A Substation Radar Layout Method Based on Target Misjudgment Detection

Abstract

The present invention provides a substation radar layout method based on target misjudgment detection, including analyzing the hot zone activity range of target small animals and non-target objects, designing the radar layout space structure; using a two-dimensional uniform grid method to perform initial random layout to obtain a number of radar layout schemes; determining the fitness function of a genetic algorithm based on constraint conditions; using the fitness function to iterate the genetic algorithm to optimize the radar layout scheme; performing cyclic iterations according to the target misjudgment detection requirements of different surfaces to obtain the optimal radar layout scheme corresponding to different surfaces. The present invention can solve the misjudgment detection situation existing in the existing substation small animal intrusion detection, thereby avoiding the problem of high false alarm rate in the process of radar detecting small animals; by analyzing the activity hot zones of target animals and non-target objects, and then laying out radars according to the hot zones, the radar layout scheme is optimized by iterative genetic algorithm through the fitness function, thereby improving the monitoring accuracy of target animals.

Description

Technical Field

[0001] The invention relates to the technical field of radar layout, and in particular to a substation radar layout method based on target misjudgment detection. Background Art

[0002] With the development of economy, the electricity consumption is increasing day by day, and the number of substations is increasing. The operation and maintenance of substations is very important. Substations are complex in environment, with many kinds of equipment and cabinets, which easily attract rodents to build nests. They like to get into the equipment function rooms in substations, chew the transmission wires, touch the switch by mistake, etc., which can easily cause equipment short circuit, insulation damage, switch malfunction and other faults. Therefore, in order to ensure the normal operation of the power system, the invasion of small animals in substations is an important problem that needs to be solved, and how to effectively realize intrusion detection is particularly important.

[0003] At present, with the rapid development of video surveillance technology and machine vision technology, many newly built smart substations and some renovated unmanned distribution rooms are gradually adopting intelligent monitoring technologies such as high-definition video surveillance.

[0004] However, the effect is poor in bad weather such as rain and snow, and it is sensitive to changes in light and cannot work around the clock. In addition, the data captured by the camera can generally only be processed by the host computer to identify and determine the intrusion information, and it is impossible to complete real-time interaction with the expelling device at the intrusion location. Therefore, the existing technology also considers using radar instead of cameras to detect invading small animals. However, in actual application, radar detection is prone to misjudgment detection, resulting in detection alarms for non-target objects such as passing staff, reducing the accuracy of small animal intrusion. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a substation radar layout method based on target misjudgment detection to solve the misjudgment detection problem existing in the existing substation small animal intrusion detection, thereby causing the problem of high false alarm rate in the process of radar detecting small animals.

[0006] The technical solution of the present invention is: a substation radar layout method based on target misjudgment detection, comprising the following steps:

[0007] S1) Analyze the hot zone activity range of target small animals and non-target objects, and design the radar layout space structure;

[0008] S2), using a two-dimensional uniform grid method to perform initial random layout and obtain several radar layout schemes;

[0009] S3), determining the fitness function of the genetic algorithm based on the constraint conditions;

[0010] S4), using the fitness function to iterate the genetic algorithm to optimize the radar layout scheme;

[0011] S5), according to the target misjudgment detection requirements of different surfaces, steps S2)-S4) are looped to obtain the optimal radar layout solutions corresponding to different surfaces.

[0012] Preferably, in step S1), designing the radar layout space structure specifically includes the following steps:

[0013] S11), analyzing the historical intrusion data of small animals in the target substation and the non-target objects that cause misjudgment detection, and obtaining the hot zone activity range of the target small animals and non-target objects;

[0014] S12), constructing a double-layer N-face structure of one layer of regular N-prisms and two layers of regular N-prism pyramids, where N∈[3,10];

[0015] S13), using 3D software to model an assembly model of a regular N-prism and a regular N-pyramid, and then obtaining the desired structure through rotation, translation, and scaling transformation;

[0016] S14), splicing the assembly models of the regular N prisms and the regular N pyramids to form a double-layer N-face radar layout space structure.

[0017] Preferably, in step S14), the first layer of radar in the constructed double-layer N-face radar layout space structure is used to detect the target small animal hot zone; the second layer of radar is used to detect the non-target object hot zone that causes misjudgment, and the misjudgment of non-target objects is reduced by processing and comparing the data returned by the two layers of radar.

[0018] Preferably, in step S2), a two-dimensional uniform grid method is used to perform initial random layout, specifically comprising the following steps:

[0019] S21), dividing the radar layout on each surface of the double-layer N-surface radar layout space structure into an N×N two-dimensional uniform grid, wherein each grid unit represents a position where a radar is to be deployed;

[0020] S22), use a length of 2N 2 The binary string represents the radar layout of the entire double-layer structure surface, where the first N 2 The first layer is represented by the N 2 The bit represents the second layer, 0 means that there is no radar deployed in the grid cell, and 1 means that there is a radar deployed in the grid cell;

[0021] S23), taking the center of the grid unit as the radar signal transmitting point and receiving point, randomly arranging the grid units on each surface of the double-layer N-surface radar layout space structure to obtain several initial radar layout schemes, and judging whether each radar layout scheme can cover the target and non-target hot spots at different angles, if so, selecting this layout scheme, if not, selecting other schemes for judgment.

[0022] Preferably, in step S3), the constraints include radar target hot zone range coverage, radar non-target hot zone coverage, radar non-target hot zone coverage, radar coverage overlap, false positive detection rate and the number of radars.

[0023] Preferably, in step S3), determining the fitness function of the genetic algorithm based on the constraint conditions specifically includes the following steps:

[0024] S31), setting the weights of the radar target hot zone coverage rate, radar non-target hot zone coverage rate, radar non-target hot zone coverage rate, radar coverage overlap, false positive detection rate and radar quantity to be ω respectively. 1 ,ω 2 ,ω 3 ,ω 4 ,ω 5 , and ω 1 +ω 2 +ω 3 +ω 4 +ω 5 =1;

[0025] S32), the expression of the fitness function based on the constraint condition is:

[0026] F=ω 1 ×C 1 (x)+ω 2 ×C 2 (x)-ω 3 ×R(x)-ω 4 ×E(x)-ω 5 ×N(x);

[0027] In the formula, x represents the radar layout plan, and the content is a 2N-length 2 The binary string, C 1 (x) represents the radar target hot zone coverage, C 2 (x) represents the coverage rate of radar non-target hotspots, R(x) represents the coverage overlap rate of the first-layer radar and the second-layer radar, E(x) represents the false detection rate, and N(x) represents the ratio of the number of radars to the grid cells.

[0028] As a preferred embodiment, in step S32), the radar target hot zone coverage C 1(x), radar non-target hot zone coverage rate C 2 The expressions of (x) are:

[0029]

[0030]

[0031] The expression of the coverage overlap ratio R(x) of the first-layer radar and the second-layer radar is:

[0032]

[0033] Where n 1 S c represents the total coverage area of ​​a layer of radar at the maximum detection distance r, S ij Represents the projected area of ​​the target hot zone within the range of elevation angle i and azimuth angle j;

[0034] n 2 S c represents the total coverage area of ​​the second-layer radar at the maximum detection distance r, S i ' j represents the projection area of ​​the non-target hot zone within the range of pitch angle i and azimuth angle j; S c represents the projected coverage area of ​​a single radar at the maximum detection distance r, P 2 represents the radar transmit power, G represents the antenna gain, λ represents the radar wavelength, σ represents the target radar cross-sectional area, k represents the Boltzmann constant, T is the system temperature, B is the receiver bandwidth, and f is the noise factor.

[0035] Preferably, in step S32), the expression of the false positive detection rate E(x) is:

[0036]

[0037] Where FP(x) is the number of times the radar mistakenly identifies a non-target as a target, FN(x) is the number of times the radar fails to detect a target, TP(x) is the number of times the radar correctly identifies a target as a target, and TN(x) is the number of times the radar correctly identifies a non-target as a non-target.

[0038] Preferably, in step S32), the expression of the ratio N(x) of the number of radars to the grid units is:

[0039]

[0040] In the formula, if radars are deployed in i grid units, then d i =1, otherwise d i =0.

[0041] Preferably, in step S4), the fitness function is used to perform genetic algorithm iteration to optimize the radar layout scheme, which specifically includes the following steps:

[0042] S41), determine the size of the primary population, set the number of individuals in the primary population, use the radar layout scheme selected in step S23) as the first generation parent population, pre-set the crossover probability and mutation probability, and perform the three sub-steps of selection, crossover, and mutation in sequence to generate the offspring population generation by generation;

[0043] S42), when executing the selection sub-step, use the tournament selection strategy, set a fixed tournament size, and randomly select a certain number of individuals from the population to compete in each selection.

[0044] Then select the individuals with the highest fitness to enter the next generation;

[0045] S43), when executing the intersection sub-step, the radar layout area is regarded as a weighted undirected graph, the nodes represent the positions of the radars, the edges represent the connections between the nodes, and the connectivity is judged by the minimum spanning tree algorithm;

[0046] S44), when executing the mutation sub-step, all individuals in the group are judged whether to mutate with the set mutation probability, two gene sequences of the same length are randomly selected, and a new individual is obtained by using an exchange mutation method;

[0047] S45) sets the number of iterations of the genetic optimization process. If the number of iterations set value is not reached, the genetic population continues to cyclically execute the genetic optimization operations described in steps S42)-S44); if the number of iterations set value is reached, the individual with the highest fitness value in the genetic population is selected as the optimal radar layout solution, and the radar layout optimization process ends.

[0048] Preferably, in step S43), the crossover sub-step is performed in the following manner:

[0049] S431), for individuals in the population, first determine whether they participate in the crossover operation, specifically: generate a random number between 0 and 1, if the random number is greater than the crossover probability, the individual directly skips the crossover operation; if the random number is less than the crossover probability, the individual is paired, and the optional paired individuals include all other individuals except itself;

[0050] S432), considering the radar layout area as a weighted undirected graph D, the position of each radar corresponds to a node in the graph, the edge between the nodes represents the connection between the radars, and the weight of the edge represents the cost of the connection;

[0051] S433), constructing a minimum spanning tree A from the undirected graph D using the Kruskal algorithm, wherein the minimum spanning tree A will contain all nodes and connect these nodes with the minimum total weight to ensure connectivity;

[0052] S434), in the crossover operation, in order to ensure that the offspring individuals after the crossover are connected, the minimum spanning tree of one of the parent individuals is selected as the basis of the crossover operation, and then the missing part of the other parent individual is added to the basic minimum spanning tree to generate a new offspring individual;

[0053] S435) Select offspring individuals with high fitness to enter the next generation population.

[0054] The beneficial effects of the present invention are:

[0055] 1. The present invention can solve the problem of misjudgment detection in the existing substation small animal intrusion detection, thereby avoiding the problem of high false alarm rate in the process of radar detecting small animals;

[0056] 2. The present invention analyzes the activity hotspots of target animals and non-target objects, and then deploys radars according to the hotspots. The present invention also iterates the genetic algorithm through the fitness function to optimize the radar layout plan, thereby improving the monitoring of target animals and avoiding interference of non-target objects on the monitoring results. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic diagram of the process of the present invention;

[0058] Figure 2 It is a schematic diagram of a flow chart of the method of the present invention iterating according to the requirements of target misjudgment detection on different surfaces; DETAILED DESCRIPTION

[0059] The specific implementation of the present invention will be further described below in conjunction with the accompanying drawings:

[0060] Example 1

[0061] like Figure 1 As shown, this embodiment provides a substation radar layout method based on target misjudgment detection to solve the misjudgment detection situation existing in the existing substation small animal intrusion detection, thereby causing the problem of high false alarm rate in the radar detection of small animals, including the following steps:

[0062] S1) Analyze the hot zone activity range of target small animals and non-target objects, and design the radar layout space structure;

[0063] S2), using a two-dimensional uniform grid method to perform initial random layout and obtain several radar layout schemes;

[0064] S3), determining the fitness function of the genetic algorithm based on the constraint conditions; in this embodiment, the constraint conditions include radar target hot zone coverage, radar non-target hot zone coverage, radar non-target hot zone coverage, radar coverage overlap, false positive detection rate and the number of radars;

[0065] S4), using the fitness function to iterate the genetic algorithm to optimize the radar layout scheme;

[0066] S5), according to the target misjudgment detection requirements of different surfaces, the steps S2)-S4) are looped to obtain the optimal radar layout scheme corresponding to different surfaces. For details, please refer to Figure 2 shown.

[0067] Example 2

[0068] On the basis of Example 1, in step S1), designing the radar layout space structure specifically includes the following steps:

[0069] S11), analyzing the historical intrusion data of small animals in the target substation and the non-target objects that cause misjudgment detection, and obtaining the hot zone activity range of the target small animals and non-target objects;

[0070] S12), constructing a double-layer N-face structure of one layer of regular N-prisms and two layers of regular N-prism pyramids, where N∈[3,10];

[0071] S13), using 3D software to model an assembly model of a regular N-prism and a regular N-pyramid, and then obtaining the desired structure through rotation, translation, and scaling transformation;

[0072] S14), splicing the assembly models of the regular N prisms and the regular N pyramids to form a double-layer N-face radar layout space structure.

[0073] In addition, the first layer of radar in the double-layer N-face radar layout space structure constructed in this embodiment is used to detect the target small animal hot zone; the second layer of radar is used to detect the non-target object hot zone that causes misjudgment, and the misjudgment of non-target objects is reduced by processing and comparing the data returned by the two layers of radar.

[0074] Example 3

[0075] On the basis of Example 1, in step S2), a two-dimensional uniform grid method is used to perform initial random layout, which specifically includes the following steps:

[0076] S21), dividing the radar layout on each surface of the double-layer N-surface radar layout space structure into an N×N two-dimensional uniform grid, wherein each grid unit represents a position where a radar is to be deployed;

[0077] S22), use a length of 2N 2The binary string encoding rules are used to represent the radar layout of the entire double-layer structure surface, where the first N 2 The first layer is represented by the N 2 The bit represents the second layer, 0 means that there is no radar deployed in the grid cell, and 1 means that there is a radar deployed in the grid cell;

[0078] S23), taking the center of the grid unit as the radar signal transmitting point and receiving point, randomly arranging the grid units on each surface of the double-layer N-surface radar layout space structure to obtain several initial radar layout schemes, and judging whether each radar layout scheme can cover the target and non-target hot spots at different angles, if so, selecting this layout scheme, if not, selecting other schemes for judgment.

[0079] Example 4

[0080] On the basis of Example 1, the fitness function of the genetic algorithm is determined based on the constraint conditions in step S3), which specifically includes the following steps:

[0081] S31), setting the weights of the radar target hot zone coverage rate, radar non-target hot zone coverage rate, radar non-target hot zone coverage rate, radar coverage overlap, false positive detection rate and radar quantity to be ω respectively. 1 ,ω 2 ,ω 3 ,ω 4 ,ω 5 , and ω 1 +ω 2 +ω 3 +ω 4 +ω 5 =1;

[0082] S32), the expression of the fitness function based on the constraint condition is:

[0083] F=ω 1 ×C 1 (x)+ω 2 ×C 2 (x)-ω 3 ×R(x)-ω 4 ×E(x)-ω 5 ×N(x);

[0084] In the formula, x represents the radar layout plan, and the content is a 2N-length 2 The binary string, C 1 (x) represents the radar target hot zone coverage, C 2 (x) represents the coverage rate of radar non-target hotspots, R(x) represents the coverage overlap rate of the first-layer radar and the second-layer radar, E(x) represents the false detection rate, and N(x) represents the ratio of the number of radars to the grid cells.

[0085] As a preferred embodiment of this invention, in step S32), the radar target hot zone coverage C 1 (x), radar non-target hot zone coverage rate C 2 The expressions of (x) are:

[0086]

[0087]

[0088] The expression of the coverage overlap ratio R(x) of the first-layer radar and the second-layer radar is:

[0089]

[0090] Where n 1 S c represents the total coverage area of ​​a layer of radar at the maximum detection distance r, S ij Represents the projected area of ​​the target hot zone within the range of elevation angle i and azimuth angle j;

[0091] n 2 S c represents the total coverage area of ​​the second-layer radar at the maximum detection distance r, S i ' j represents the projection area of ​​the non-target hot zone within the range of pitch angle i and azimuth angle j; S c represents the projected coverage area of ​​a single radar at the maximum detection distance r, P 2 represents the radar transmit power, G represents the antenna gain, λ represents the radar wavelength, σ represents the target radar cross-sectional area, k represents the Boltzmann constant, T is the system temperature, B is the receiver bandwidth, and f is the noise factor.

[0092] The expression of the false positive detection rate E(x) is:

[0093]

[0094] Where FP(x) is the number of times the radar mistakenly identifies a non-target as a target, FN(x) is the number of times the radar fails to detect a target, TP(x) is the number of times the radar correctly identifies a target as a target, and TN(x) is the number of times the radar correctly identifies a non-target as a non-target.

[0095] The expression of the ratio N(x) of the number of radars to the grid cells is:

[0096]

[0097] In the formula, if radars are deployed in i grid units, then d i=1, otherwise d i =0.

[0098] Example 5

[0099] On the basis of Example 1, in step S4), the fitness function is used to perform genetic algorithm iteration to optimize the radar layout scheme, which specifically includes the following steps:

[0100] S41), determine the size of the primary population, set the number of individuals in the primary population, use the radar layout scheme selected in step S23) as the first generation parent population, pre-set the crossover probability and mutation probability, and perform the three sub-steps of selection, crossover, and mutation in sequence to generate the offspring population generation by generation;

[0101] S42), when executing the selection sub-step, use the tournament selection strategy, set a fixed tournament size, randomly select a certain number of individuals from the population to compete in each selection, and then select the individuals with the highest fitness to enter the next generation;

[0102] S43), when executing the intersection sub-step, the radar layout area is regarded as a weighted undirected graph, the nodes represent the positions of the radars, the edges represent the connections between the nodes, and the connectivity is judged by the minimum spanning tree algorithm;

[0103] S44), when executing the mutation sub-step, all individuals in the group are judged whether to mutate with the set mutation probability, two gene sequences of the same length are randomly selected, and a new individual is obtained by using an exchange mutation method;

[0104] S45) sets the number of iterations of the genetic optimization process. If the number of iterations set value is not reached, the genetic population continues to cyclically execute the genetic optimization operations described in steps S42)-S44); if the number of iterations set value is reached, the individual with the highest fitness value in the genetic population is selected as the optimal radar layout solution, and the radar layout optimization process ends.

[0105] As a preferred embodiment of this invention, in step S43), the crossover sub-step is specifically performed in the following manner:

[0106] S431), for individuals in the population, first determine whether they participate in the crossover operation, specifically: generate a random number between 0 and 1, if the random number is greater than the crossover probability, the individual directly skips the crossover operation; if the random number is less than the crossover probability, the individual is paired, and the optional paired individuals include all other individuals except itself;

[0107] S432), considering the radar layout area as a weighted undirected graph D, the position of each radar corresponds to a node in the graph, the edge between the nodes represents the connection between the radars, and the weight of the edge represents the cost of the connection;

[0108] S433), constructing a minimum spanning tree A from the undirected graph D using the Kruskal algorithm, wherein the minimum spanning tree A will contain all nodes and connect these nodes with the minimum total weight to ensure connectivity;

[0109] S434), in the crossover operation, in order to ensure that the offspring individuals after the crossover are connected, the minimum spanning tree of one of the parent individuals is selected as the basis of the crossover operation, and then the missing part of the other parent individual is added to the basic minimum spanning tree to generate a new offspring individual;

[0110] S435) Select offspring individuals with high fitness to enter the next generation population.

[0111] The above embodiments and descriptions are only for illustrating the principles and best embodiments of the present invention. Without departing from the spirit and scope of the present invention, the present invention may be subject to various changes and improvements, all of which fall within the scope of the present invention to be protected.

Claims

1. A substation radar layout method based on target misjudgment detection, characterized in that: The following steps are involved: S1), analyzing the hot zone activity range of the target small animals and non-target objects, and designing the radar layout space structure; wherein, designing the radar layout space structure specifically includes the following steps: S11), analyzing the historical intrusion data of small animals in the target substation and the non-target objects that cause misjudgment detection, and obtaining the hot zone activity range of the target small animals and non-target objects; S12), constructing a double-layer N-face structure of one layer of regular N-prisms and two layers of regular N-prism pyramids, where N∈[3,10]; S13), using 3D software to model an assembly model of a regular N-prism and a regular N-pyramid, and then obtaining the desired structure through rotation, translation, and scaling transformation; S14), splicing the assembly models of the regular N prisms and the regular N pyramids to form a double-layer N-face radar layout space structure; wherein the first layer of radar in the constructed double-layer N-face radar layout space structure is used to detect the target small animal hot zone; the second layer of radar is used to detect the non-target object hot zone that causes misjudgment, and the misjudgment of non-target objects is reduced by processing and comparing the data returned by the two layers of radars; S2), using a two-dimensional uniform grid method to perform initial random layout and obtain several radar layout schemes; S3), determining the fitness function of the genetic algorithm based on the constraint conditions; S4), using the fitness function to iterate the genetic algorithm to optimize the radar layout scheme; S5), according to the target misjudgment detection requirements of different surfaces, steps S2)-S4) are looped to obtain the optimal radar layout solutions corresponding to different surfaces.

2. The substation radar layout method based on target misjudgment detection according to claim 1 is characterized in that: In step S2), a two-dimensional uniform grid method is used to perform an initial random layout, specifically comprising the following steps: S21), dividing the radar layout on each surface of the double-layer N-surface radar layout space structure into an N×N two-dimensional uniform grid, wherein each grid unit represents a position where a radar is to be deployed; S22), use a length of 2N 2 The binary string encoding is used to represent the radar layout of the entire double-layer structure surface, where the first N 2 The first layer is represented by the N 2 The bit represents the second layer, 0 means that there is no radar deployed in the grid cell, and 1 means that there is a radar deployed in the grid cell; S23), taking the center of the grid unit as the radar signal transmitting point and receiving point, randomly arranging the grid units on each surface of the double-layer N-surface radar layout space structure to obtain several initial radar layout schemes, and judging whether each radar layout scheme can cover the target and non-target hot spots at different angles, if so, selecting this layout scheme, if not, selecting other schemes for judgment.

3. The substation radar layout method based on target misjudgment detection according to claim 1 is characterized in that: In step S3), the constraints include radar target hot zone coverage, radar non-target hot zone coverage, radar non-target hot zone coverage, radar coverage overlap, false positive detection rate and the number of radars.

4. The substation radar layout method based on target misjudgment detection according to claim 3 is characterized in that: In step S3), the fitness function of the genetic algorithm is determined based on the constraint conditions, which specifically includes the following steps: S31), setting the weights of the radar target hot zone coverage rate, the radar non-target hot zone coverage rate, the radar non-target hot zone coverage rate, the radar coverage overlap, the false positive detection rate and the number of radars to ω1, ω2, ω3, ω4 and ω5 respectively, and ω1+ω2+ω3+ω4+ω5=1; S32), the expression of the fitness function based on the constraint condition is: F=ω1×C1(x)+ω2×C2(x)-ω3×R(x)-ω4×E(x)-ω5×N(x); In the formula, x represents the radar layout plan, and the content is a 2N-length 2 C1(x) represents the coverage rate of radar target hotspot, C2(x) represents the coverage rate of radar non-target hotspot, R(x) represents the coverage overlap rate of the first-layer radar and the second-layer radar, E(x) represents the false positive detection rate, and N(x) represents the ratio of the number of radars to the grid cells.

5. The substation radar layout method based on target misjudgment detection according to claim 4 is characterized in that: In step S32), the expressions of the radar target hot zone coverage rate C1(x) and the radar non-target hot zone coverage rate C2(x) are respectively: The expression of the coverage overlap ratio R(x) of the first-layer radar and the second-layer radar is: Where n1S c represents the total coverage area of ​​a layer of radar at the maximum detection distance r, S ij Represents the projected area of ​​the target hot zone within the range of elevation angle i and azimuth angle j; n2S c represents the total coverage area of ​​the second-layer radar at the maximum detection distance r, S i ' j represents the projection area of ​​the non-target hot zone within the range of pitch angle i and azimuth angle j; S c represents the projected coverage area of ​​a single radar at the maximum detection distance r, P2 represents the radar transmit power, G represents the antenna gain, λ represents the radar wavelength, σ represents the target radar cross-sectional area, k represents the Boltzmann constant, T is the system temperature, B is the receiver bandwidth, and f is the noise factor.

6. The substation radar layout method based on target misjudgment detection according to claim 4 is characterized in that: In step S32), the expression of the false positive detection rate E(x) is: Where FP(x) is the number of times the radar mistakenly identifies a non-target as a target, FN(x) is the number of times the radar fails to detect a target, TP(x) is the number of times the radar correctly identifies a target as a target, and TN(x) is the number of times the radar correctly identifies a non-target as a non-target.

7. The substation radar layout method based on target misjudgment detection according to claim 4 is characterized in that: In step S32), the expression of the ratio N(x) of the number of radars to the grid units is: In the formula, if the radar is deployed in the i-th grid unit, then d i =1, otherwise d i =0.

8. The substation radar layout method based on target misjudgment detection according to claim 1 is characterized in that: In step S4), the fitness function is used to iterate the genetic algorithm to optimize the radar layout scheme, which specifically includes the following steps: S41), determine the size of the primary population, set the number of individuals in the primary population, use the radar layout scheme selected in step S23) as the first generation parent population, pre-set the crossover probability and mutation probability, and perform the three sub-steps of selection, crossover, and mutation in sequence to generate the offspring population generation by generation; S42), when executing the selection sub-step, use the tournament selection strategy, set a fixed tournament size, randomly select a certain number of individuals from the population to compete in each selection, and then select the individuals with the highest fitness to enter the next generation; S43), when executing the intersection sub-step, the radar layout area is regarded as a weighted undirected graph, the nodes represent the positions of the radars, the edges represent the connections between the nodes, and the connectivity is judged by the minimum spanning tree algorithm; S44), when executing the mutation sub-step, all individuals in the group are judged whether to mutate with the set mutation probability, two gene sequences of the same length are randomly selected, and a new individual is obtained by using an exchange mutation method; S45) sets the number of iterations of the genetic optimization process. If the number of iterations set value is not reached, the genetic population continues to cyclically execute the genetic optimization operations described in steps S42)-S44); if the number of iterations set value is reached, the individual with the highest fitness value in the genetic population is selected as the optimal radar layout solution, and the radar layout optimization process ends.

9. The substation radar layout method based on target misjudgment detection according to claim 8 is characterized in that: In step S43), the crossover sub-step is specifically performed in the following manner: S431), for individuals in the population, first determine whether they participate in the crossover operation, specifically: generate a random number between 0 and 1, if the random number is greater than the crossover probability, the individual directly skips the crossover operation; if the random number is less than the crossover probability, the individual is paired, and the optional paired individuals include all other individuals except itself; S432), considering the radar layout area as a weighted undirected graph D, the position of each radar corresponds to a node in the graph, the edge between the nodes represents the connection between the radars, and the weight of the edge represents the cost of the connection; S433), constructing a minimum spanning tree A from the undirected graph D using the Kruskal algorithm, wherein the minimum spanning tree A will contain all nodes and connect these nodes with the minimum total weight to ensure connectivity; S434), in the crossover operation, in order to ensure that the offspring individuals after the crossover are connected, the minimum spanning tree of one of the parent individuals is selected as the basis of the crossover operation, and then the missing part of the other parent individual is added to the basic minimum spanning tree to generate a new offspring individual; S435) Select offspring individuals with high fitness to enter the next generation population.

Images