and the uncertainty of other unmanned aerial vehicle location information shared for information sharing.

In addition, the control constant (k) and the damping coefficient (c) are a kind of design variable and can be set according to the nature of the task to be performed by the UAVs, which will be described later.

은 자기자신을 제외한 모든 무인기들과 작용하는 밀어내는 제어명령(

)과 잡아당기는 제어명령(

)의 합벡터로, 전술한 바와 같이 이웃하는 다른 무인기와의 상대거리가 기준거리(

) 보다 짧을 경우에는 충돌을 방지하기 위해 밀어내는 제어명령(

)이 작용하고, 이웃하는 다른 무인기와의 상대거리가 기준거리(

) 이상인 경우에는 대형을 유지하기 위해 잡아당기는 제어명령(

)이 작용할 수 있다.Control command of the unmanned aerial vehicle generated in the collision avoidance step (S52)

is a push control command (

) and pulling control commands (

), as described above, the relative distance with other UAVs is the reference distance (

) is shorter than the control command (

) acts, and the relative distance with other neighboring UAVs is the reference distance (

), the control command to pull to maintain the formation (

) may work.

On the other hand, the mission boundary, which is the boundary line of the area where the UAV performs the mission, is assumed to be a curve and can be expressed as a curve equation as in Equation (8) below, and the order of the curve is determined in the mission boundary maintenance step (S54) Control commands can be generated.

... (8)

... (8)

Here, a _n , a _n-1 ,..., a ₀ are the equation coefficients of the curve of the mission boundary assumed as the curve.

Based on Equation (8), the shortest distance between the i-th UAV and the mission boundary by Equation (9) below through It is possible to calculate the minimum value when

... (9)

... (9)

는 i번째 무인기와 임무경계 사이의 최단 위치벡터이고,

는 최단위치에 해당하는 임무경계의 위치좌표이며, d_ib는 최단위치에서의 i번째 무인기와 임무경계 사이의 상대거리이고, x_i, y_i는 최단위치에서의 iㅂ번째 무인기의 위치좌표이다.here,

is the shortest position vector between the i-th UAV and the mission boundary,

is the position coordinates of the mission boundary corresponding to the lowest unit value, d _ib is the relative distance between the i-th UAV and the mission boundary at the lowest unit value, and x _i , y _i are the position coordinates of the i-th UAV at the lowest unit value. .

)을 산출하는데 사용되는데, 상기 임무경계 유지단계(S54)에서의 제어명령(

)은 아래의 (10)식과 같이 정의될 수 있다.The relative distance (d _ib ) between the mission boundary and the UAV derived by Equation (9) above is a control command (

) is used to calculate the control command (

) can be defined as Equation (10) below.

... (10)

... (10)

는 각 무인기와 임무경계 사이에 작용하는 제어명령의 합벡터이고, N은 운용되는 무인기의 개수, k는 제어상수, d_ib는 임무경계와 무인기 사이의 상대거리 및

는 무인기와 임무경계의 충돌을 방지하기 위한 무인기와 임무경계 사이의 기준거리를 뜻하는 것으로, 무인기와 임무경계 사이에는 오직 충돌을 방지하기 위한 밀어내는 제어명령만 작용할 수 있다.here,

is the sum vector of control commands acting between each UAV and the mission boundary, N is the number of operated UAVs, k is the control constant, d _ib is the relative distance between the mission boundary and the UAV and

is the reference distance between the UAV and the mission boundary to prevent collision between the UAV and the mission boundary.

)은 상기 (10)식에서 확인할 수 있는 바와 같이, 무인기와 임무경계 사이의 상대거리(d_ib)가 기설정된 기준거리(

) 보다 긴 경우에는 아무런 제어명령(

)도 작용하지 않고, 오직 무인기와 임무경계 사이의 상대거리(d_ib)가 기설정된 기준거리(

) 이하인 경우에만 임무경계와의 충돌 회피를 위한 임무경계로부터 밀어내는 제어명령(

)을 적용하게 된다.That is, the control command in the mission boundary maintenance step (S54) (

) is the reference distance ( d _ib ) between the UAV and the mission boundary is preset

), no control command (

) does not work, only the relative distance (d _ib ) between the UAV and the mission boundary is the preset reference distance (

) or less, the control command to push from the mission boundary to avoid collision with the mission boundary (

) will be applied.

On the other hand, the design variables used in the collision avoidance step (S52) and the mission boundary maintenance step (S54), that is, the control constant (k) and the damping coefficient (c) should be set according to the nature of the mission to which a plurality of UAVs are deployed. In order to set these design variables, the change in the relative distance between the UAVs according to the design variables can be calculated through simulation.

) 만큼의 거리에서 정지하지 않고, 기준거리(

)에 근사한 위치에서 진동하며 거리를 유지하였고, 기준거리(

)로부터 ±0.002m에서 진동하며 상대거리(d_ij)가 늘어났다 줄어들었다 하는 특성을 보여 기준거리(

)로부터 처음 ±0.002m인 지점을 수렴시점으로 가정하였다.In more detail, the relative distance change according to the control constant (k) and the damping coefficient (c) was measured under the simulation conditions shown in FIG. 6 and below (Table 1). (

) without stopping at a distance as much as the reference distance (

) vibrated at a position close to the distance and maintained the distance, and the reference distance (

), it vibrates at ±0.002 m, and the relative distance (d _ij ) increases and decreases, showing that the reference distance (

), the first point ±0.002 m from the convergence point was assumed.

) 도달정도

로 나타낸 것을 도 7의 (a)에 도시하였다.First, the relative distance (d _ij ) was measured while the control constant (k) was changed to 1, 3, 5, 10, 20 while the damping coefficient (c) was fixed at 0.5. The reference distance of the unmanned aerial vehicle (i, j) (

) degree of reach

It is shown in Fig. 7 (a).

)에 도달하는 속도가 빨랐지만, 다시 멀어지려는 제어명령(

) 또한 커서 진동폭이 크게 나타났다.In addition, the convergence time is shown as shown in (Table 2) below. As the control constant (k) increases, the relative distance (d _ij ) between the two UAVs becomes the reference distance (

) was reached quickly, but the control command (

), and the amplitude of vibration was large.

) 도달정도

로 나타낸 것을 도 7의 (b)에 도시하였다.Next, the relative distance (d _ij ) was measured while the damping coefficient (c) was changed to 0.1, 0.3, 0.5, 0.7, and 0.9, respectively, in a state where the control constant (k) was fixed at 5, and the result was similarly measured with time The reference distance of the two UAVs (i, j) according to

) degree of reach

It is shown in FIG. 7(b).

)에 도달한 시간은 차이가 없었지만, 댐핑상수(c)가 증가할수록 진동폭이 점점 작아져 결국 상대거리(d_ij)가 기준거리(

)에 수렴하는 시간은 점점 빨라지는 것을 확인할 수 있다.At this time, the convergence time is shown as shown in Table 3 below. Since the control constant (k) is fixed, as the damping constant (c) increases, the relative distance (d _ij ) between the two UAVs is the first reference distance (

There was no difference in the time to reach ), but as the damping constant (c) increased, the amplitude of vibration became smaller and, eventually, the relative distance (d _ij ) became the reference distance (

), it can be seen that the convergence time is getting faster.

As a result of the above simulation, when the UAV needs to form a formation with high density and fast, the control constant (k) and damping coefficient (c) are set high. It can be seen that the control constant (k) and damping coefficient (c) should be set low. It may be set and input to the swarm algorithm together with the location and speed data of the UAV in the aforementioned data input step (S10).

)과 잡아당기는 제어명령(

) 및 임무경계 유지단계(S54)에서의 밀어내는 제어명령(

)이 혼합된 합벡터 방향으로 최종 제어명령, 즉 가속도를 산출할 수 있고, 이와 같이 산출된 가속도가 경로점 유도단계(S30)에서 경로점으로 전환되어 경로점 명령단계(S40)에서 각 무인기로의 제어명령으로 생성될 수 있다.The control commands in the collision avoidance step (S52) and the mission boundary maintenance step (S54) as described above may be utilized for the acceleration calculation in the acceleration calculation step (S20). In the cluster algorithm, the collision avoidance step (S52) ) in the push-out control command (

) and pulling control commands (

) and the pushing control command in the mission boundary maintenance step (S54) (

) can calculate the final control command, ie, acceleration, in the direction of the mixed sum vector, and the calculated acceleration is converted to a path point in the path point induction step (S30), and is then transferred to each UAV in the path point command step (S40). It can be created with a control command of

In addition, as described above, the operation module of each UAV includes a swarm algorithm, a command module, and a Pursuit Guidance so that UAVs can generate control commands by themselves and fly autonomously. , thereby minimizing central control and minimizing dependence on communication.

On the other hand, in order to confirm the large change due to the movement of multiple UAVs according to the flight control method according to the present invention, the control constant (k) and the damping coefficient ( c) was changed and a total of three simulations were performed.

Overall, regardless of the size of the control constant (k) and damping coefficient (c), the phenomenon of aggregation and collision avoidance of unmanned aerial vehicles with random positions and speeds through control commands repeatedly appears, eventually resulting in one large It was possible to confirm through the simulation that it formed and flew.

In addition, in the simulation in which both the control constant (k) and the damping coefficient (c) are set to be large, the speed of aggregation between UAVs is fast and the formation is less disturbed when it collides with the mission boundary, while the control constant (k) And in the simulation in which the size of the damping coefficient (c) is set small, the formation is formed slowly due to the slow aggregation speed between the UAVs at the beginning, and the phenomenon that the formation is disturbed and re-formed during mission boundary or collision avoidance with other UAVs seemed

In addition, in the simulation in which the size of the control constant (k) is set large and the size of the damping coefficient (c) is set small, the phenomenon that UAVs gather quickly at the beginning, but fail to form a formation at once, and spread again and then quickly agglomerate. could check

Through the simulation results as described above, it is possible to determine the fluidity of the large-scale change process of the UAV by adjusting the values of the design variables, i.e., the control constant (k) and the damping factor (c), according to the nature of the mission using a plurality of UAVs. Able to know.

Therefore, according to the flight control method according to the present invention as described above, autonomous flight of multiple UAVs is possible by converting the acceleration calculated through a swarm algorithm that can be mounted on individual UAVs into waypoint information and calculating it as a waypoint command. Therefore, in disaster situations with high environmental risks such as earthquakes and fires, multiple UAVs can be operated to efficiently search for survivors, and each UAV can use information from nearby UAVs to autonomously act. By forming a formation, the increase in the computational amount due to the increase in the number of UAVs used for the mission can be minimized, and even if some UAVs fail their mission, the remaining UAVs can perform the mission, so the robustness of the mission can be improved. In addition, it is possible to minimize the central control, so it is possible to minimize the dependence on communication and the fatigue of the controller. Accordingly, it has various advantages such as optimizing the fluidity of the large-scale change process of the UAV.

Although the above-described embodiments have been described with respect to the most preferred example of the present invention, it is not limited to the above embodiment, and the flight control method according to the present invention is performed in the main server of the central control center to control the flight of a plurality of UAVs. It will be apparent to those skilled in the art that various modifications can be made without departing from the technical spirit of the present invention.

The present invention relates to a method for controlling autonomous flight of a plurality of unmanned aerial vehicles using swarm intelligence, and more particularly, the manpower and repulsion of unmanned aerial vehicles in swarm flight by receiving position and speed information of other unmanned aerial vehicles, mission information and control constants, etc. , by making it possible to control in the direction of narrowing the distance between UAVs and the direction away from them, thereby enabling autonomous flight without central control of individual UAVs or separate leading UAVs, thereby reducing environmental risks such as earthquakes and fires. It relates to a method for controlling autonomous flight of multiple UAVs using swarm intelligence that enables more efficient performance of tasks such as searching for survivors in a major disaster situation.

S10: data input step S20: acceleration calculation step
S30: path point induction step S40: path point command step
S50: Formation and maintenance phase S52: Collision avoidance phase
S54: Mission boundary maintenance phase

Claims (9)

In the flight control method of multiple UAVs using swarm intelligence,
a data input step of receiving the location and speed data of the unmanned aerial vehicle and inputting it into a swarm algorithm;
an acceleration calculation step of calculating the acceleration of the UAVs using the data input to the swarm algorithm;
a waypoint induction step of converting the acceleration of the UAV calculated by the swarm algorithm into waypoint information using a waypoint guide module (Waypoint Guidance); and
and a path point command step of calculating a path point command of each UAV using the path point switched by the path point induction module.
The method of claim 1,
Autonomous flight control of multiple UAVs using swarm intelligence, characterized in that the position and speed data of the UAV that has been moved to the route point transmitted in the route point command step are input back into the cluster algorithm and used to generate a new route point Way.
The method of claim 1,
In the data input step, the mission information of the UAV and the control constant and damping coefficient according to the mission information are additionally input to the cluster algorithm.
The method of claim 1,
Autonomous flight of multiple UAVs using cluster intelligence, characterized in that it further comprises a formation and maintenance step in which a plurality of UAVs form and maintain a formation on their own through the exchange of location and speed information by short-range communication between the UAVs. control method.
5. The method of claim 4,
The large formation and maintenance step is
A collision avoidance step of avoiding collisions with each other by using the relative distance and relative speed information between the UAVs;
and a mission boundary maintenance step of preventing the unmanned aerial vehicle from departing from the mission boundary by using the relative distance information between the unmanned aerial vehicle and the mission boundary.
6. The method of claim 5,
In the collision avoidance step,

Autonomous flight control method of multiple UAVs using swarm intelligence, characterized in that the UAVs form their own formation by
(here,

is the sum vector of control commands with other UAVs acting on the i-th UAV, N is the number of UAVs operated, k is a control constant, c is a damping factor,

is the relative distance between the i-th UAV and the j-th UAV and

is the reference distance for maintaining the formation between UAVs.)
7. The method of claim 6,
The control constant (k) and the damping coefficient (c) are set larger as the formation speed of the UAVs increases and the density increases.
6. The method of claim 5,
In the mission boundary maintenance step,

Autonomous flight control method of multiple UAVs using swarm intelligence, characterized in that the UAVs do not deviate from the mission boundary by themselves.
(here,

is the control command sum vector acting between the i-th UAV and the mission boundary, N is the number of UAVs operated, k is the control constant,

is the relative distance between the i-th UAV and the mission boundary and

is the reference distance for collision avoidance between the UAV and the mission boundary.)
6. The method of claim 5,
The method for controlling autonomous flight of a plurality of UAVs using swarm intelligence, characterized in that the collision avoidance step and the mission boundary maintenance step can be utilized for the acceleration calculation in the acceleration calculation step.

Images