KR102207084B1 - Method for Cloud-Based Drone Navigation for Efficient Battery Charging in Drone Networks:CBDN - Google Patents

Abstract

In the present specification, in a drone network, a method performed in a traffic control device to allocate a flight path to a drone is at least one drone to which a flight path is assigned or at least one quick battery charging machine (QCM) Obtaining traffic information on the drone network from the network, the traffic information including at least one of first information of the at least one drone on the drone network or second information of the at least one QCM; Determining an optimal flight path using an algorithm for setting a flight path in consideration of the at least one drone based on the traffic information; And allocating the optimal flight path to the drone, wherein the optimal flight path is a path for minimizing the total flight time of the drones constituting the drone network.

Description

Cloud-Based Drone Navigation for Efficient Battery Charging in Drone Networks:CBDN}

The present invention relates to a drone navigation method in a drone network, and in more detail, by collecting traffic information of drones included in a drone network, and providing the drones with an optimal flight path considering battery charging, the overall drone It is about how to find global drone service routes that minimize the flight time of the people.

Recently, various services using drones have appeared. A kind of network consisting of drones and a drone management system is called a drone network. This drone network can include a large number of drones, and a system that manages a large number of drones included in the drone network is called a drone network management system. The drone network management system allocates flight routes to destinations to drones in the drone network. A large number of drones in service in this drone network have a limited battery capacity and perform short-range or long-distance flights using the flight path allocated from the drone network management system. At this time, especially when drones fly long distances, drones can reach the set destination by charging the battery in QCM via Quick Battery-Charging Machines (QCM) due to limited battery capacity. However, in the case where a drone must go through QCM to reach a set destination, the existing drone network management system according to the prior art allocates the flight path to the destination to the drones without considering whether other drones use QCM. I did it. In this way, the existing network system allocates flight paths via QCM for battery charging to drones without considering whether other drones use QCM, so that a large number of drones use a specific QCM for battery charging. In the specific QCM, a drone that arrives late according to the order of arrival of the drones causes congestion, such as a long waiting time for battery charging. Congestion in such a specific QCM increases the flight time of the overall drones in the drone service, making it difficult to efficiently operate the drone service.

The matters described in this background are prepared to enhance an understanding of the background of the invention, and may include matters not known in the prior art to those of ordinary skill in the field to which this technology belongs.

An object of the present specification is to provide a method of allocating an efficient drone service path considering battery charging to drones in flight included in a drone network.

In addition, an object of the present specification is to provide a method for finding global drone service routes that can shorten the overall flight time of drones by reducing congestion at the drone battery charging station to drones in flight included in the drone network.

The technical problems to be achieved in the present specification are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those of ordinary skill in the technical field to which the present invention belongs from the following description. I will be able to.

In the present specification, in a drone network, a method performed in a traffic control device to allocate a flight path to a drone is at least one drone to which a flight path is assigned or at least one quick battery charging machine (QCM) Obtaining traffic information on the drone network from the network, the traffic information including at least one of first information of the at least one drone on the drone network or second information of the at least one QCM; Determining an optimal flight path using an algorithm for setting a flight path in consideration of the at least one drone based on the traffic information; And allocating the optimal flight path to the drone, wherein the optimal flight path is a path for minimizing the total flight time of the drones constituting the drone network.

In addition, in the present specification, the first information includes at least one of average speed information, speed standard deviation information, departure location information, destination information, current location information, or moving direction information of the at least one drone, and the second information It may include at least one of current state of the at least one QCM, charging reservation information, and charging standby time information.

In addition, in the present specification, the determining of an optimal flight path includes: initializing a value stored in a memory of the traffic control device; Determining whether the drone is charged with a battery; Determining a QCM set, which is information on QCMs that can be reached with the remaining battery capacity of the drone; In each QCM, determining k first paths based on the individual waiting time and the QCM set indicating information on the total waiting time of the drone, wherein the first paths determine a path based on the individual waiting time, It is a route that took the least time; In each of the QCMs, the individual waiting time and the global waiting time, which is information representing the sum of the total waiting time of the at least one drone, is applied to the first routes to calculate the required time for each k routes. can do.

In addition, in the present specification, when it is determined that charging the battery is not necessary, determining an optimal flight path; It may further include the step of determining the shortest flight path to the destination not via the QCM.

In addition, in the present specification, the first flight paths may be determined using a flight time variable and a charging time variable.

In addition, in the present specification, there is provided a traffic control device for allocating a flight path to a drone, comprising: an RF unit for transmitting and receiving a radio signal; And a processor functionally connected to the RF unit, wherein the processor includes traffic information on the drone network from at least one drone or at least one quick battery charging machine (QCM) to which a flight path is assigned. Obtaining, the traffic information includes at least one of first information of the at least one drone on the drone network or second information of the at least one QCM; Determining an optimal flight path using an algorithm for setting a flight path in consideration of the at least one drone based on the traffic information; And allocating the optimal flight path to the drone, wherein the optimal flight path is a traffic control device that is a path that minimizes the total flight time of drones constituting the drone network.

The present invention has the effect of allocating an optimal service path considering battery charging to drones in flight on a drone network.

In addition, the present invention has the effect of minimizing the average flight time of the overall drones by allocating efficient service paths considering battery charging to drones in flight on a drone network.

The effects obtainable in the present specification are not limited to the above-mentioned effects, and other effects not mentioned can be clearly understood by those of ordinary skill in the art from the following description. will be.

1 is a diagram showing an example of a drone network architecture proposed in the present specification.
2 is a diagram showing an example of a local QCM optimization technique that only considers charging its own battery without considering other drones in service proposed in the present specification.
3 is a diagram showing an example of a global QCM selection optimization technique that also considers other drones included in the drone network proposed in the present specification.
4 is a diagram showing an example showing that the drone proposed in the present specification can obtain various routes for flying from a departure point to a destination.
5 is a diagram showing a graph showing a set of QCMs that a drone proposed in the present specification can reach as a remaining battery capacity at a starting point.
6 is a diagram showing an example for aiding understanding of a method of calculating a global waiting time proposed in the present specification.
7 is a diagram showing a flow chart of an optimal flight path allocation algorithm in consideration of other drones proposed in the present specification.
8 is a diagram showing the total flight time of a drone required when flying along a path allocated according to a flight path allocation algorithm proposed in the present specification.
9 is a diagram showing the total flight time in consideration of the additional waiting time of other drones on the drone network, which is required when flying along a path allocated according to the flight path allocation algorithm proposed in the present specification.
10 is a diagram showing a drone network simulation snapshot proposed in the present specification.
11 is a diagram showing a simulation result of QCM usage rate for each algorithm in the drone network proposed in the present specification.
12 is a diagram showing a simulation result on the effect of the performance of the present invention according to the number of drones proposed in the present specification.
13 is a diagram showing a simulation result for the effect of the performance of the present invention according to the speed of the drone proposed in the present specification.
14 is a diagram showing a simulation result for the effect of the performance of the present invention according to the standard deviation of the speed of the drone proposed in the present specification.
15 is a diagram showing a simulation result on the effect of the performance of the present invention according to the number of QCMs proposed in the present specification.
16 is a diagram showing a simulation result for the effect of the performance of the present invention according to the performance of the QCM proposed in the present specification
17 is a diagram showing a simulation result of the effect of the performance of the present invention according to the weight of a delivery article proposed in the present specification
18 is a diagram showing an example of an operation of a cloud-based navigation traffic control center (TTC) proposed in the present specification.
19 illustrates a block diagram of a traffic control apparatus to which the methods proposed in the present specification can be applied.

Since the present invention can apply various transformations and have various embodiments, specific embodiments are illustrated in the drawings and will be described in detail in the detailed description. However, this is not intended to be limited to a specific embodiment of the present invention, it is to be understood to include all conversions, equivalents, and substitutes included in the spirit and scope of the present invention. In describing the present invention, when it is determined that a detailed description of a related known technology may obscure the subject matter of the present invention, a detailed description thereof will be omitted.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. These terms are used only for the purpose of distinguishing one component from another component.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In the present application, terms such as "comprise" or "have" are intended to designate the presence of features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, but one or more other features. It is to be understood that the presence or addition of elements or numbers, steps, actions, components, parts, or combinations thereof, does not preclude in advance.

Drone network architecture

1 shows an example of a drone network architecture. Referring to FIG. 1, a Traffic Control Center (TTC, hereinafter referred to as TTC) 101, a Quick Battery Charging Machine (QCM, hereinafter referred to as QCM) 102, and a drone departing from the starting point. A drone network architecture composed of 103, a starting point 104, and a destination 105 are shown.

Specifically, the TTC 101, a core computing and storage node of the drone cloud, is a cloud-based management system that maintains the state of the drone 103 and the QCM 102 for location management. The TCC 101 provides the latest traffic statistics information of the drone 103 and QCM 102. The traffic statistics information of the drone includes information such as the starting point 104, the destination 105, the average speed of the drone 103 according to the current location and direction of movement, and the standard deviation of the speed of the drone 103, and the QCM 102 The traffic statistics information of the QCM 102 includes information such as current QCM 102 status, charging reservation information, and charging standby time.

The QCM 102 is a charging station that charges the battery of the drone 103. The QCM 102 periodically reports the current QCM 102 status, charge reservation information, charge standby time information, and the like to the TCC 101. The QCM 102 is randomly deployed throughout the drone 103 network.

The drone 103 is an unmanned aerial vehicle (UAV) that flies from a departure point to a destination. The drone 103 can know the delay of the waiting time for charging the battery in each QCM 102 through communication with the TCC 101. In order to reduce battery consumption due to frequent communication with the TCC 101, the drone 103 communicates with the TCC 101 only once at the starting point 104, except in an emergency.

As shown in FIG. 1, when the battery capacity is limited and the battery needs to be charged during service, the QCM 102 may charge the battery to continue providing the service. The drone service may refer to various services for delivering goods and information to the destination 105, such as a delivery service using the drone 103, an alarm service, or a pesticide spraying service. Hereinafter, it is used in the same sense.

The concept of cloud-based drone navigation

Hereinafter, the concept of Cloud-Based Drone Navigation (CBDN) will be described. Cloud-based drone navigation uses a charging reservation system in each QCM. Cloud-based drone navigation aims to reduce the total flight time of drones on the drone network. The total flight time is calculated as the sum of the waiting time in QCM, charging time in QCM and flight time from origin to destination.

Cloud-based drone navigation does not provide flight paths at the level of individual drones, but provides drones with optimal flight paths at the level of drone networks provided by TCC. In other words, it allocates flight paths taking into account other drones on the drone network. This is called a global QCM optimization technique.

FIG. 2 is an example showing a local QCM optimization technique, which is a technique of allocating flight paths by considering only their own battery charge without considering other drones in drone service. In FIG. 2, drones in service are indicated as d1 to d8, and QCMs available to drones in service are indicated as q1 to q7. For convenience of explanation, the present invention is limited to five drones d1 to d5. Referring to FIG. 2, when QCM is allocated to drones in service without considering the paths of other drones adjacent to it, congestion occurs due to a large number of drones in QCMs allocated to other adjacent drones.

At this time, the origin and destination of these drones (d1 to d5) may be different.

In order to reach the destination, each drone may start from the starting point sequentially or at different times, select one QCM from q1 to q5 on the determined route, select a battery, and charge the battery.

For convenience of explanation, it is limited to the case where the drones (d1 to d5) need to charge the battery on the route from the origin to the destination. Drones go through the QCM to charge the battery along the way to the destination. Drones (d1~d5) are assumed to have departed from the starting point in order of d1, d2, d3, d4, d5 for drone service.

As shown in FIG. 2, when the local QCM selection optimization technique does not take into account other drones, each of the drones (d1 to d5) charges the battery via the QCM indicated by q1, and the route to the destination is the fastest. May correspond to the path.

In this case, if the drones (d1~d5) arrive at QCM q1 in the order of d5, d4, d3, d2, d1 in the opposite order of departure, each drone at the time of departure (d1~d5) depends on the local QCM selection optimization technique. Although the route to the destination via QCM q1 was selected as the fastest flight route, the remaining d1-d4 drones except for the first drone d5, may not be the fastest flight route.

In other words, when drones departing late (e.g., d2~d5) select (and reserve) the QCM to charge the battery via the route to the destination by applying a local QCM selection optimization technique at the point of departure, they start first. This is because other drones that have chosen (and reserved) QCM are not considered, and have been assigned the optimal route by considering only their own battery charge. This optimal path calculation method causes congestion due to the waiting time for charging the battery in QCM q1.

The drone d5 that arrives first at QCM q1 does not have a waiting time for battery charging, so it can charge the battery as soon as it arrives at QCM q1, and when the battery charge is complete, the set optimal route to the destination (i.e., the least You can fly along the time required).

This path is actually the optimal path because there is no waiting time in QCM q1. However, the drones arriving late at q1 (d1~d4), when they arrive at q1, first arrive at QCM q1 to charge the battery, but if there is a drone that has not yet finished charging the battery, a waiting time for battery charging occurs.

Therefore, at the time of departure, the route from q1 to charging the battery and flying to the destination was calculated as the optimal route, but in reality, the calculated optimal route may not be the best route due to the occurrence of waiting time for battery charging. have. In addition, the additional waiting time for the drones (d1 to d4) arriving late at QCM here increases the overall flight time of the drones included in the drone network.

In order to solve this problem, the present invention proposes a cloud-based drone navigation system that allocates flight paths through a global QCM selection optimization technique that also considers other drones in service.

3 shows an example of a global QCM selection optimization technique that also considers other drones included in the drone network. The global QCM selection optimization technique is a technique that collects traffic information of drones and QCMs using a cloud-based drone network management system, and allocates flight paths considering efficient battery charging to drones based on the collected traffic information. The above information collection and flight path assignment are performed in the TCC of the cloud-based drone network system.

The traffic information of the drones includes information such as drone speed information, speed standard deviation information, departure location information, destination information, current location information, and movement direction information.

The traffic information of the QCM may include current QCM status, charging reservation information in QCM, charging waiting time information in QCM, and the like.

In more detail, the global QCM selection optimization technique is a technique that uses a faster path compared to the additional waiting time of other drones that may occur due to battery charging of a specific drone and the bypass path using other QCMs. In FIG. 3, drones in service are indicated by d1 to d8, and QCM is indicated by q1 to q7.

For convenience of explanation, the description is limited to drones of d1 to d5. Each of the drones (d1~d5) starts from the starting point in the order d1, d2, d3, d4, d5. Unlike the case where the local QCM selection optimization technique of FIG. 2 is applied, the drones of FIG. 3 (d1 to d5) receive a flight path considering other drones at the start point from the TCC of the cloud-based drone network management system.

As a result of receiving the flight path through this method, the QCM that the drones (d1 to d5) pass through on the flight path to the destination may be set differently for each drone. For example, as shown in FIG. 3, drones d1 and d5 are assigned q1, drone d3 is assigned q2, and drones d2 and d4 are assigned q3.

As each drone is assigned a different QCM, the time required for waiting until the battery charge of the drones that arrive late at a specific QCM to complete charging of the batteries can be greatly reduced or may not occur at all.

As described above, applying the global QCM selection optimization technique does not cause congestion due to waiting time in a specific QCM, and has the effect of minimizing the overall flight time of drones included in the drone network.

That is, through the global QCM selection optimization technique, an excessive number of drones can be avoided from passing through a specific QCM, and congestion due to the waiting time for battery charging in the corresponding QCM is not caused. This leads to the effect of minimizing the flight time of all drones included in the drone network.

However, from the individual point of view of each drone, if the global QCM selection optimization technique is followed, there may be a problem of giving up the optimal route that can be used by each drone to avoid congestion in a specific QCM.

For example, suppose there is an optimal route (Route 1) that takes the least amount of time to reach the destination among various routes to reach the destination of a drone (Drone 1). In this route 1, there is a QCM (QCM1) that drone 1 must pass through to charge the battery, but many other drones may also choose this QCM1. According to the global QCM selection optimization technique, congestion with other drones is expected in QCM1 to which drone 1 is assigned, and to avoid such congestion, a route (path 2) that bypasses to another QCM (QCM2) will be assigned to drone 1. will be. Drone 1, which has been assigned the new route 2, cannot use route 1, which was the optimal route to the destination, and chooses the newly allocated route 2, which requires a longer travel time.

In other words, from the standpoint of each individual drone, for the purpose of reducing the overall flight time of the drones included in the network, there may be a situation in which they must give up the route that was best for them. However, in terms of the overall drone network, individual drones may have their own. By abandoning the optimal route of the drones, the overall flight time of drones using the drone network can be reduced, which enables efficient operation of the entire drone network.

As shown above, compared to the local QCM selection optimization technique that can bring about a large increase in the overall flight time of drones, the drone global QCM selection optimization technique using a cloud-based management system can be seen to have a clear advantage.

In the following, an algorithm applied to a global QCM selection optimization technique using a cloud-based drone network management system, that is, an algorithm for cloud-based drone navigation, will be described in more detail.

Specific design of cloud-based drone navigation

4 is a diagram showing an example showing that a drone can obtain various routes for flying from a departure point to a destination. 4, there are 4 routes that a drone can obtain. As shown in Fig. 4, the drone can reach its destination through various routes.

Therefore, using a cloud-based navigation system to allocate the most suitable route among these various routes to a drone is of importance in terms of efficient drone network operation and battery saving of individual drones in the drone network.

In the following, before looking at an algorithm for a specific design of a cloud-based drone navigation system, a method of calculating a global waiting time will be first described. The global standby time takes into account not only the waiting time for charging the battery of the drone that is currently assigned a flight path from the cloud-based management system, but also the waiting time for charging the battery of other drones caused by the drone that is currently assigned the flight path. It's a time value.

Specifically, suppose that the current drone 1 is assigned a flight path via QCM 1 from cloud-based drone navigation. QCM 1 is reserved for battery charging for other drones (e.g., Drone 2 and Drone 3). Drones arrive at QCM 1 in the order of drone 2, drone 1, and drone 3. Drone 2 is currently charging the battery in QCM 1. Before drone 1 was assigned a flight route via QCM 1, when drone 2 finished charging the battery, drone 3 tried to charge the battery. However, since Drone 1 arrives before Drone 3, Drone 3 has to wait until the battery of Drone 1 is fully charged, even though the battery of Drone 2 is completely charged. In this way, the waiting time for drone 1's battery charging (i.e., the waiting time for drone 2 to finish charging) and the additional waiting time for drone 3 (i.e., drone 1 arriving before drone 3, drone 3 The time plus the time to wait for additional charging) corresponds to the global standby time.

The formula for calculating the global waiting time is given below. 5 is a diagram illustrating an example for helping understanding a method of calculating a global waiting time. In FIG. 5, the unit of time follows the unit of minute (min).

Look at the meaning of the above formula. Wg represents the global waiting time you want to find. Wi represents the individual waiting time in QCM of a drone that is currently assigned a flight path. Kerr represents the number of drones in the j-th zone of FIG. 5. Wo represents the overlapped time between the charging time of the drone currently reserved in QCM and the charging time of a late arrival drone. Gi represents the index of the section where there is no drone charging in QCM. A method of calculating the global waiting time through the example of Equation 1 and FIG. 5 may be described as follows.

Looking at the waiting time occurring in the first area (area 1) 541 of the QCM, charging of the four drones 511, 512, 513, and 514 is reserved from 5 minutes to 230 minutes. A new drone arrives at QCM at 15 minutes (501), and the drone that arrives at 15 minutes (501) can charge the battery from 65 minutes, when the charging of the drone (511) that was being charged first ends at 65 minutes. Here, an individual waiting time of 50 minutes occurs for the drone arriving at 15 minutes (501).

When the charging of the drone 511 that arrived first ends in 65 minutes, the drone that arrived at 15 minutes 501 starts charging the battery for 75 minutes from 65 minutes. The drone (501), which arrived at 15 minutes, starts charging, and the drone (512), which is scheduled to charge the battery from 65 minutes, the drone (513), which is scheduled to charge the battery from 110 minutes, and the battery is charged from 230 minutes. The charging time of the drones 514 reserved for the following is delayed by 75 minutes each. Due to the 501 charging of the drone that arrived in 15 minutes, the total flight time of the drone network was increased by 225 (3*75) minutes.

The waiting time occurring in the second area (area 2) 542 is examined. The time to finish charging of all drones in the first area 541 is 305 (230+75) minutes. The expected charging start time of the second area 542 is 260 minutes, but the actual charging start time is 305 minutes after 45 minutes. Accordingly, the charging of the drone 521 scheduled to charge the battery from 260 minutes and the drone 522 scheduled to charge the battery from 305 minutes are delayed by 45 minutes, respectively. Due to the delayed time in the first area, it further increases the total flight time of the drone network by 90 (45 * 2) minutes.

The waiting time occurring in the third area (area 3) 543 is examined. The time to finish charging of all drones in the second area 542 is 425 (380+45) minutes. Accordingly, the expected charging start time of the third area 543 was 420 minutes, but the actual charging start time is 425 minutes, which is delayed by 5 minutes. Accordingly, charging of the drone 531 scheduled to charge the battery from 420 minutes and the drone 532 reserved to charge the battery from 495 minutes are delayed by 5 minutes, respectively. Due to the delay in the second zone 542, the total flight time of the drone network is further increased by 10 (5 * 2).

As above, the global waiting time, which is the sum of all additional waiting times occurring in the first to third areas, is 375 minutes. Equation 1 expresses that the global waiting time is calculated in this way.

Hereinafter, a detailed algorithm of a cloud-based drone network navigation operation will be described. The table below is the algorithm of a cloud-based drone network.

Hereinafter, the meaning of the algorithm variables shown in Table 1 and the meaning of each line will be described.

Dcur: It represents a drone departing from the starting point.

Cbattery: Shows the remaining battery capacity of the drone departing from the starting point.

Bw: Indicates the weight of the drone's package departing from the starting point.

Qall: Represents all QCM sets on the drone network.

Gu,v: This is a graph showing the QCM that can be reached with the current battery on the path from the origin u to the destination v.

u: indicates the location of the source.

v: indicates the location of the destination.

Pu,v: represents the route from the origin u to the destination v.

Pi: Of the k paths from the origin to the destination derived through the k-shortest path function, the i-th shortest path is the sum of the flight distance from the origin to the destination and the distance to the QCM in the graph of Gu and v Represents. For example, out of seven routes, the route of the third shortest distance is denoted by P3.

Represents the estimated total flight time of the Ti:Pi route (flight time, charge time, individual charge wait time and global charge wait time sum).

Tu,v: represents the total flight time from the origin u to the destination v.

Fi: Shows the expected flight time from the origin to the destination.

Ci: represents the sum of the expected charging times in QCMs included in the Pi path.

Wi: This represents the sum of the expected global waiting times in QCMs included in the Pi path.

Hereinafter, the meaning of each line will be described.

1: The traffic control center (TCC) of the cloud-based drone navigation system receives information from drones for starting the current drone service and performs battery charging scheduling.

2.: Initialize variables. At this time, the variables described above are initialized. 3: The step of determining whether the distance to the destination where the drone should fly is the distance that requires battery charging. At this time, the flight time of the drone depends on the weight of the package that the drone has to carry to its destination.

4: If it is determined that the distance from line3 to the destination the drone should fly is a route that can be taken without charging the battery, enter the Pdirect value representing the route from the departure point to the destination without charging the battery in the value of the Pu,v variable. Where u is the starting point and v is the destination. In addition, Pu,v is a variable representing the path from the origin u to the destination v.

5: Returns the Pu,v value of which the Pdirect value is input.

6-8: If the distance that the drone must fly to its destination in line 3 is the distance that requires battery charging, the set of QCMs that can be reached with the battery is obtained for efficient battery charging allocation.

9: Create a graph using the set obtained in lines 6-8. For example, this graph may be a graph showing the set of QCMs that a specific drone with a starting point u and a destination v can reach with the current battery.

6 is a diagram illustrating a graph drawn using the derived QCM set.

10: Use the k-Shortest-path algorithm on the graph created in line 9 to obtain k paths to destinations in the order of the most efficient paths. The K-Shortest-Path algorithm can use two variables. For example, flight time (time1), which is the sum of the time it takes the drone to travel its route from its origin to QCM (time1) and the time it takes the drone to travel its route from QCM to its destination (time2). You can use the +time2) variable and the charging time variable, which is the charging time in QCM. The sum of these two possible variables derives k paths in small order. Each of the derived k paths may be given an index from 1 to k.

11-12: k paths can be represented as variables such as {P1, P2, P3??Pk}. Compute_Travel_Time_of_Path(P_i) of Line 12 represents a function that calculates the time (Ti) required for each path from P1 to Pk. The variable Ti representing the time spent on the i-th path Pi can be expressed as the sum of three variables. The three variables are flight time (time1), which is the sum of the time it takes the drone to travel the route from the origin to the QCM (time1) and the time it takes the drone to travel the route from the QCM to the destination (time2). The +time2) variable and the charging time variable, which is the charging time in QCM, can be used, and it can also be a global waiting time variable, which is the waiting time in QCM considering other drones included in the network. And by summing these three variables flight time, charging time, and global waiting time, the time Ti- spent on the i-th path Pi can be obtained.

13: Compare the variable Tu,v representing the time taken from the origin u to the destination v, and the time Ti taken in the route Pi derived in line 12. At this time, in line 2 above, all variables are initialized, and infinite values are input to all variables as initial values.In this case, the initial comparison step, that is, the T1 value and the initial time taken considering the global waiting time in the path P1 In the step of comparing Tu,v where the value is set, the T1 value will be input to the memory as the Tu,v value in line 14 below.

14: When T1 is smaller than Tu,v, the Ti value is input to Tu,v. For example, if the required time T2 in the route P2 is stored in Tu,v and the required time T3 in the route P3 is compared with Tu,v, if the value of T3 is less than Tu,v, Tu The value of T3 will be entered in the ,v value. Conversely, if the value of T3 is greater than Tu,v, the value stored in the memory in Tu,v will not be changed.

15: Compared with the values of Tu,v, the Pi path, which is the path of Ti, which takes less time, is modified to the path from the origin u to the destination v

16-19: Returns the last selected path, Pu,v. The finally set Pu,v corresponds to the optimal route considering the flight time of the drone, charging time in QCM, and global waiting time.

20: End the algorithm.

7 is an example of a flowchart showing the above-described algorithm. To apply the global QCM selection optimization technique, all variables are initialized (S7010).

At this time, all variables include the variables mentioned in step 1 of FIG.

In the next step, it is determined whether or not the drone battery needs to be charged to reach the destination u (S7020). If battery charging is not required, the Pdirect value is input into the route Pu,v from the source u to the destination v (S7032), and the procedure ends (S7091).

When battery charging is required, in the next step, a Qset, a set of QCMs that can be reached with the remaining battery capacity of the drone, is derived, and a graph is created using the derived Qset (S7031).

Next, k paths from the departure point to the destination via QCM are derived (S7040).

In consideration of the global waiting time, the time Ti required when following the i-th path Pi is calculated (S70570).

When following the path Pi, a procedure for comparing whether the time required Ti is less than the currently set Tu,v is performed (S7060).

If Ti is less than Tu,v, a Pi value is input to Pu,v (S7071).

When Ti is greater than Tu,v, the value of Pu,v is not changed (S7072).

In the next step, it is determined whether the value of i is equal to k (S7080).

If the k value is equal to the i value, the procedure ends (S7091).

If the value of k is greater than the value of i, 1 is added to the current value of i (S7092), and the process returns to the S7050 procedure, and the subsequent procedures from S7050 are repeated as described above.

Cloud-based drone navigation collects traffic information of QCM with other drones in flight by TCC, and based on the collected information, the drone network's drone network's drone network is based on the previously salpin cloud-based drone navigation algorithm (i.e., the global QCM selection optimization technique applied algorithm). It is a system that implements a method of allocating routes that can minimize the overall flight time.

Comparison with other algorithms for QCM selection and flight path assignment

The four paths 401 to 404 shown in FIG. 4 are selected through different algorithms. The algorithm used to allocate the four paths of FIG. 4 is as follows.

(1)Shortest-Flight-Wait-Time(SFWT)(Route 1), (2)Shortest-Flight-Time(SFT)(Route 2), (3) Individual Reservation Navigation(IRN)(Route 3), (4) ) Cloud-Based Drone Navigation (CBDN) (Route 4) was used.

In more detail, the (1)Shortest-Flight-Wait-Time (SFWT) algorithm used to allocate the route 1 401 of FIG. 4 is used in QCM to utilize the statistical waiting time for charging the drone's battery. Algorithm.

The (2) Shortest-Flight-Time (SFT) algorithm used to allocate the route 2 402 of FIG. 4 is an algorithm considering only the flight time of drones.

The route 3 (403) (3) Individual Reservation Navigation (IRN) algorithm of FIG. 4 is a QCM reservation system algorithm that considers only the waiting time for charging its own (ie, a drone currently assigned a flight route) battery in QCM.

The (4) Cloud-Based Drone Navigation (CBDN) algorithm used to allocate route 4 (404) in FIG. It is a system algorithm.

8 is a diagram showing the total flight time (8010 to 8040) of the drone required when flying along a path allocated according to the previously described algorithm.

In more detail, Figure 8 does not take into account the additional waiting time of other drones (i.e., does not take into account the global waiting time), and only the waiting time for charging the battery of itself (i.e., a drone assigned a route). It is a diagram showing the considered flight time.

Specifically, looking at the overall result of FIG. 8, it can be seen that when only the own waiting time is considered, when setting the flight path by applying the IRN system algorithm, the optimum flight path can be allocated from the standpoint of an individual drone.

This result is a result of not considering that the waiting time in QCM of other drones may increase due to individual drones. Therefore, from the perspective of the entire drone network, the drones currently assigned the flight paths are assigned the flight paths through the IRN algorithm, resulting in an increase in the overall flight time of all drones in the network. This is a bad result in terms of efficient operation of the entire drone network.

Therefore, for efficient operation of the drone network, a method of reducing the total flight time of the entire drone network is suitable. Hereinafter, the result of applying the global waiting time considering not only the waiting time of the drone itself, but also the battery charging waiting time of other drones will be described.

9 is a diagram showing the total flight time (9010 to 9040) of the drone required when flying along a path allocated according to the above-described algorithm. In more detail, FIG. 9 is a diagram showing the total flight time in consideration of the additional waiting time of other drones (ie, taking into account the global waiting time).

Specifically, looking at the overall result of FIG. 9, when the flight path is set by applying the CBDN system algorithm taking into account the waiting times of other drones, the most efficient network operation is possible from the perspective of the entire drone network.

For example, assume that there are currently 3 drones (Drones A to C) flying in the drone network, and there are 2 QCMs available for drones (QCM A and QCM B). Although the destination of each drone is different, it is assumed that all three drones must recharge the battery once via QCM A to reach the destination.

Currently, if the flight times of drones A to C are respectively Ta, Tb, and Tc, the total flight time of all drones on the current network (i.e., net flight time from departure to destination, charging standby time, and charging time) The total flight time of the network is T=Ta+Tb+Tc.

Currently, drone D heading to its destination is assigned a flight path at the starting point. Excluding the waiting time for charging in the case where drone D passes through QCM A and when passing through QCM B, it is said that the time required from the departure point to the destination is the same as 60 minutes in both cases.

That is, the sum of the time required to move from the origin to the QCM (20 minutes), the charging time in the QCM (20 minutes), and the time required to move from the QCM to the destination (20) is the same.

It is assumed that the waiting time in QCM A of drone D is 20 minutes, and the waiting time in QCM B is 30 minutes. The flight time of drone D is called Td.

Below, we look at the overall flight time of the drone network. From the perspective of drone D, if you consider only its own charging wait time, the route to the destination via QCM A, which has less waiting time, will be the most suitable route. At this time, the flight time Td of drone D is 80 (20+20+20+20) minutes.

However, suppose that if drone D uses QCM A, the waiting time of drones A to C increases by 20 minutes. This will increase the global waiting time by 60 minutes. In this case, the total flight time of the network T(T1)=Ta+Tb+Tc+80(Td)+60 (global waiting time).

Conversely, if drone D is assigned a route that considers the waiting time of other drones and uses QCM B, the flight time Td of drone D at this time is 90 minutes (20+20+20+30). The flight time of drone D has been increased, but the flight time of drones A to C does not increase because other drones do not use QCM B and only use QCM A. In other words, the global waiting time is zero. In this case, the total flight time of the network T(T2)=Ta+Tb+Tc+90(Td)+0 (global waiting time).

As a result of applying the CBDN system algorithm, the total flight time of the network is reduced by 50 minutes. Therefore, allocation of flight paths using the CBDN algorithm system may assign longer flight paths to individual drones, but it can be seen that the network has the effect of minimizing the overall flight time.

As described above, cloud-based drone navigation (CBDN) is a technique that finds the optimal path that can minimize the average flight time of the overall drones, rather than finding its own optimal path.

Experimental examples according to application of various variables and other algorithms

Hereinafter, when various environments and various variables are applied to the method of the present invention, examples of experiments in which performance evaluation is performed on how the performance of the method of the present invention are affected will be described.

In order to clearly reveal the advantageous effects of the present invention over other algorithms examined above, a comparison with other algorithm systems will be described.

In order to evaluate the performance of the present invention, an experiment was conducted using OMNeT++ simulation, and the drone specification of the Matrice 200 Series developed by DJI was used for drone simulation. The simulation setting values used are shown in Table 2 below.

The performance evaluation experiment of this method was based on the simulation settings in Table 2. 10 is a diagram showing a drone network simulation snapshot. 11 shows the QCM usage rate for each applied algorithm in this experiment. Figure 10 (d) shows the result of the CBDN system algorithm of the present invention. The CBDN system algorithm allocates QCM to drones in consideration of efficient battery charging to avoid congestion in QCM. Therefore, it can be seen that the usage rate of each QCM is evenly distributed, and the advantageous effect of using the CBDN system algorithm and the QCM on the drone network in a balanced and efficient manner is well revealed.

Hereinafter, experimental examples according to various variables in various environments will be described. In the graphs of Figs. 12 to 17 below, the y-axis of the graph (a) of Figs. 12 to 17 represents the average flight time of the drones included in the drone network, and the y-axis of the graph (b) of Figs. 12 to 17 is an average Shows the QCM usage rate. Hereinafter, for convenience of explanation, descriptions of the y-axis of the graphs (a) and (b) of FIGS. 12 to 17 are omitted.

(1) Effect of the number of drones

12 is a diagram showing the effect of the performance of the present invention according to the number of drones. The x-axis of the graphs (a) and (b) of FIG. 12 represents the number of drones included in the drone network. Analyzing the results of the graph (a) of FIG. 12, it can be seen that in the case of the CBDN system algorithm, the average flight time of drones is the shortest when compared with other algorithms. In addition, unlike other systems, despite the increase in the number of drones, it can be seen that the flight time of drones on the drone network does not change significantly accordingly. In addition, when analyzing the results of the graph of FIG. 12 (b), CBDN has the highest QCM average usage rate when compared with other algorithms. In addition, it can be seen that as the number of drones increases, the average QCM usage rate increases, and the increase is the largest compared to other algorithms.

(2) Influence of drone speed

13 is a diagram showing the effect of the performance of the present invention according to the speed of the drone. The x-axis of the graphs (a) and (b) of FIG. 13 represents the speed of the drone. Analyzing the results of the graph (a) of FIG. 13, it can be seen that in the case of the CBDN system algorithm, the average flight time of drones is the shortest when compared with other algorithms. In addition, analyzing the results of the graph of FIG. 13 (b), CBDN has the highest QCM average usage rate when compared with other algorithms. In addition, it can be seen that as the speed of the drone increases, the average QCM usage rate increases, and the increase is the largest compared to other algorithms.

(3) The effect of the standard deviation of the speed of the drone

14 is a diagram showing the effect of the performance of the present invention according to the size of the standard deviation of the speed of the drone. The x-axis in the graphs (a) and (b) of FIG. 14 represents the magnitude of the standard deviation of the speed of the drone. When analyzing the results of the graph of FIG. 14 (a), it can be seen that in the case of the CBDN system algorithm, the average flight time of drones is the shortest when compared with other algorithms. In addition, when analyzing the result of the graph of FIG. 14 (b), CBDN has the highest QCM average usage rate when compared with other algorithms.

(4) Impact of the number of QCMs

15 is a diagram showing the effect of the performance of the present invention according to the number of QCMs. The x-axis of the graphs (a) and (b) of FIG. 15 represents the number of QCMs included in the drone network. Analyzing the results of the graph (a) of FIG. 15, it can be seen that in the case of the CBDN system algorithm, the average flight time of drones is the shortest when compared with other algorithms. In addition, analyzing the results of the graph of FIG. 15B, CBDN has the highest QCM average usage rate when compared with other algorithms. In addition, it can be seen that as the number of drones increases, the average QCM usage rate decreases, and the reduction is the largest compared to other algorithms. The fact that the decrease in the QCM average usage rate according to the increase in the number of QCMs is large means that the drones are using a large number of QCMs evenly, so this result is an experimental result that shows the advantageous effects of the CBDN system algorithm.

(5) Impact of QCM performance

16 is a diagram showing the effect of the performance of the present invention according to the performance of the QCM. The x-axis of the graphs (a) and (b) of FIG. 16 represents the performance of QCM (ie, the time taken to charge the drone's battery from a discharged state to a fully charged state). Analyzing the results of the graph (a) of FIG. 16, it can be seen that in the case of the CBDN system algorithm, the average flight time of drones is the shortest compared to other algorithms. In addition, when analyzing the results of the graph of FIG. 16 (b), CBDN has the highest QCM average usage rate when compared with other algorithms.

(6) Effect of the weight of the delivered item

17 is a diagram showing the effect of the performance of the present invention according to the weight of the delivered article. The x-axis of the graphs (a) and (b) of FIG. 17 represents the weight of the shipment. Analyzing the results of the graph of FIG. 17 (a), it can be seen that in the case of the CBDN system algorithm, the average flight time of drones is the shortest when compared with other algorithms. In addition, when analyzing the result of the graph of FIG. 17 (b), CBDN has the highest QCM average usage rate when compared with other algorithms.

18 is an example showing an operation of a traffic control center (TCC) in cloud-based navigation. That is, in the drone network, the method performed by the traffic control device to allocate flight paths to the drones,

A step of acquiring traffic information on the drone network from at least one drone or at least one quick battery charging machine (QCM) to which a flight path is assigned is performed (S1810).

The traffic information may include at least one of first information of the at least one drone on the drone network or second information of the at least one QCM.

Next, based on the traffic information, a step of determining an optimal flight path is performed using an algorithm for setting a flight path in consideration of the at least one drone (S1820).

Next, the step of allocating the optimal flight path to the drone is performed (S1830).

The optimal flight path is a path that minimizes the total flight time of the drones constituting the drone network.

19 illustrates a block diagram of a traffic control apparatus to which the methods proposed in the present specification can be applied.

Referring to FIG. 19, the traffic control apparatus includes a processor 1910, a memory 1920, an RF unit 1930, and a control unit 1940.

The processor 1910 implements the functions, processes and/or methods proposed in FIGS. 1 to 18 above. The memory 1920 is connected to the processor and stores various information for driving the processor. The RF unit 1930 is connected to the processor and transmits/receives information from drones and QCMs. The control unit 1940 controls the operation of the traffic control device.

Furthermore, although each drawing has been described separately for convenience of description, it is also possible to design a new embodiment by merging the embodiments described in each drawing. In addition, designing a computer-readable recording medium in which a program for executing the previously described embodiments is recorded is within the scope of the present invention according to the needs of those skilled in the art.

The direction-based device search method according to the present specification is not limitedly applicable to the configuration and method of the embodiments described as described above, but the embodiments are all or part of each embodiment so that various modifications can be made. It may be configured in combination.

Meanwhile, the direction-based device search method of the present specification may be implemented as a code that can be read by a processor in a recording medium that can be read by a processor provided in a network device. The processor-readable recording medium includes all types of recording devices that store data that can be read by the processor. Examples of recording media that can be read by the processor include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, etc., and also include those implemented in the form of carrier waves such as transmission through the Internet. . Further, the processor-readable recording medium is distributed over a computer system connected through a network, so that the processor-readable code can be stored and executed in a distributed manner.

In addition, although the preferred embodiments of the present specification have been illustrated and described above, the present specification is not limited to the specific embodiments described above, and the technical field to which the present invention belongs without departing from the gist of the present invention claimed in the claims. In addition, various modifications can be implemented by those of ordinary skill in the art, and these modifications should not be understood individually from the technical spirit or prospect of the present invention.

In addition, in this specification, both the product invention and the method invention are described, and the description of both inventions may be supplementally applied as necessary.

Claims (10)

In a drone network, the method performed by a traffic control device to assign a flight path to a drone is:
Obtaining traffic information on the drone network from at least one drone or at least one quick battery charging machine (QCM) to which a flight path is assigned,
The traffic information includes at least one of first information of the at least one drone on the drone network or second information of the at least one QCM;
Determining an optimal flight path using an algorithm for setting a flight path in consideration of the at least one drone based on the traffic information; And
Allocating the optimal flight path to the drone; Including
The optimal flight path is a path that minimizes the total flight time of the drones constituting the drone network,
The steps to determine the optimal flight path are:
Initializing a value stored in a memory of the traffic control device;
Determining whether the drone is charged with a battery;
Determining a QCM set, which is information on QCMs that can be reached with the remaining battery capacity of the drone;
Determining k first paths based on the individual waiting time and the QCM set indicating information on the total waiting time of the drone in each QCM,
The first paths are paths that take a minimum amount of time when a path is determined based on the individual waiting time; And
Calculating a required time for each k paths by applying a global waiting time to the first paths;
Including,
The global waiting time is based on 1) the individual waiting time in each of the QCMs and 2) the charging time of the drones reserved for each QCM and the charging time of the drones arriving late at each of the QCMs. Indicating the waiting time of the entire at least one drone, method.
The method of claim 1, wherein the first information includes at least one of average speed information, speed standard deviation information, departure location information, destination information, current location information, or movement direction information of the at least one drone,
The second information includes at least one of a current state of the at least one QCM, charging reservation information, and charging standby time information.
delete The method of claim 1,
If it is determined that charging the battery is not necessary, the step of determining the optimal flight path is:
The method further comprising the step of determining the shortest flight path to the destination not via the QCM.
The method of claim 1,
The first paths are determined using a flight time variable and a charging time variable.
In a drone network, in a traffic control device to allocate a flight path to a drone,
An RF unit for transmitting and receiving wireless signals; And
Including a processor functionally connected to the RF unit,
The processor obtains traffic information on the drone network from at least one drone or at least one quick battery charging machine (QCM) to which a flight path is allocated,
The traffic information includes at least one of first information of the at least one drone on the drone network or second information of the at least one QCM,
Based on the traffic information, an optimal flight path is determined using an algorithm for setting a flight path in consideration of the at least one drone,
Allocating the optimal flight path to the drone,
The optimal flight path is a path that minimizes the total flight time of the drones constituting the drone network,
To determine the optimal flight path,
Initialize the value stored in the memory of the traffic control device,
Determine whether the drone is charged with the battery,
Determine a QCM set, which is information of QCMs that can be reached with the remaining battery capacity of the drone,
In each QCM, k first paths are determined based on the individual waiting time and the QCM set indicating information on the total waiting time of the drone,
The first routes are routes that take a minimum amount of time when a route is determined based on the individual waiting time,
Applying the global waiting time to the first routes to calculate the required time for each k routes,
The global waiting time is based on 1) the individual waiting time in each of the QCMs and 2) the charging time of the drones reserved for each QCM and the charging time of the drones arriving late at each of the QCMs. Traffic control device indicating the waiting time of the entire at least one drone.
The method of claim 6,
The first information includes at least one of average speed information, speed standard deviation information, departure location information, destination information, current location information, or movement direction information of the at least one drone,
The second information includes at least one of a current state of the at least one QCM, charging reservation information, and charging waiting time information.
delete The method of claim 6,
The processor, in order to determine an optimal flight path when it is determined that battery charging is not required,
A traffic control device that determines the shortest flight path to a destination not via QCM.
The method of claim 6,
The first routes are determined by using a flight time variable and a charging time variable.

Images