A method for monitoring electronic fence of unmanned aerial vehicle

Abstract

The invention relates to the technical field of route trajectory allocation, and discloses an electronic fence monitoring method for unmanned aerial vehicles. The method first divides a monitoring area, allocates a unmanned aerial vehicle to each sub-area to perform a patrol mission, and then uses an optimization algorithm to generate multiple candidate patrol path plans according to factors such as unmanned aerial vehicle performance constraints and mission requirements. Each plan consists of multiple paths, and coordinate sequence coding is used to represent the flight trajectory of each unmanned aerial vehicle in the sub-area. Finally, each plan is evaluated according to indicators such as coverage rate and path length, and the optimal plan is selected to implement monitoring. The invention can efficiently solve the optimal patrol path of unmanned aerial vehicles in the monitoring area through measures such as area division, diversified path generation, coding optimization and iterative update, thereby improving the monitoring coverage rate and the patrol path length of the unmanned aerial vehicle formation, and has positive application value.

Description

Technical Field

[0001] The present invention relates to the field of route track allocation, and more specifically, to an electronic fence monitoring method for a drone. Background Art

[0002] Electronic fence is the most advanced perimeter anti-theft alarm system. With its unique performance, electronic fence is widely used in substations, power plants, water plants, factories, industrial centers, industrial and mining enterprises, material warehouses, residential areas, villa areas, schools, airports, aquaculture and livestock farms, government agencies, key cultural relics sites, military facilities, prisons, detention centers and other places with walls or that need walls; for large areas, drones are currently used in conjunction with electronic fences to conduct all-round monitoring of the area, achieving long-term, continuous and efficient coverage monitoring of the monitored area.

[0003] The existing monitoring method is to have drones patrol and monitor evenly along the prescribed navigation route. The coverage of this method will change with the movement of the drone, and thus it is impossible to ensure that the coverage of the drone monitoring is maximized. In addition, when electronic fence monitoring is applied to actual scenarios, it is also necessary to consider the impact of the drone energy consumption factor on the drone's intelligent collaborative formation control scheme, which in turn makes the existing drone's intelligent collaborative formation control scheme unreasonable and the drone's patrol path scheme is not the most optimal. Summary of the invention

[0004] The present invention provides an electronic fence monitoring method for an unmanned aerial vehicle, which solves the technical problems raised in the above-mentioned background technology.

[0005] The present invention provides an electronic fence monitoring method for a drone, comprising the following steps:

[0006] Step 101, the entire monitoring area is evenly divided into multiple sub-areas, and each sub-area is monitored by a drone;

[0007] Step 102: randomly generate multiple drone patrol path plans, where one drone patrol path plan includes a trajectory of each drone flying in the sub-area it is responsible for;

[0008] Step 103, obtaining a patrol path feature vector by encoding a plurality of randomly generated UAV patrol path schemes;

[0009] Step 104, calculating the fitness value of the patrol path feature vector of the UAV patrol path plan;

[0010] Step 105: Select an update method to update the patrol path feature vector of the UAV patrol path plan;

[0011] Step 106: Select the patrol path feature vector with the largest fitness value obtained after the update as the benchmark feature vector, and then select the current UAV patrol path plan generated when the benchmark feature vector was last updated;

[0012] Step 107: Generate patrol paths for drones in all sub-areas according to the arrangement order of the waypoint coordinates in each sub-area in the current drone patrol path plan.

[0013] In a preferred embodiment, in step 101, the method of evenly dividing the entire monitoring area into N sub-areas includes the following steps:

[0014] Step 1: Collect a top view of the monitoring area, establish a coordinate system with the lower left corner of the top view as the origin, and obtain all vertices of the boundary of the monitoring area;

[0015] Step 2: Connect all vertices with a closed convex polygon and obtain the x value of all vertices in the convex polygon. max and x min and max and min ;

[0016] Among them, x max Indicates the maximum horizontal coordinate of the vertex in the monitoring area, x min Indicates the minimum horizontal coordinate of the vertex in the monitoring area, y max Indicates the maximum vertical coordinate of the vertex in the monitoring area, y min Indicates the minimum ordinate among the vertices in the monitoring area;

[0017] Step 3: Construct the minimum bounding rectangle of the monitoring area, with (x min ,y min ) as the coordinate of the lower left corner of the minimum enclosing rectangle, (x max ,y max ) as the upper right corner coordinate of the minimum enclosing rectangle, and construct the minimum enclosing rectangle of the monitoring area;

[0018] Step 4: Take the minimum bounding rectangle of the monitoring area as the initial node area, and initialize the number of divisions t to 0;

[0019] Step 5: Divide the node area into N sub-areas (N is a positive integer greater than 3). If S 子 ≥S 阈 , S 子 represents the projected area of ​​the sub-region, S 阈 If the projection area threshold of the set area is reached, the sub-area is used as the node area for the next round of division, and the number of iterations is accumulated by one. Otherwise, the t-th sub-area is output as the sub-area for the final monitoring area division.

[0020] In a preferred embodiment, in step 101, the area calculation formula of the sub-region is:

[0021]

[0022] in, represents the projection area of ​​the t-th sub-region, Represents the coordinates of the i-th vertex of the t-th divided sub-region, Represents the i+1th vertex coordinate of the sub-area divided for the tth time. For the last vertex, the coordinates of subsequent vertices return to the first vertex.

[0023] In a preferred embodiment, in step 102, the constraints of the generated drone patrol path plan are:

[0024] The distance between adjacent waypoints does not exceed the maximum flight distance limit of the drone;

[0025] The distance between two drones does not exceed the monitoring diameter of the drones;

[0026] The distance between two drones is not less than the disturbance range diameter of the drones.

[0027] In a preferred embodiment, in step 104, the calculation formula of the fitness value is as follows:

[0028]

[0029] in represents the fitness value of the patrol path feature vector of the kth UAV patrol path plan, ω1 and ω2 represent the weight coefficient of the UAV endurance time and the weight coefficient of the UAV's monitoring coverage of the monitoring area, respectively. ω1 and ω2 are adjusted according to the system's requirements for the UAV's endurance time and the UAV's monitoring coverage of the monitoring area, ω1∈[0,1], ω2∈[0,1], T k represents the total duration of the UAV patrol path in the kth UAV patrol path plan, C k It represents the monitoring coverage of the monitoring area in the kth UAV patrol path plan, and σ represents the Sigmoid function, which is an S-type function.

[0030] In a preferred embodiment, in step 104, the calculation formula for the total duration of the drone patrol path in the drone patrol path plan is:

[0031] T k =L k / v

[0032] Among them, L krepresents the distance of all UAV patrol paths in the kth UAV patrol path plan, and v represents the rated average flight speed of the UAV;

[0033] The calculation formula for the distance of all drone patrol paths is:

[0034]

[0035] in, represents the distance of the UAV patrol path in the jth sub-area in the kth UAV patrol path plan, N represents the number of sub-areas, M represents the number of waypoints in the jth sub-area, represents the coordinates of the ith waypoint in the jth sub-area of ​​the kth UAV patrol path plan, represents the coordinates of the i+1th waypoint in the jth sub-area of ​​the kth UAV patrol path plan;

[0036] The calculation formula for the monitoring coverage of the monitoring area in the drone patrol path plan is:

[0037] C k =G k / Z

[0038] Among them, G k It represents the number of grids covered by all UAV patrol paths, and Z represents the total number of grids in the monitoring area.

[0039] In a preferred embodiment, in step 104, the method for calculating the monitoring coverage of the monitoring area in the drone patrol path plan includes: step 1, first discretizing the monitoring area into Z small grids, the range of the monitoring area is (x min ,y min ) to (x max ,y max );

[0040] The formula for calculating the number of grid rows is:

[0041] W h =ceil((y max -y min ) / d)

[0042] Among them, W h Indicates the number of rows of the network after the monitoring area is discretized, y max Indicates the maximum vertical coordinate of the vertex in the monitoring area, y min represents the minimum ordinate of the vertex in the monitoring area, and d represents the side length of the grid;

[0043] The formula for calculating the number of grid columns is:

[0044] W l=ceil((x max -x min ) / d)

[0045] Among them, W l represents the number of columns of the network after the monitoring area is discretized, x max Indicates the maximum horizontal coordinate of the vertex in the monitoring area, x min represents the minimum horizontal coordinate of the vertex in the monitoring area, and d represents the side length of the grid;

[0046] Step 2: traverse all patrol paths and determine whether the grid where each waypoint it passes is covered;

[0047] For each drone patrol path:

[0048] Extract all waypoint coordinates in the path;

[0049] For each grid cell, check whether the rectangular area of ​​the grid cell intersects with any segment of the path. If so, mark the grid cell as "covered";

[0050] Step 3: Count the number of all grid cells marked as "covered" G;

[0051] Step 4: Calculate the monitoring coverage of the monitoring area in the drone patrol path plan.

[0052] In a preferred embodiment, in step 106, the first update condition is:

[0053] R sk <0.5, P<1;

[0054] The second update condition is:

[0055] R sk <0.5, P ≥ 1;

[0056] The third update condition is:

[0057] R sk ≥0.5;

[0058] in, H F represents the set of patrol path feature vectors in all schemes before updating, H k represents the patrol path feature vector in the kth solution, H pk represents the mean vector of the patrol path feature vector in the k-th solution before updating, T represents vector transpose, and sigmoid represents the S-type function;

[0059] P=2a·r1-a

[0060] C=2r2

[0061] In the formula, the range of r1 and r2 is [0, 1], r1 and r2 are randomly generated, and a is the convergence factor;

[0062] a=2-2t / T max

[0063] Where, T max is the maximum number of iterations;

[0064] The update method that meets the first update condition is as follows:

[0065] x i (t+1)=x * (t)-P·(C·x * (t)-x i (t))

[0066] x * (t) represents the historical feature vector with the maximum fitness value after completing t updates, x i (t) represents the updated historical feature vector;

[0067] The update method that meets the second update condition is as follows:

[0068] x i (t+1)=x suij (t)-P·(C·x suij (t)-x i (t))

[0069] x i (t) represents the updated historical feature vector, x suij (t) represents a random historical feature vector that has completed t updates;

[0070] The update method that meets the third update condition is as follows:

[0071] x i (t+1)=[C·x * (t)-x i (t)]·e L cos(2πL)x * (t)

[0072] In the formula, the value range of L is [-1, 1], L is a random number, and e is a natural constant.

[0073] An electronic fence monitoring system for drones, including the following modules:

[0074] The area division module is used to divide the entire monitoring area into multiple sub-areas evenly;

[0075] The path plan generation module is used to randomly generate multiple UAV patrol path plans;

[0076] A scheme coding generation module is used to generate patrol path feature vectors by coding a plurality of randomly generated UAV patrol path schemes;

[0077] A fitness value calculation module is used to calculate the fitness value of the patrol path feature vector of the UAV patrol path plan;

[0078] An updating module, used to update the patrol path feature vector of the UAV patrol path plan;

[0079] The decision output module generates patrol path plans for drones in all sub-areas based on the patrol path feature vector of the updated drone patrol path plan.

[0080] A computer storage medium is used to store computer-readable instructions, which can execute the above-mentioned electronic fence monitoring method of a drone when executed by a computer system.

[0081] The beneficial effects of the present invention are that the present invention can efficiently solve the optimal patrol path of drones in the monitoring area through measures such as area division, diversified path generation, coding optimization and iterative updating, thereby improving the monitoring coverage and the patrol path length of the drone formation, and has positive application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 The present invention is a flowchart of the electronic fence monitoring method of the unmanned aerial vehicle.

[0083] Figure 2 It is a system block diagram of the electronic fence monitoring system of the drone of the present invention.

[0084] In the figure: 100, area division module; 200, path plan generation module; 300, plan code generation module; 400, fitness value calculation module; 500, update module; 600, decision output module. DETAILED DESCRIPTION

[0085] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the contents of this specification. Each example may omit, replace or add various processes or components as needed. In addition, the features described relative to some examples may also be combined in other examples.

[0086] like Figure 1As shown, a method for monitoring an electronic fence of a drone includes the following steps:

[0087] Step 101, the entire monitoring area is evenly divided into N sub-areas, and each sub-area is monitored by a drone;

[0088] Collect historical flight trajectory data of the UAV in the monitoring area and the data of the monitoring area. The historical flight trajectory data of the UAV in the monitoring area includes the rated flight speed information of the UAV, the monitoring altitude information of the UAV, the telemetry data of the UAV, the environmental data of the UAV, etc. The data of the monitoring area includes the geographic information of the monitoring area, vector data such as roads, pipelines, and regional boundaries in the monitoring area, and waypoint distribution data of the monitoring area.

[0089] In one embodiment of the present invention, the method of evenly dividing the entire monitoring area into N sub-areas (all drones fly at the same altitude) includes the following steps:

[0090] Step 1: Collect a top view of the monitoring area, establish a coordinate system with the lower left corner of the top view as the origin, and obtain all vertices of the boundary of the monitoring area;

[0091] Step 2: Connect all vertices with a closed convex polygon and obtain the x value of all vertices in the convex polygon. max and x min and max and min ;

[0092] Among them, x max Indicates the maximum horizontal coordinate of the vertex in the monitoring area, x min Indicates the minimum horizontal coordinate of the vertex in the monitoring area, y max Indicates the maximum vertical coordinate of the vertex in the monitoring area, y min Indicates the minimum ordinate among the vertices in the monitoring area;

[0093] Step 3: Construct the minimum bounding rectangle of the monitoring area, with (x min ,y min ) as the coordinate of the lower left corner of the minimum enclosing rectangle, (x max ,y max ) as the upper right corner coordinate of the minimum enclosing rectangle, and construct the minimum enclosing rectangle of the monitoring area;

[0094] Step 4: Take the minimum bounding rectangle of the monitoring area as the initial node area, and initialize the number of divisions t to 0;

[0095] Step 5: Divide the node area into N sub-areas (N is a positive integer greater than 3). If S 子 ≥S 阈 , S 子represents the projected area of ​​the sub-region, S 阈 If the projection area threshold of the set area is reached, the sub-area is used as the node area for the next round of division, and the number of iterations is accumulated by one. Otherwise, the t-th sub-area is output as the sub-area for the final monitoring area division.

[0096] The area of ​​the sub-region is calculated as:

[0097]

[0098] in, represents the projection area of ​​the t-th sub-region, Represents the coordinates of the i-th vertex of the t-th divided sub-region, Indicates the coordinates of the i+1th vertex of the sub-area divided for the tth time. For the last vertex, the coordinates of subsequent vertices return to the first vertex. S 阈 The size of is determined by the monitoring range and endurance of the monitoring drone. At least the endurance of the drone must be able to meet the requirements of S 阈 Complete three patrols in a sub-area of ​​size.

[0099] Step 102: randomly generate multiple drone patrol path plans, where one drone patrol path plan includes a trajectory of each drone flying in the sub-area it is responsible for;

[0100] A coordinate system is established with the lower left side of the top view of the entire monitoring area as the origin. The coordinate system is the same as the coordinate system established by dividing the sub-areas in step 101. A series of waypoints are arranged in each sub-area. Each waypoint represents a relatively important monitoring area in the sub-area. Each waypoint is represented by its two-dimensional coordinates. The trajectory of each drone flying in the sub-area it is responsible for is composed of a series of waypoint coordinates arranged in the order in which the drone arrives at the waypoints.

[0101] The constraints of the generated UAV patrol path plan are:

[0102] 1. The distance between adjacent waypoints does not exceed the maximum flight distance limit of the drone;

[0103] 2. The distance between two drones does not exceed the monitoring diameter of the drones;

[0104] 3. The distance between two drones shall not be less than the diameter of the drone’s disturbance range.

[0105] Step 103, obtaining a patrol path feature vector by encoding a plurality of randomly generated UAV patrol path schemes;

[0106] The patrol path feature vector of the kth UAV patrol path plan is expressed as They represent the UAV patrol path feature vectors in the 1st, 2nd, jth, and Nth sub-areas in the kth UAV patrol path plan respectively;

[0107] The feature vector of the UAV patrol path in the jth sub-area is expressed as It is represented as the coordinates of the waypoint that the j-th UAV passes through for the i-th time. It is represented as the coordinates of the waypoint that the j-th UAV passes through for the Mth time;

[0108] It should be noted that the “()” and “,” in the above patrol path feature vector are only separators for convenience of explanation and are not included in the calculation process of the following steps, where M represents the number of waypoints in each sub-area;

[0109] For example, the monitoring area is divided into three sub-areas, and the number of waypoints in the three sub-areas is 10, 18, and 16, respectively. Then the patrol path feature vector of the UAV patrol path plan contains the coordinates of 10, 18, and 16 waypoints, respectively.

[0110] Step 104: Calculate the fitness value of the patrol path feature vector of the UAV patrol path plan. The calculation formula of the fitness value is as follows:

[0111]

[0112] in represents the fitness value of the patrol path feature vector of the kth UAV patrol path plan, ω1 and ω2 represent the weight coefficient of the UAV endurance time and the weight coefficient of the UAV's monitoring coverage of the monitoring area, respectively. ω1 and ω2 are adjusted according to the system's requirements for the UAV's endurance time and the UAV's monitoring coverage of the monitoring area, ω1∈[0,1], ω2∈[0,1], T k represents the total duration of the UAV patrol path in the kth UAV patrol path plan, C k represents the monitoring coverage of the monitoring area in the k-th UAV patrol path plan, σ represents the Sigmoid function, which is an S-type function;

[0113] The calculation formula for the total duration of the drone patrol path in the drone patrol path plan is:

[0114] T k =L k / v

[0115] Among them, L k represents the distance of all UAV patrol paths in the kth UAV patrol path plan, and v represents the rated average flight speed of the UAV;

[0116] The calculation formula for the distance of all drone patrol paths is:

[0117]

[0118] in, represents the distance of the UAV patrol path in the jth sub-area in the kth UAV patrol path plan, N represents the number of sub-areas, M represents the number of waypoints in the jth sub-area, represents the coordinates of the ith waypoint in the jth sub-area of ​​the kth UAV patrol path plan, represents the coordinates of the i+1th waypoint in the jth sub-area of ​​the kth UAV patrol path plan;

[0119] The calculation formula for the monitoring coverage of the monitoring area in the drone patrol path plan is:

[0120] C k =G k / Z

[0121] Among them, G k It represents the number of grids covered by all UAV patrol paths, and Z represents the total number of grids in the monitoring area.

[0122] In one embodiment of the present invention, a method for obtaining the total number of grids in a monitoring area and the number of grids covered by all drone patrol paths is provided:

[0123] Step 1: Discretize the monitoring area into Z small grids. The range of the monitoring area is x min ,y min ) to (x max ,y max );

[0124] The formula for calculating the number of grid rows is:

[0125] W h =seil((y max -x min ) / d)

[0126] Among them, W h Indicates the number of rows of the network after the monitoring area is discretized, y max Indicates the maximum vertical coordinate of the vertex in the monitoring area, y min represents the minimum ordinate of the vertex in the monitoring area, and d represents the side length of the grid;

[0127] The formula for calculating the number of grid columns is:

[0128] W l =ceil((x max -x min ) / d)

[0129] Among them, W l represents the number of columns of the network after the monitoring area is discretized, x max Indicates the maximum horizontal coordinate of the vertex in the monitoring area, x min represents the minimum horizontal coordinate of the vertex in the monitoring area, and d represents the side length of the grid;

[0130] Step 2: traverse all patrol paths and determine whether the grid where each waypoint it passes is covered;

[0131] For each drone patrol path:

[0132] (1) Extract the coordinates of all waypoints in the path;

[0133] (2) For each grid cell, check whether the rectangular area of ​​the grid cell intersects with any segment of the path. If so, mark the grid cell as “covered”.

[0134] Step 3: Count the number of all grid cells marked as "covered" G;

[0135] Step 4: Calculate the monitoring coverage of the monitoring area in the drone patrol path plan.

[0136] Step 105: Select an update method to update the patrol path feature vector of the UAV patrol path plan;

[0137] The first update condition is:

[0138] R sk <0.5, P<1;

[0139] The second update condition is:

[0140] R sk <0.5, P ≥ 1;

[0141] The third update condition is:

[0142] R sk ≥0.5;

[0143] in, H F represents the set of patrol path feature vectors in all schemes before updating, H k represents the patrol path feature vector in the kth solution, H pk represents the mean vector of the patrol path feature vector in the k-th solution before updating, T represents vector transpose, and sigmoid represents the S-type function;

[0144] P=2a·r1-a

[0145] C=2r2

[0146] In the formula, the range of r1 and r2 is [0, 1], r1 and r2 are randomly generated, and a is the convergence factor;

[0147] a=2-2t / T max

[0148] Where, T max is the maximum number of iterations;

[0149] The update method that meets the first update condition is as follows:

[0150] x i (t+1)=x * (t)-P·(C·x * (t)-x i (t))

[0151] x * (t) represents the historical feature vector with the maximum fitness value after completing t updates, x i (t) represents the updated historical feature vector;

[0152] The update method that meets the second update condition is as follows:

[0153] x i (t+1)=x suij (t)-P·(C·x suij (t)-x i (t))

[0154] x i (t) represents the updated historical feature vector, x suij (t) represents a random historical feature vector that has completed t updates;

[0155] The update method that meets the third update condition is as follows:

[0156] x i (t+1)=[C·x * (t)-x i (t)]·e L cos(2πL)x * (t)

[0157] In the formula, the value range of L is [-1, 1], L is a random number, and e is a natural constant;

[0158] Step 106: Select the patrol path feature vector with the largest fitness value obtained after the update as the benchmark feature vector, and then select the current UAV patrol path plan generated when the benchmark feature vector was last updated;

[0159] The selected updated historical feature vector is the updated T max The historical feature vector of T max The value is 100;

[0160] Step 107: Generate patrol paths for drones in all sub-areas according to the arrangement order of the waypoint coordinates in each sub-area in the current drone patrol path plan.

[0161] The monitoring area is divided into four sub-areas, each of which is patrolled by a drone; the baseline feature vector contains the codes of the four drone patrol path plans in a coding format, for example:

[0162] {(10,20)(30,40)(50,10)(20,60);(60,15)(40,25)(20,35)(80,45);

[0163] (30,70)(50,50)(70,30)(10,10);(15,5)(35,25)(55,45)(25,65)}.

[0164] Follow these steps to decode the drone patrol path in each sub-area from the code:

[0165] 1. Separate the encoded string by “;” to get 4 segments, corresponding to 4 drone paths respectively;

[0166] 2. For each encoded string:

[0167] A. Press "()" to separate and get the coordinates of 4 waypoints;

[0168] B. Separate each waypoint coordinate by "," to get the x and y coordinate values;

[0169] C. Connect all waypoints in sequence to form the patrol path of the drone;

[0170] 3. Corresponding to each sub-area, we finally get:

[0171] UAV patrol path in sub-area 1: (10,20)-(30,40)-(50,10)-20,60);

[0172] UAV patrol path in sub-area 2: (60,15)-(40,25)-(20,35)-80,45);

[0173] The drone patrol path in sub-area 3: (30,70)-(50,50)-(70,30)-10,10);

[0174] UAV patrol path in sub-area 4: (15,5)-(35,25)-(55,45)-25,65);

[0175] The above path is composed of a series of two-dimensional coordinate points. The drone will patrol in the sub-area by passing through each waypoint in turn according to this path.

[0176] like Figure 2 As shown, an electronic fence monitoring system for a drone includes the following modules:

[0177] The area division module 100 is used to divide the entire monitoring area into multiple sub-areas on average;

[0178] A path plan generation module 200 is used to randomly generate multiple drone patrol path plans;

[0179] A scheme coding generation module 300 is used to perform scheme coding on a plurality of randomly generated UAV patrol path schemes to generate a patrol path feature vector;

[0180] A fitness value calculation module 400 is used to calculate the fitness value of the patrol path feature vector of the UAV patrol path plan;

[0181] An updating module 500, used to update a patrol path feature vector of a patrol path plan of a drone;

[0182] The decision output module 600 generates patrol path plans for drones in all sub-areas based on the patrol path feature vector of the updated drone patrol path plan.

[0183] The present invention provides a computer storage medium for storing computer-readable instructions, which can execute the aforementioned electronic fence monitoring method for a drone when executed by a computer system.

[0184] The above describes an embodiment of the present embodiment, but the present embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present embodiment, ordinary technicians in this field can also make many forms, all of which are within the protection of the present embodiment.

Claims

1. A method for monitoring an electronic fence of a drone, characterized in that: The following steps are involved: Step 101, the entire monitoring area is evenly divided into multiple sub-areas, and each sub-area is monitored by a drone; Step 102: randomly generate multiple drone patrol path plans, where one drone patrol path plan includes a trajectory of each drone flying in the sub-area it is responsible for; Step 103, obtaining a patrol path feature vector by encoding a plurality of randomly generated UAV patrol path schemes; Step 104, calculating the fitness value of the patrol path feature vector of the UAV patrol path plan; Step 105: Select an update method to update the patrol path feature vector of the UAV patrol path plan; Step 106: Select the patrol path feature vector with the largest fitness value obtained after the update as the benchmark feature vector, and then select the current UAV patrol path plan generated when the benchmark feature vector was last updated; Step 107: Generate patrol paths for drones in all sub-areas according to the arrangement order of the waypoint coordinates in each sub-area in the current drone patrol path plan; In step 101, the method of evenly dividing the entire monitoring area into N sub-areas includes the following steps: Step 1: Collect a top view of the monitoring area, establish a coordinate system with the lower left corner of the top view as the origin, and obtain all vertices of the boundary of the monitoring area; Step 2: Connect all vertices with a closed convex polygon and obtain the x value of all vertices in the convex polygon. max and x min and max and min ; Among them, x max Indicates the maximum horizontal coordinate of the vertex in the monitoring area, x min Indicates the minimum horizontal coordinate of the vertex in the monitoring area, y max Indicates the maximum ordinate of the vertex in the monitoring area, y min Indicates the minimum ordinate among the vertices in the monitoring area; Step 3: Construct the minimum bounding rectangle of the monitoring area, with x min ,y min As the coordinate of the lower left corner of the minimum enclosing rectangle, x max ,y max As the upper right corner coordinate of the minimum enclosing rectangle, the minimum enclosing rectangle of the monitoring area is constructed; Step 4: Take the minimum bounding rectangle of the monitoring area as the initial node area, and initialize the number of divisions t to 0; Step 5: Divide the node area into N sub-areas. If S 子 ≥S 阈 , S 子 represents the projected area of ​​the sub-region, S 阈 If the projection area threshold of the set area is reached, the sub-area is used as the node area for the next round of division, and the number of iterations is accumulated by one. Otherwise, the t-th sub-area is output as the sub-area for the final monitoring area division; The area of ​​the sub-region is calculated as: in, represents the projection area of ​​the t-th sub-region, Represents the coordinates of the i-th vertex of the t-th divided sub-region, Represents the coordinates of the i+1th vertex of the sub-region divided for the tth time; In step 102, the constraints of the generated UAV patrol path plan are: The distance between adjacent waypoints does not exceed the maximum flight distance limit of the drone; The distance between two drones does not exceed the monitoring diameter of the drones; The distance between two drones is not less than the diameter of the drone’s disturbance range; In step 104, the calculation formula of the fitness value is as follows: in represents the fitness value of the patrol path feature vector of the kth UAV patrol path plan, ω1 and ω2 represent the weight coefficient of the UAV endurance time and the weight coefficient of the UAV's monitoring coverage of the monitoring area, respectively. ω1 and ω2 are adjusted according to the system's requirements for the UAV's endurance time and the UAV's monitoring coverage of the monitoring area, ω1∈[0,1], ω2∈[0,1], T k represents the total duration of the UAV patrol path in the kth UAV patrol path plan, C k represents the monitoring coverage of the monitoring area in the k-th UAV patrol path plan, σ represents the Sigmoid function, which is an S-type function; In step 104, the calculation formula for the total duration of the drone patrol path in the drone patrol path plan is: T k =L k / v Among them, L k represents the distance of all UAV patrol paths in the kth UAV patrol path plan, and v represents the rated average flight speed of the UAV; The calculation formula for the distance of all drone patrol paths is: in, represents the distance of the UAV patrol path in the jth sub-area in the kth UAV patrol path plan, N represents the number of sub-areas, M represents the number of waypoints in the jth sub-area, represents the coordinates of the ith waypoint in the jth sub-area of ​​the kth UAV patrol path plan, represents the coordinates of the i+1th waypoint in the jth sub-area of ​​the kth UAV patrol path plan; The calculation formula for the monitoring coverage of the monitoring area in the drone patrol path plan is: C k =G k / Z Among them, G k represents the number of grids covered by all drone patrol paths, and Z represents the total number of grids in the monitoring area; In step 104, the method for calculating the monitoring coverage of the monitoring area in the drone patrol path plan includes: step 1, first discretizing the monitoring area into Z small grids, the range of the monitoring area is (x min ,y min ) to (x max ,y max ); The formula for calculating the number of grid rows is: W h =ceil((y max -y min ) / d) Among them, W h It represents the number of rows of the network after the monitoring area is discretized, and d represents the side length of the grid; The formula for calculating the number of grid columns is: W l =ceil((x max -x min ) / d) Among them, W l It represents the number of columns of the network after the monitoring area is discretized, and d represents the side length of the grid; Step 2: traverse all patrol paths and determine whether the grid where each waypoint it passes is covered; For each drone patrol path: Extract all waypoint coordinates in the path; For each grid cell, check whether the rectangular area of ​​the grid cell intersects with any segment of the path. If so, mark the grid cell as "covered"; Step 3: Count the number of all grid cells marked as "covered" G; Step 4: Calculate the monitoring coverage of the monitoring area in the UAV patrol path plan; In step 106, the first update condition is: R sk <0.5,P<1; The second update condition is: R sk <0.5,P≥1; The third update condition is: R sk ≥0.5； in, H F represents the set of patrol path feature vectors in all schemes before updating, H k represents the patrol path feature vector in the kth solution, H pk represents the mean vector of the patrol path feature vector in the k-th solution before updating, T represents vector transpose, and sigmoid represents the S-type function; P=2a·r1-a C＝2r2 In the formula, the range of r1 and r2 is [0, 1], r1 and r2 are randomly generated, and a is the convergence factor; a=2-2t / T max Where, T max is the maximum number of iterations; The update method that meets the first update condition is as follows: x i (t+1)=x * (t)-P·(C·x * (t)-x i (t)) x * (t) represents the historical feature vector with the maximum fitness value after completing t updates, x i (t) represents the updated historical feature vector; The update method that meets the second update condition is as follows: x i (t+1)=x suij (t)-P·(C·x suij (t)-x i (t)) x i (t) represents the updated historical feature vector, x suij (t) represents a random historical feature vector that has completed t updates; The update method that meets the third update condition is as follows: x i (t+1)=[C·x * (t)-x i (t)]·e L cos(2πL)x * (t) In the formula, the value range of L is [-1, 1], L is a random number, and e is a natural constant.

2. An electronic fence monitoring system for a drone using the electronic fence monitoring method for a drone as claimed in claim 1, characterized in that: Includes the following modules: A region division module (100), used for evenly dividing the entire monitoring area into a plurality of sub-areas; A path plan generating module (200), used for randomly generating a plurality of UAV patrol path plans; A scheme coding generation module (300) is used to perform scheme coding on a plurality of randomly generated drone patrol path schemes to generate patrol path feature vectors; A fitness value calculation module (400) is used to calculate the fitness value of the patrol path feature vector of the unmanned aerial vehicle patrol path plan; An updating module (500), used for updating a patrol path feature vector of a patrol path plan of a drone; The decision output module (600) generates patrol path plans for drones in all sub-areas based on the patrol path feature vector of the updated drone patrol path plan.

3. A computer storage medium for storing computer-readable instructions, characterized in that: When executed by a computer system, the computer-readable instructions are capable of executing the electronic fence monitoring method for a drone according to claim 1.

Images