KR102279956B1 - 3D optimal surveillance trajectory planning Method and Apparatus for multi-UAVs using particle swarm optimization with surveillance area priority - Google Patents

Abstract

Provided are a method and device for optimizing three-dimensional surveillance trajectory setting of multiple unmanned aerial vehicles (UAV) using particle swarm optimization theory based on surveillance area priority. The method for optimizing three-dimensional surveillance trajectory setting of multiple unmanned aerial vehicles (UAV) using particle swarm optimization theory based on surveillance area priority provided in the present invention comprises: a step of deriving an optimal trajectory for each monitoring drones (MDr) using particle swarm optimization (PSO) theory; a step of obtaining the best flight waypoint sequence from all possible trajectories using a multipurpose suitability function; a step of detecting the presence of a collision using coordinated agents coevolution (CAC); a step of generating an optimal trajectory for one flight time if CAC is satisfied; and a step of updating a surveillance area importance (SAI) value prior to the next flight.

Description

3D optimal surveillance trajectory planning method and Apparatus for multi-UAVs using particle swarm optimization with surveillance area priority

The present invention relates to a method and apparatus for planning a 3D optimal security monitoring path for a plurality of UAVs using a minority optimization theory.

Unmanned Aerial Vehicles (UAVs) have been used for decades in both military and civilian applications in areas such as remote sensing, sensor data acquisition, UAV-enabled communications, relays for ad hoc networks, disaster monitoring, flood area monitoring, and wildfire tracking. is becoming more and more important over time. However, the misuse of drones poses a serious threat to public safety due to the lack of proper legislation. Monitoring Drones (MDr) are usually battery-operated, so the possible monitoring path of the MDr must be carefully designed, taking into account physical constraints. The problem of UAV route planning as a fundamental element of unmanned aerial vehicle autonomous control modules has been studied for decades. It can be formulated as an optimization problem to find a feasible route from the origin to the destination. And an optimal path is usually associated with a path that maximizes (or minimizes) a particular optimization index (eg, energy consumption, path length, etc.) of a particular task.

In order to derive the optimal route, many researchers have proposed numerous route planners. In the prior art, a path planner was proposed based on a simulation annealing algorithm to obtain an almost optimal path in a two-dimensional (2D) radar constraint environment. In another prior art, an optimal UAV path for collecting sensing data from a wireless sensor network was derived by utilizing the A* algorithm. However, such a proposal (designed for a 2D environment) is difficult to apply to the 3D operating space because more constraints need to be modeled to obtain the optimal path. For 3D path planning, algorithms such as D*, RRT (Rapidly exploring Random Tree), bio-inspired algorithm, or EA (Evolutionary Algorithm) are used. Since finding an optimal solution to a path problem is a non-deterministic polynomial time-complete problem, EA is optimal due to its advantages when dealing with very complex 3D path planning problems. Another prior art employs a state-of-the-art variant of the Differential Evolutionary (DE) algorithm to solve the 3D path planning problem. For multiple UAV route planning, Besada-Portas is based on multiple coordinated agents coevolution EA (MCACEA) with 11 optimization indices and constraints in the optimization criteria. Suggested route planner. Another prior art proposed a Variable Neighborhood Descend (VND)-enhanced particle swarm optimization (PSO) path planner for multiple UAVs in an agricultural application scenario, but only operating time and path length were considered as optimization targets.

An object of the present invention is to provide a security monitoring area importance update mechanism for deriving a three-dimensional optimal monitoring path for a plurality of surveillance drones and a path planner based on a small swarm optimization theory. In addition, we propose a multi-purpose fitness function according to energy consumption, flight risk and monitoring area priority to evaluate the route generated by the proposed route planner.

In one aspect, the 3D monitoring route setting optimization method of multiple unmanned aerial vehicles (UAV) using the monitoring area priority-based small swarm optimization theory proposed in the present invention uses the Particle Swarm Optimization (PSO) theory. to derive the optimal route for each monitoring drone (MDr), obtain the best flight stop sequence from all possible routes using a multi-purpose fitness function, and collide using CAC (Coordinated Agents Coevolution) detecting the presence of , generating an optimal route for one flight time if the CAC is satisfied, and updating the Surveillance Area Importance (SAI) value before the next flight.

The step of deriving the optimal path for each monitoring drone using the small swarm optimization theory is to generate a velocity vector and a position vector for each monitoring drone, and to determine the particle size through PSO using the velocity vector and position vector. Determining individual peaks and peaks of all entities, and improving the speed and position of each entity through individual peaks and peaks of all entities to derive an optimal route, monitoring drones to be applicable to autonomous flying devices An optimal path is derived to prevent collisions between the two.

Before the next flight, the step of updating the SAI value is a step of increasing the SAI value when the presence of illegal drones (IDr) is detected in the area (cell) where the MDr event is detected. increasing the SAI value for the rapidly changing area compared to the SAI value and decreasing the SAI value for the previously monitored area to avoid repeated monitoring for the same area.

When the presence of IDr is detected in the cell in which the MDr event is detected, the step of increasing the SAI value is when IDr is detected in a specific cell, monitoring neighboring cells to prevent IDr from entering the area, and Increases the SAI value of a cell and its neighboring cells for that cell.

In order to avoid repeated monitoring for the same area, the step of reducing the SAI value for the previously monitored area increases the SAI value for the cell when the cell is not monitored for more than a predetermined flight time, and For cells, adjust the cell's SAI value to reduce the SAI value.

In another aspect, the apparatus for optimizing the three-dimensional surveillance route setting of multiple unmanned aerial vehicles (UAV) using the surveillance area priority-based prime swarm optimization theory proposed in the present invention is based on the Particle Swarm Optimization (PSO) theory. Using CAC (Coordinated Agents Coevolution), a prediction unit that derives the optimal route for each monitoring drone (MDr) and obtains the best flight stop sequence from all possible routes using a multi-purpose fitness function is used. It includes a detector that detects the presence of collisions by doing this, and generates an optimal route for one flight time if CAC is satisfied, and an update unit that updates the Surveillance Area Importance (SAI) value before the next flight.

According to embodiments of the present invention, it is possible to derive an optimal monitoring path for a plurality of surveillance drones to monitor a specific operational area and to detect the presence of illegal drones. Using the proposed path planner, the optimal path can be obtained from all possible paths according to the proposed fit function. In the proposed multi-purpose fitness function, not only Energy Consumption (EC) but also UAV maneuverability, flight risk and monitoring area priority are considered as cost determinants, making the proposed approach different from many other approaches in UAV route planning. has

1 is a diagram for explaining an operation scenario according to an embodiment of the present invention.
2 is a diagram illustrating an operating area and a path of an MDr according to an embodiment of the present invention.
3 is a diagram illustrating a representation of the terrain being monitored according to an embodiment of the present invention.
4 is a diagram for explaining UAV mobility parameters according to an embodiment of the present invention.
5 is a flowchart for explaining a method for planning a 3D optimal security monitoring path for a plurality of UAVs using a minority swarm optimization theory according to an embodiment of the present invention.
6 is a diagram illustrating the configuration of an apparatus for planning a 3D optimal security monitoring path for a plurality of UAVs using a minority swarm optimization theory according to an embodiment of the present invention.

The use of unmanned aerial vehicles (UAVs) has been considered a promising technology for both military and civilian applications. However, the misuse of illegal drones poses a serious threat to social safety as the relevant laws and regulations are not fully in place. In the present invention, we propose a security monitoring area importance update mechanism and a path planner based on a small group optimization theory to derive a three-dimensional (3D) optimal monitoring path for a plurality of surveillance drones. We also propose a multi-purpose fitness function according to energy consumption, flight risk and monitoring area priority to evaluate the route generated by the proposed route planner.

According to embodiments of the present invention, it is possible to derive an optimal monitoring path for a plurality of surveillance drones to monitor a specific operational area and to detect the presence of illegal drones. To solve this problem, we propose a route planner according to PSO and monitoring area priority. Additionally, it extends route planning into a 3D environment. Using the proposed path planner, the optimal path can be obtained from all possible paths according to the proposed fit function. In the proposed multi-purpose fitness function, not only Energy Consumption (EC) but also UAV maneuverability, flight risk and surveillance area priority are considered as cost determinants. In this respect, the proposed approach is different from many other approaches in UAV route planning.

The simulation results according to the embodiment of the present invention show that the route generated by the proposed route planner can preferentially visit important areas while obtaining high suitability in various real situations. Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram for explaining an operation scenario according to an embodiment of the present invention.

1 , MDr (Monitoring Drones) having similar specifications is used to monitor the entire operating area to detect the presence of IDr (Illegal Drones) in the 3D operating space. While monitoring, we establish that MDr cannot enter any area prohibited by regulation. To avoid being recognized and destroyed as a hostile drone, MDr is pre-recognized as MDr in Ground-based Detection System (GBDS) for detecting ground-based drones. Furthermore, it is established that the MDr communicate with each other in a UAV-to-ground link or ad hoc manner. Thus, MDr can share information with each other while performing flight tasks.

2 is a diagram illustrating an operating area and a path of an MDr according to an embodiment of the present invention.

In the implementation, the entire operating area is divided into several small unit areas called cells as shown in FIG. 2 , where the area marked in red represents the restricted area. It is assumed that the MDr can cover four cells at a specific location (i.e., waypoint) depending on the coverage gradient of the onboard camera imaging sensor (i.e., the area shown in blue).

3 is a diagram illustrating a representation of the terrain being monitored according to an embodiment of the present invention.

The landscape is represented by adopting a transformation of the Foxhole Shekel function to mimic the real terrain, and it is formulated as Equation (1).

(1)

(One)

과

는 지형 형태를 변화시키기 위해 활용한다. UAV에 대한 경로 계획 분야에서 널리 받아들여지는 벤치마크가 없기 때문에, 이러한 지형을 채택했다. 왜냐하면 이러한 지형은 지역의 경관에 고도가 높은 산, 빌딩과 같은 형태를 나타낼 수 있기 때문이다.Here, the parameter

and

is used to change the topography. Since there is no widely accepted benchmark in the field of route planning for UAVs, this terrain has been adopted. This is because such topography can represent high-altitude mountains and buildings in the local landscape.

In an implementation according to an embodiment of the present invention, the path generated by the optimization algorithm is a sequence of three-dimensional waypoints. Thus, feasible paths are encoded as vectors. Here, the element w _i =( x _i , y _i , z _i ) represents the i-th waypoint as in (2).

(2)

(2)

Here, N _w is the number of waypoints of the feasible route.

Next, we propose a multi-objective fitness function composed of 8 optimization indices to evaluate the path generated by the proposed multiple UAV path planning algorithm. To emphasize the importance of different optimization indices in the optimization process, we divide them into two groups and assign them different priorities: (I) constraints that the UAV must meet due to physical limitations (e.g., terrain, prohibited areas, angle of rotation, flight inclination, avoidance of multiple UAV collisions) and (II) optimization objects that must be maximized according to specific mission criteria (eg, energy consumption, flight hazard, and importance of surveillance area).

<Table 1>

Table 1 shows these classifications and their calculation formulas. For all possible UAV routes, a route with a high suitability value is always preferred. Therefore, we formulate the fitness function as

(3)

(3)

Here, the F _objective is an objective function whose value must be maximized to derive the optimal path. The remaining parts correspond to constraints that must be satisfied before planning a route. More details will be described below.

One of the optimization criteria is the objective function, which is used to improve the quality of the route planning. We define the objective function as the weighted component of energy consumption, flight risk, and surveillance area importance. Therefore, the objective function is formulated as indicated in (4).

(4)

(4)

Here, F _EC , F _FR and F _SAI are defined in the range [0,1], and w _i ( i = 1, 2, 3) reflects the significant difference while evaluating the candidate path as the weight of each factor. Intuitively, routes with lower energy and flight risk, but with higher surveillance area importance, are more desirable. Thus, while energy consumption and flight risk are defined as negative values, the importance of the monitoring area is positive.

Small drones are usually battery-powered, meaning they have to finish monitoring before they consume all of their energy. Therefore, feasible routes with low fuel consumption are always preferred. In the present invention, it is assumed that the UAV speed is constant during the operating time. EC can be formulated as

(5)

(5)

(6)

(6)

(7)

(7)

(8)

(8)

Here, EC _i is the fuel consumption from the i- th stop to the ( i +1) stop. P _u is the energy consumption at speed v in units of time. t _i,j+1 is the flight time from the i- th waypoint to the ( i+1) waypoint. d _i,i+1 is the 3D flight distance between the i-th waypoint and the ( i +1) waypoint, and max EC is the normalization constant value, which is formulated as follows.

(9)

(9)

Here, X , Y and Z represent the x, y, and z axes in the operating space, respectively.

MDr due to its physical properties (eg small and light) makes it vulnerable to weather conditions (eg rain, snow) during surveillance operations. In addition, flying altitude can be another great risk, as MDr can be accidentally destroyed by strong winds at high altitude. Based on the above scenario, we define the following two types of Flight Risk (FR).

은 그것들의 환경 가치의 합으로 정의된다.Because of the strong random nature of environmental risks, it is difficult to build accurate mathematical models. For simplicity, we generated random environmental risk values for each waypoint. and r, the environmental risk between the i- th stop and the ( i +1) stop

is defined as the sum of their environmental values.

(10)

(10)

는 상수 제어 파라미터이다.here,

is a constant control parameter.

Flight risk is a position-dependent parameter that only increases or decreases depending on weather conditions during flight and UAV flight altitude. The total flight risk can be calculated as in formulas (11) to (13).

(11)

(11)

(12)

(12)

Here, FR _i is the flight risk from the i- th waypoint to the ( i +1) waypoint, W _ER and W _AR are the weights of the environmental risk and the flight altitude risk, respectively, and max FR is expressed as follows.

(13)

(13)

는 최대 환경 위험을 나타낸다.here,

represents the maximum environmental risk.

When an MDr is designated to perform a specific security monitoring mission, it is first to monitor an important area, that is, an area with a higher Surveillance Area Importance (SAI) value. Therefore, SAI values are introduced to characterize different surveillance area priorities between cells. In an implementation according to an embodiment of the present invention, each cell is assigned a random SAI value, which is applied to the entire grid on the map. Therefore, the normalized SAI value of the feasible path can be calculated by formulas (14) to (16).

(14)

(14)

(15)

(15)

(16)

(16)

Here, SAI _i ( t ) and v _cellx ( t ) are the SAI values of the i- th waypoint and cell x for flight time t, respectively. N ( i ) is the set of cells that can be monitored at the i-th waypoint. N _n is the number of N ( _i ), and V _max is the maximum SAI value.

A constraint function is used to evaluate the validity of a path. If they are satisfied, then each constraint equals zero and if not, a negative penalty value Q is chosen. By choosing the Q value to be less than -1, we ensure that the fitness value of a feasible path is always greater than any non-feasible path. Therefore, if all constraint functions are satisfied during the optimization process, a feasible trajectory can always be obtained.

MDr cannot literally pass through terrain (eg collide with a mountain). Therefore, the flight altitude of the MDr must be higher than the altitude of the terrain. Use the described terrain function Altd ( x , y ) to determine the elevation of a location ( x , y ). Then, the terrain constraint can be described as follows.

(17)

(17)

In some specific areas (eg sensitive government areas), MDr cannot be entered by regulation. Legal pathways should be carefully designed to avoid such restricted areas. For simplicity, it is assumed that such forbidden areas are rectangular. The Forbidden Area Constraint (FAC) can be formulated as:

(18)

(18)

Here, l _x and l _y are the lower limits of the x and y coordinates of the j- th forbidden zone, respectively, and u _x and u _y are the upper limits of the x and y coordinates of the j- th forbidden zone, respectively.

4 is a diagram for explaining UAV mobility parameters according to an embodiment of the present invention.

As shown in Fig. 4(a), the rotation angle is defined as the horizontal angle between the previous direction and the current direction. Practical paths must be reasonably smooth to allow the UAV to maneuver easily. Therefore, the rotation angle of the UAV must be less than the maximum allowable rotation angle. This constraint can be formulated as

(19)

(19)

는 i번째 경유지(x _i , y _i , z _i )에서의 회전각이며,

는 최대 허용 회전각이다.

의 공식화는 다음과 같다.here,

is the rotation angle at the i- th waypoint ( x _i , y _i , z _{i ),}

is the maximum allowable rotation angle.

The formulation of is as follows.

(20)

(20)

는 표준 벡터이다.here,

is the standard vector.

As shown in Fig. 4(b), similar to the rotation angle in the horizontal direction, the flight inclination was introduced to show the UAV maneuverability in the vertical direction, that is, the glide and climb angles. Flight slope is defined as the slope between the horizontal direction between the current waypoint and the next waypoint. The flight slope must be within the range of maximum glide and climb angles. The Flying Slope Constraint (FSC) can be formulated as (21).

(21)

(21)

와

는 각각 허용 가능한 최대 활공각과 상승각이며, r _i 는 i번째 경유지의 비행경사이다. r _i 는 (22)와 같이 공식화할 수 있다.here,

Wow

are the maximum allowable glide and climb angles, respectively, and r _i is the flight slope of the i-th waypoint. r _i can be formulated as (22).

(22)

(22)

According to an embodiment of the present invention, the focus is on route planning for a plurality of UAVs. When multiple UAVs are used for complex surveillance missions, the path must be carefully designed to avoid collisions between the UAVs, which is essential for mission performance. For the two separate routes, the UAV must maintain a minimum safe distance to avoid collisions. This constraint can be described as follows.

(23)

(23)

는 q번째 UAV 경로의 i번째 경유지와 q번째 UAV 경로의 j번째 경유지 사이의 직교 거리이다.where d _min is the minimum safety distance to avoid collision,

is the orthogonal distance between the i- th stop of the q-th UAV path and the j-th stop of the q-th UAV path.

5 is a flowchart illustrating a method for optimizing a three-dimensional monitoring route setting of a plurality of unmanned aerial vehicles (UAV) using a small swarm optimization theory based on a monitoring area priority according to an embodiment of the present invention.

The proposed 3D monitoring route setting optimization method for multiple unmanned aerial vehicles (UAV) using the small swarm optimization theory based on the priority of the monitoring area uses the Particle Swarm Optimization (PSO) theory to monitor drones (Monitoring Drones). ; deriving an optimal route for MDr (510), obtaining the best flight stop sequence from all possible routes using a multi-objective fitness function (520), the existence of a collision using CAC (Coordinated Agents Coevolution); detecting (530), generating an optimal route for one flight time (540) if CAC is satisfied, and updating (550) the survival area importance (SAI) value before the next flight. .

In step 510, an optimal path for each monitoring drone is derived using the small swarm optimization theory. The velocity vector and position vector are generated for each monitoring drone, and the individual maximum value of the object and the maximum value of all objects are determined through PSO using the velocity vector and the position vector. Then, the optimal path is derived by improving the speed and position of each individual through the individual peak values of the individual and the highest values of all entities.

In step 520, the best flight waypoint sequence is obtained from all possible routes using the multi-objective fitness function.

In step 530, the presence of a collision is detected using the CAC, and if the CAC is satisfied, in step 540, an optimal route is generated for one flight time.

Then, in step 550, before the next flight, the SAI value is updated. Step 550 is a step of increasing the SAI value when the presence of illegal drones (IDr) is detected in the cell in which the MDr event is detected, the SAI value is rapidly changing compared to the past average SAI value increasing the SAI value for the area and decreasing the SAI value for the previously monitored area to avoid repeated monitoring for the same area.

When the presence of IDr is detected in the cell in which the MDr event is detected, the step of increasing the SAI value is when IDr is detected in a specific cell, monitoring neighboring cells to prevent IDr from entering the area, and Increases the SAI value of a cell and its neighboring cells for that cell. In addition, the step of reducing the SAI value for the previously monitored area to avoid repeated monitoring for the same area increases the SAI value for the cell when the cell is not monitored for more than a predetermined flight time, and monitoring Adjust the SAI value of the cell to reduce the SAI value for the cell that received . The 3D optimal security monitoring path planning method for multiple UAVs using the swarm optimization theory proposed below will be described in more detail.

According to an embodiment of the present invention, a monitoring area priority update mechanism based on event detection is proposed. When an MDr detects the presence of an IDr in a particular area (i.e., an event is detected in that area), it will either pay more attention to that area during the next flight or share information with other MDr, depending on intuition. MDr should also pay more attention to regions where SAI values are changing rapidly compared to historical average SAI values. Moreover, it is necessary to increase the SAI value so that the MDr can cover the area during the next flight for an area that is not monitored for more than a certain number of flights. In order to increase the monitoring efficiency, the SAI value for the previously monitored area is reduced to avoid repeated monitoring of the same area. Based on the above requirements, we define four cases for updating the SAI value.

1) When IDr is detected in a specific cell, the neighboring cell is monitored to prevent IDr from entering the area. Therefore, the SAI value of that cell and its adjacent cell is increased.

2) If the SAI value of a cell changes rapidly compared to the previous average SAI value, the SAI value of that cell is adjusted.

3) If the cell is not monitored for more than a certain flight time, the SAI value should be increased.

4) If the cell is monitored, the SAI value should be reduced.

Based on the explanation above, we formulate the cases as follows.

일 때, i번째 경유지에서 감시된 셀들에서 이벤트가 발생하였고, 그 셀들의 SAI 값은 (24)를 활용하여 업데이트할 수 있다.In the first case,

When , an event has occurred in the monitored cells at the i- th waypoint, and the SAI value of the cells may be updated by using (24).

(24)

(24)

Here, SAI _i ( t ) is the SAI value of the i- th waypoint during the t- th flight time , v _cellx ( t ) represents the SAI value of the cell x during the t-th flight time , th _event and v _max are the events, respectively. The detection threshold and the maximum SAI value.

In the second case, if the absolute value between the current SAI value of cell x and the previous average SAI value is greater than the predetermined update threshold th _update , the SAI value is returned as the maximum value[ v _max , f _cellx ( t )] Since it is updated by value, we formulate the second case as (25) and (26).

(25)

(25)

(26)

(26)

는 첫 번째부터 (t - 2)번째 비행 시간까지의 평균 SAI 값을 나타내며,

는 셀 x의 초기 SAI 값이며

는 상수 제어 매개변수이다. here,

represents the average SAI value from the first to the ( t - 2) time of flight,

is the initial SAI value of cell x

is a constant control parameter.

In the third case, if the cell is not monitored for more than a certain time-of-flight threshold (27), the SAI value is increased.

(27)

(27)

In the fourth case, when the cell is being monitored, (28) is calculated to decrease the SAI value.

(28)

(28)

The present invention proposes a route planner for a plurality of UAVs based on the standard PSO and the surveillance area-first update mechanism described above. The proposed planner first derives an optimal path for each MDr using PSO. Here, the proposed multi-objective fitness function is used to obtain the best flight waypoint sequence from all possible routes. The CAC is then used to detect the presence of a collision. If the CAC is satisfied, the planner generates an optimal route for one flight time. Finally, the SAI value will be updated before the next flight. More details are provided in the next subsection.

Particle Swarm Optimization (PSO) theory is a widely used evolutionary heuristic search algorithm to solve optimization problems. This theory was developed based on the behavior of hunting flocks of birds. In the process of finding food, each bird updates its speed (step size) and position based on information gathered from itself and from the flock. In PSO, each entity corresponds to a randomly initialized candidate solution. Then, at each iteration, the velocity and position of each entity is updated based on information about the previous velocity, the highest position occupied by the entity (individual influence), and the highest position occupied by any entity in the group (social group). The mathematical formula is:

Assume that the number of objects is P, the dimension of objects is D, and the number of repetitions is N. For the i-th entity, x _i = ( x _i1 , x _i2 , ..., x _iD ) and v _i = ( v _i1 , v _i2 , ..., v _iD ) represent the velocity and position vectors, respectively. In a standard PSO, there are two kinds of cost values. That is, there are P _i,best , which is the individual highest value described in (29), and S _{best, which} is the highest value of all individuals.

(29)

(29)

Once the two cost values are determined, the velocity and position of each entity in each dimension is improved using (28).

(30)

(30)

In (30), r ₁ and r ₂ are random values between 0 and 1. w is the inertia parameter, which reflects the effect of the previous iteration rate on the current iteration. c ₁ and c ₂ represent self-awareness and social awareness indicating the ability to inherit from the individual itself and from the whole group.

In the proposed planner, a feasible flight path consists of a stopover sequence and line segments. During implementation, an 8-way path is adopted. In the present invention, the entire operational area is divided into unit cell areas.

The proposed planner process is shown in Algorithm 1.

Initially, the estimated surveillance flight time is set to cover the entire operational area, and SAI values for all cells are initialized. Next, an optimal path for a plurality of MDr is derived using PSO, which corresponds to step 5 to step 33. During the optimization process, we first randomly generate the position and velocity vectors of N _par objects, and initialize P _t,Best and S _best to x _t and x _{Npar , respectively.} Next, the position and velocity vectors of each object are improved using the formula (27). Then, the newly updated entity is evaluated using the proposed multi-objective fitness function, which consists of several objective and constraint functions described above.

After that, we update P _t,Best and S _best based on the fitness values of all entities. Then, with the iteration number set equal to N _iter , the optimal path of the first MDr is stored. Finally, in order to prevent overlapping of waypoints between multiple MDrs, such as repeated monitoring of the same area, the SAI value of the monitored cell is reduced before starting the next route planning. Therefore, monitoring efficiency is improved. Since our study mainly focuses on collision-free path planning for multiple UAVs, we apply collision avoidance constraints to determine whether collisions can occur between MDr paths. If the CAC is not satisfactory, return to step 5. An optimal path for a plurality of MDr can be obtained only when CAC is satisfied. After that, the SAI value should be updated using formulas (22) to (25) before the next flight. Finally, when the flight time is equal to N _flight , all optimal planned routes are generated for a plurality of MDr covering the entire operating area except for the restricted area.

6 is a diagram illustrating the configuration of an apparatus for optimizing a three-dimensional monitoring route setting of a plurality of unmanned aerial vehicles (UAV) using a small swarm optimization theory based on a monitoring area priority according to an embodiment of the present invention.

The apparatus 600 for optimizing the setting of a three-dimensional monitoring route of a plurality of unmanned aerial vehicles (UAV) using a small number optimization theory based on the priority of the monitoring area includes a prediction unit 610 , a detection unit 620 , and an update unit 630 .

The predictor 610 , the detector 620 , and the updater 630 may be configured to perform steps 510 to 550 of FIG. 5 .

The prediction unit 610 derives an optimal path for each monitoring drone by using a small swarm optimization theory. The velocity vector and position vector are generated for each monitoring drone, and the individual maximum value of the object and the maximum value of all objects are determined through PSO using the velocity vector and the position vector. Then, the optimal path is derived by improving the speed and position of each individual through the individual peak values of the individual and the highest values of all entities.

The prediction unit 610 obtains the best flight waypoint sequence from all possible routes by using the multi-objective fitness function.

The detector 620 detects the presence of a collision by using the CAC, and generates an optimal route for one flight time when the CAC is satisfied.

The update unit 630 updates the SAI value before the next flight. The update unit 630 increases the SAI value when the presence of illegal drones (IDr) is detected in the cell in which the MDr event is detected. In addition, the SAI value is increased for an area where the SAI value is rapidly changing compared with the historical average SAI value, and the SAI value for the previously monitored area is decreased to avoid repeated monitoring for the same area.

When IDr is detected in a specific cell, the updater 630 monitors an adjacent cell in order to prevent IDr from entering the corresponding area, and increases the SAI value of the corresponding cell and the adjacent cell for the corresponding cell. In addition, the step of reducing the SAI value for the previously monitored area to avoid repeated monitoring for the same area increases the SAI value for the cell when the cell is not monitored for more than a predetermined flight time, and monitoring Adjust the SAI value of the cell to reduce the SAI value for the cell that received .

The device described above may be implemented as a hardware component, a software component, and/or a combination of the hardware component and the software component. For example, devices and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), It may be implemented using one or more general purpose or special purpose computers, such as a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that can include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

The software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or device, to be interpreted by or to provide instructions or data to the processing device. may be embodied in The software may be distributed over networked computer systems, and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and carry out program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

As described above, although the embodiments have been described with reference to the limited embodiments and drawings, various modifications and variations are possible from the above description by those skilled in the art. For example, the described techniques are performed in a different order than the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (8)

deriving an optimal path for each monitoring drone (MDr) using a particle swarm optimization (PSO) theory;
obtaining a best flight waypoint sequence from all possible routes using a multi-objective fitness function;
detecting the presence of a collision using Coordinated Agents Coevolution (CAC);
generating an optimal route for one flight time if CAC is satisfied; and
Steps to update SAI (surveillance area importance) values before next flight
A 3D optimal surveillance path planning method comprising a. According to claim 1,
The step of deriving the optimal path for each monitoring drone using the small swarm optimization theory is:
Generate velocity vector and position vector for each monitoring drone, determine individual maximum value of object and maximum value of all objects through PSO using velocity vector and position vector, individual maximum value of object and maximum value of all objects It derives the optimal path by improving the speed and position of each entity through the value, and derives the optimal path for collision prevention between monitoring drones to be applicable to autonomous flying devices.
How to plan a 3D optimal surveillance path. According to claim 1,
Before the next flight, the steps to update the SAI value are:
increasing the SAI value when the presence of illegal drones (IDr) is detected in the cell in which the MDr event is detected;
increasing the SAI value for regions where the SAI value is rapidly changing compared to the historical average SAI value; and
Decreasing the SAI value for the previously monitored area to avoid repeated monitoring for the same area.
A 3D optimal surveillance path planning method comprising a. 4. The method of claim 3,
When the presence of IDr is detected in the cell in which the MDr event is detected, increasing the SAI value comprises:
When IDr is detected in a specific cell, it monitors the neighboring cell to prevent IDr from entering the area, and increases the SAI value of that cell and the neighboring cell for that cell.
How to plan a 3D optimal surveillance path. 4. The method of claim 3,
The step of reducing the SAI value for a previously monitored area to avoid repeated monitoring for the same area is:
When a cell is not monitored for more than a predetermined flight time, the SAI value of the cell is adjusted to increase the SAI value for the cell and decrease the SAI value for the monitored cell.
How to plan a 3D optimal surveillance path. Using the Particle Swarm Optimization (PSO) theory to derive the optimal route for each Monitoring Drones (MDr), and to obtain the best flight waypoint sequence from all possible routes using the multi-purpose fitness function. predictor;
a detection unit that detects the presence of a collision using Coordinated Agents Coevolution (CAC) and generates an optimal route for one flight time when the CAC is satisfied; and
Updater to update SAI (surveillance area importance) values before the next flight
A 3D optimal surveillance path planning device comprising a. 7. The method of claim 6,
the prediction unit,
Generate velocity vector and position vector for each monitoring drone, determine individual maximum value of object and maximum value of all objects through PSO using velocity vector and position vector, individual maximum value of object and maximum value of all objects It derives the optimal path by improving the speed and position of each entity through the value, and derives the optimal path for collision prevention between monitoring drones so that it can be applied to autonomous flying devices.
3D Optimal Security Surveillance Path Planning Device. 7. The method of claim 6,
update department,
When the presence of illegal drones (IDr) is detected in the cell in which the MDr event is detected, the SAI value is increased;
increasing the SAI value for regions where the SAI value is rapidly changing compared to the historical average SAI value; and
Decrease the SAI value for previously monitored areas to avoid repeated monitoring for the same area.
3D Optimal Security Surveillance Path Planning Device.

Images