Electric power station building personnel intrusion detection method based on wireless channel state information

Abstract

The invention relates to an electric power station building personnel intrusion detection method based on wireless channel state information, comprising the following steps: deploying WiFi-CSI sensors in an electric power station building, and collecting CSI data of multiple subcarriers in real time through a mutual sniffing mode; performing wavelet transform processing on the CSI data by using an adaptive denoising algorithm to realize separation of a personnel activity signal and an electric equipment interference signal of the power station building; constructing a phase-amplitude correction model, eliminating the influence of a multi-metal structure on the multipath effect of the CSI signal, and extracting CSI features related to personnel movement; dividing the station building into a plurality of detection areas, and identifying personnel activity signals and area positions by adopting a dynamic matching algorithm; and further judging an intrusion detection result in combination with an access control state and equipment operation parameters. According to the invention, the requirements of real-time performance, reliability, high precision and low cost of personnel intrusion detection of the power station building can be met.

Description

Technical Field

[0001] The present invention relates to the field of power substation personnel intrusion detection, and in particular to a power substation personnel intrusion detection method based on wireless channel state information. Background Art

[0002] As the core hub of the power grid, the physical security of power substations directly impacts the stability of grid operations. Unauthorized access can lead to equipment damage, misoperation, or malicious sabotage, potentially causing regional power outages or equipment failures, resulting in significant economic losses and safety incidents. Therefore, detecting intrusions into power substations is crucial to maintaining the safe operation of the power system.

[0003] Traditional methods for detecting intrusion include infrared sensors and cameras. Infrared sensor-based methods require precise sensor placement to avoid blind spots, increasing installation difficulty and cost. They are also susceptible to environmental factors such as lighting, temperature, and obstruction. Camera-based methods also require precise camera placement to avoid blind spots, but they also pose privacy concerns and are susceptible to lighting.

[0004] In recent years, intrusion detection technology based on WiFi channel state information (CSI) has gained momentum in home environments. This technology, which analyzes WiFi signal changes to detect human activity, offers the advantages of low cost and non-contact. However, existing WiFi-CSI technology has not yet been applied to power substations. The unique environment (such as electromagnetic interference, multiple metal reflections, and enclosed spaces) significantly affects signal propagation, making it difficult to directly adapt general methods. Summary of the Invention

[0005] In order to solve the above problems, the purpose of the present invention is to provide a method for detecting personnel intrusion in power substations based on wireless channel state information, which solves the problem that the existing WiFi-CSI technology has not yet been applied to power substations, and its special environment (such as electromagnetic interference, multi-metal reflection, and closed space) has a significant impact on signal propagation, making it difficult for general methods to be directly adapted.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A method for detecting personnel intrusion into a power substation based on wireless channel state information comprises the following steps:

[0008] S1: Deploy WiFi-CSI sensors in the power substation and place them in the same local area network as the existing WiFi devices in the power substation, enabling communication and data transmission;

[0009] S2: The WiFi-CSI sensors in the power station room use MAC codes to identify each other to sniff WiFi signals in real time, and then parse and process CSI data packets;

[0010] S3: Preprocess the collected CSI data packets using an adaptive denoising algorithm;

[0011] S4: Based on the phase-amplitude correction model, the phase and amplitude of the CSI data affected by multipath signals are corrected for the preprocessed CSI data, signal fluctuations caused by non-target disturbances are eliminated, and CSI features related to human activities are extracted;

[0012] S5: Divide the distribution station into multiple detection areas based on the specific functional areas of the distribution station. Use a dynamic matching algorithm to identify whether human activity occurs and locate the areas where human activity occurs based on CSI features related to human activity.

[0013] S6: Perform linkage composite verification based on the results of the dynamic matching algorithm in combination with the access control status or equipment operating parameters (such as current or voltage changes).

[0014] Furthermore, the CSI data includes the amplitude and phase of 62 subcarriers with timestamps, 11 of which are null subcarriers. Before subsequent data processing, the null subcarriers in the CSI data are removed, that is, 51 valid subcarriers.

[0015] Further preprocessing is as follows:

[0016] The CSI signal is divided into multiple time regions in chronological order according to the time interval Δt, and the average value u of the CSI amplitude in each time region is calculated. amp (t region ) and standard deviation α amp (t region )calculate:

[0017]

[0018] Where N is the number of CSI signal samples in the time zone; CSI amp,i (t region ) is the amplitude of the i-th sample;

[0019] According to the average value u of the amplitude in this time region amp (t region ) and standard deviation α amp (t region ), construct the normalized basis of the amplitude, and for each CSI signal amplitude sample in the time region, use the following formula for normalization:

[0020]

[0021] Symlet wavelet and 3-layer decomposition are used to perform wavelet decomposition on the normalized CSI amplitude data. Wavelet decomposition decomposes the original CSI signal into sub-band signals in different frequency ranges.

[0022] Dynamically adjust the threshold according to the local characteristics of the CSI signal, perform threshold processing on the sub-band signals in different frequency ranges, and perform wavelet reconstruction on the sub-band signals in different frequency ranges after threshold processing;

[0023] Based on the frequency distribution characteristics of the signal and the difference in frequency range between electrical equipment signals and human activity signals in actual scenarios, a reasonable frequency threshold (for example, 10 Hz) is set as the basis for separation, and human activity signals and electrical equipment signals are separated through frequency analysis.

[0024] Furthermore, the phase-amplitude correction model is as follows:

[0025] Calculate the phase difference of each subcarrier data of the CSI signal, the phase difference Δφ of the Nth subcarrier at time t N (t) are as follows:

[0026] Δφ N (t) = φ N (t)-φ re (t)

[0027] where φ N (t) represents the phase difference of the Nth subcarrier at time t, φ re (t) represents the phase of the reference subcarrier at time t;

[0028] According to the phase difference, the phase of the subcarrier is corrected to obtain the phase φ of the Nth subcarrier after phase correction at time t N-correct (t):

[0029] φ N-correct (t) = φ N (t)-Δφ N (t)

[0030] For each subcarrier’s CSI data, calculate the ratio of the Nth subcarrier amplitude at time t to the reference signal amplitude:

[0031]

[0032] Among them A N (t) represents the amplitude of the Nth subcarrier at time t, A re (t) represents the amplitude of the reference subcarrier at time t;

[0033] According to the amplitude ratio, the amplitude of the Nth subcarrier at time t is corrected:

[0034]

[0035] Furthermore, the dynamic matching algorithm is as follows:

[0036] Collect CSI signal templates for areas without personnel activities and areas with personnel activities in advance, and obtain real-time CSI signals. The specific structures of the CSI signal template T and the real-time CSI signal S are as follows:

[0037] T=[t1,t2,...,t n ]

[0038] Each t j is a vector containing the amplitude and phase of 51 subcarriers;

[0039] S=[s1,s2,....,s m ]

[0040] Each of these s i is a vector containing the amplitude and phase of 51 subcarriers;

[0041] A threshold is set to calculate the optimal matching total similarity between the real-time CSI signal and the CSI signal template indicating no human activity in the area. If the optimal matching total similarity is higher than the threshold, it is considered that there is no human activity. If the optimal matching total similarity is lower than the threshold, it is considered that there may be human activity.

[0042] The real-time CSI signal is compared with the CSI signal templates of each area with human activity for the best matching total similarity, and the area with the highest best matching total similarity among all area templates is selected as the area of ​​human activity.

[0043] Furthermore, the steps for calculating the total similarity of the CSI optimal match are as follows:

[0044] For each subcarrier k (k = 1, 2, ..., 51), the Euclidean distance of its amplitude and phase is expressed as:

[0045]

[0046] in, and is the amplitude and phase of the kth subcarrier of the ith signal in the real-time CSI signal S, and is the amplitude and phase of the kth subcarrier of the jth signal in the template CSI signal T;

[0047] In order to obtain the total similarity between two signal points, the similarities of all subcarriers are summed and averaged. The total similarity calculation formula is as follows:

[0048]

[0049] Then, the dynamic matching algorithm DTW is used to obtain the optimal matching total similarity. An m×n similarity matrix D is defined, where D[i][j] represents the optimal matching total similarity between the first i signals of the real-time CSI signal S and the first j signals of the template signal T. The recursive formula is as follows:

[0050]

[0051] Start from the lower right corner of the matrix D[m][n] and backtrack to find the optimal matching path. According to the recursive formula, select the path direction corresponding to the minimum value.

[0052] Finally, the optimal matching total similarity D[i][j] between the CSI signal and the CSI signal template is obtained in real time by calculation.

[0053] Furthermore, linkage composite verification is carried out as follows: the power station intrusion detection system automatically obtains the access control status signal and equipment operating parameters. When a personnel activity signal is detected, if the access control status is open, it is judged as normal personnel activity; if the access control status is closed, it is judged as personnel intrusion activity. In addition, whether there is abnormal fluctuation in the equipment operating parameters is combined to indicate whether the personnel intrusion activity has an impact on the electrical equipment.

[0054] A power substation personnel intrusion detection system based on wireless channel state information includes a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically performs the steps of the power substation personnel intrusion detection method based on wireless channel state information as described above.

[0055] The present invention has the following beneficial effects:

[0056] 1. This invention utilizes existing WiFi infrastructure to deploy power substations, reducing the hardware cost of intrusion detection in power substations and the complexity of installation and deployment, making it easy to popularize. It also has the function of ensuring intrusion detection in dark environments, is not affected by light, and has stability.

[0057] 2. The present invention adopts the wavelet transform of the adaptive denoising algorithm and uses the phase-amplitude correction model to greatly reduce the interference of electrical equipment and metal cabinets brought by the power station on the CSI signal, and makes the personnel intrusion detection method reliable by combining the access control status or equipment operation parameters; by obtaining the CSI signal in real time, it can identify whether the power station is invaded in real time, which is real-time. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 Flow chart of the method of the present invention;

[0059] Figure 2 Schematic diagram of device communication domain in one embodiment of the present invention. DETAILED DESCRIPTION

[0060] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0061] refer to Figure 1 In this embodiment, a method for detecting personnel intrusion in a power substation based on wireless channel state information is provided. Multiple WiFi-CSI sensors are deployed to sniff each other and collect CSI data. An adaptive denoising algorithm is used to separate personnel activity signals and electrical equipment signals from the CSI data. A phase-amplitude correction model is used to eliminate the influence of multipath effects and extract CSI features related to personnel movement. The key features of CSI personnel movement are sent to a dynamic matching algorithm for detection. If no personnel signal is detected, it is determined that there is no personnel activity. If a personnel signal is detected, whether an intrusion has occurred is determined based on whether the access control is open or closed. If the access control is open, it is normal personnel activity. If the access control is closed, whether an intrusion activity has occurred and has an impact on the electrical equipment is determined based on the parameter status of the electrical equipment. If the parameter status is normal, it is determined that a personnel intrusion activity has occurred. If the parameter status is abnormal, it is determined that a personnel intrusion activity has occurred and has an impact on the electrical equipment. The details are as follows:

[0062] S1: Based on the equipment layout and electromagnetic environment of the power station, the location and direction of the WiFi-CSI sensors should be reasonably arranged. The sensors should be placed in areas with frequent human activity, such as equipment operation areas and inspection passages. Taking a power distribution station as an example, the sensors were installed in the aisles between the power station equipment using brackets. The horizontal distance from the sensors to the metal equipment should be kept at 1 meter, and the installation height should be 1.3 meters. The sensor antenna should be installed perpendicular to the ground. A total of 8 WiFi-CSI sensors were deployed to cover all areas with frequent human activity, such as Figure 2 As shown, the existing WiFi router in the power station is used to make all WiFi-CSI sensors in the same local area network and able to communicate and transmit data;

[0063] S2: WiFi-CSI sensors within the same local area network use MAC codes to identify each other and sniff WiFi signals in real time, then parse and process CSI data packets. The collected CSI data includes the amplitude and phase of 62 subcarriers with timestamps, 11 of which are null subcarriers. Before subsequent data processing, the null subcarriers in the CSI data must be removed, resulting in 51 valid subcarriers. To address the complex electromagnetic interference characteristics within power substations, this paper designs an adaptive denoising algorithm to preprocess the collected CSI data.

[0064] Preferably, in this embodiment, the preprocessing is as follows:

[0065] Determine the time interval Δt = 10s and divide the CSI signal into multiple time zones in time sequence. The average value u of the CSI amplitude in each time zone is calculated. amp (t region ) and standard deviation α amp (t region )calculate:

[0066]

[0067] Where N is the number of CSI signal samples in the time zone; CSI amp,i (t region ) is the amplitude of the i-th sample;

[0068] According to the average value u of the amplitude in this time region amp (t region ) and standard deviation α amp (t region ), construct the normalized basis of the amplitude, and for each CSI signal amplitude sample in the time region, use the following formula for normalization:

[0069]

[0070] Symlet wavelet and 3-layer decomposition are used to perform wavelet decomposition on the normalized CSI amplitude data. Wavelet decomposition decomposes the normalized CSI signal into sub-band signals of different frequency ranges.

[0071] To address high-frequency noise interference from some equipment in power scenarios, an adaptive threshold is used to dynamically adjust the threshold based on the local characteristics of the CSI signal. Sub-band signals in different frequency ranges are thresholded and then reconstructed using wavelet transforms. This removes some of the high-frequency noise interference while retaining the CSI signal's resident electrical equipment and human activity signal characteristics.

[0072] After high-frequency noise removal, the signal's frequency characteristics are further analyzed to separate human activity signals from equipment interference signals. Based on the signal's frequency distribution characteristics and the frequency range differences between electrical equipment signals and human activity signals in actual scenarios, a frequency threshold of 10Hz is set as the basis for separation. Frequency analysis is then used to separate human activity signals from electrical equipment signals.

[0073] S3: To address the multipath effect caused by multiple metal structures in the power station, the present invention constructs a phase-amplitude correction model. This model corrects the phase and amplitude of the multipath signal, eliminates signal fluctuations caused by non-target disturbances, and extracts CSI features related to human movement. The specific steps are as follows:

[0074] Select a subcarrier with high signal strength and small phase change from the processed CSI signal as a reference subcarrier;

[0075] Calculate the phase difference of each subcarrier data of the CSI signal, the phase difference Δφ of the Nth subcarrier at time t N (t) are as follows:

[0076] Δφ N (t) = φ N (t)-φ re (t)

[0077] where φ N (t) represents the phase difference of the Nth subcarrier at time t, φ re (t) represents the phase of the reference subcarrier at time t;

[0078] According to the phase difference, the phase of the subcarrier is corrected to obtain the phase φ of the Nth subcarrier after phase correction at time t N-correct (t):

[0079] φ N-correct (t) = φ N (t)-Δφ N (t)

[0080] For each subcarrier’s CSI data, calculate the ratio of the Nth subcarrier amplitude at time t to the reference signal amplitude:

[0081]

[0082] Among them A N (t) represents the amplitude of the Nth subcarrier at time t, A re (t) represents the amplitude of the reference subcarrier at time t;

[0083] According to the amplitude ratio, the amplitude of the Nth subcarrier at time t is corrected:

[0084]

[0085] S4: Divide the distribution station into multiple detection areas according to the specific functional areas of the distribution station or according to the more precise subdivision of the area according to the grid; taking a distribution station as an example, it can be divided into four detection areas according to the functional areas: high-voltage equipment area, low-voltage equipment area, control room, and auxiliary facilities area; further subdivide the four functional areas in the actual distribution station and divide the areas according to the equipment and equipment auxiliary space in each functional area. The size of each grid is about 3-6m 2 ; Use dynamic matching algorithm to identify whether human activity occurs and locate the area where human activity occurs based on the extracted human activity signal features;

[0086] Preferably, in this embodiment, the specific steps of the dynamic matching algorithm are as follows:

[0087] Collect CSI signal templates for areas without personnel activities and areas with personnel activities in advance, and obtain real-time CSI signals. The specific structures of the CSI signal template T and the real-time CSI signal S are as follows:

[0088] T=[t1,t2,...,t n ]

[0089] Each t j is a vector containing the amplitude and phase of 51 subcarriers;

[0090] S=[s1,s2,....,s m ]

[0091] Each of these s i is a vector containing the amplitude and phase of 51 subcarriers;

[0092] A threshold is set and the best match total similarity is calculated between the real-time CSI signal and the CSI signal template with no human activity in the area. If the best match total similarity is higher than the threshold, it is considered that there is no human activity. If the best match total similarity is lower than the threshold, it is considered that there may be human activity. The steps for calculating the best match total similarity of CSI are as follows:

[0093] For each subcarrier k (k = 1, 2, ..., 51), the Euclidean distance of its amplitude and phase can be expressed as:

[0094]

[0095] in, and is the amplitude and phase of the kth subcarrier of the ith signal in the real-time CSI signal S, and is the amplitude and phase of the kth subcarrier of the jth signal in the template CSI signal T;

[0096] In order to obtain the total similarity between two signal points, the similarities of all subcarriers are summed and averaged. The total similarity calculation formula is as follows:

[0097]

[0098] Then, the dynamic matching algorithm DTW is used to obtain the optimal matching total similarity. An m×n similarity matrix D is defined, where D[i][j] represents the optimal matching total similarity between the first i signals of the real-time CSI signal S and the first j signals of the template signal T. The recursive formula is as follows:

[0099]

[0100] Start from the lower right corner of the matrix D[m][n] and backtrack to find the optimal matching path. According to the recursive formula, select the path direction (match, insert or delete) corresponding to the minimum value.

[0101] Finally, the optimal matching total similarity D[i][j] between the CSI signal and the CSI signal template is obtained in real time by calculation.

[0102] If the value is lower than the threshold, the real-time CSI signal is compared with the CSI signal templates of each area with human activity for the best matching total similarity. The area with the highest best matching total similarity among all area templates is selected as the area with human activity. The specific formula for calculating the best matching total similarity is the same as described above.

[0103] S5: Combine the dynamic matching algorithm results with the access control status or equipment operating parameters (such as current or voltage changes) for linkage and composite verification. The power station intrusion detection system automatically obtains the access control status signal and equipment operating parameters. When a human activity signal is detected, if the access control status is open, it is judged as normal human activity. If the access control status is closed, it is judged as human intrusion activity. In addition, whether there is abnormal fluctuation in the equipment operating parameters is combined to indicate whether the human intrusion activity has an impact on the electrical equipment.

[0104] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0105] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0106] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0107] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0108] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other manner. Any person skilled in the art may utilize the above-disclosed technical content to modify or modify the present invention into equivalent embodiments. However, any simple modifications, equivalent variations, and modifications to the above embodiments that do not depart from the technical content of the present invention and are based on the technical essence of the present invention remain within the scope of protection of the present invention.

Claims

1. A method for detecting personnel intrusion into a power station based on wireless channel status information, characterized in that: The following steps are involved: S1: Deploy WiFi-CSI sensors in the power substation and place them in the same local area network as the existing WiFi devices in the power substation, enabling communication and data transmission; S2: The WiFi-CSI sensors in the power station use MAC codes to identify each other to achieve real-time mutual sniffing of WiFi signals, and then parse and process the CSI data packets. The adaptive denoising algorithm is used to pre-process the collected CSI data packets. S3: Based on the phase-amplitude correction model, the phase and amplitude of the preprocessed CSI data affected by multipath signals are corrected to eliminate signal fluctuations caused by non-target disturbances and extract CSI features related to human activities. S4: Divide the distribution station into multiple detection areas based on the specific functional areas of the distribution station. Use a dynamic matching algorithm to identify whether human activity occurs and locate the areas where human activity occurs based on CSI features related to human activity. S5: Perform linkage compound verification based on the dynamic matching algorithm results combined with the access control status or equipment operating parameters.

2. The method for detecting personnel intrusion into a power substation based on wireless channel state information according to claim 1, characterized in that: The CSI data includes the amplitude and phase of 62 subcarriers with time stamps, 11 of which are null subcarriers. Before subsequent data processing, the null subcarriers in the CSI data are removed, that is, 51 valid subcarriers.

3. The method for detecting personnel intrusion into a power substation based on wireless channel state information according to claim 1, characterized in that: The preprocessing is as follows: The CSI signal is divided into multiple time regions in chronological order according to the time interval Δt, and the average value u of the CSI amplitude in each time region is calculated. amp (t region ) and standard deviation α amp (t region )calculate: Where N is the number of CSI signal samples in the time zone; CSI amp,i (t region ) is the amplitude of the i-th sample; According to the average value u of the amplitude in this time region amp (t region ) and standard deviation α amp (t region ), construct the normalized basis of the amplitude, and for each CSI signal amplitude sample in the time region, use the following formula for normalization: Symlet wavelet and 3-layer decomposition are used to perform wavelet decomposition on the normalized CSI amplitude data. Wavelet decomposition decomposes the original CSI signal into sub-band signals in different frequency ranges. Dynamically adjust the threshold according to the local characteristics of the CSI signal, perform threshold processing on the sub-band signals in different frequency ranges, and perform wavelet reconstruction on the sub-band signals in different frequency ranges after threshold processing; Based on the frequency distribution characteristics of the signal and the difference in frequency range between electrical equipment signals and human activity signals in actual scenarios, a reasonable frequency threshold is set as the basis for separation, and human activity signals and electrical equipment signals are separated through frequency analysis.

4. The method for detecting personnel intrusion into a power substation based on wireless channel state information according to claim 1, characterized in that: The phase-amplitude correction model is as follows: Calculate the phase difference of each subcarrier data of the CSI signal, the phase difference Δφ of the Nth subcarrier at time t N (t) are as follows: Df N (t)=φ N (t)-φ re (t) where φ N (t) represents the phase difference of the Nth subcarrier at time t, φ re (t) represents the phase of the reference subcarrier at time t; According to the phase difference, the phase of the subcarrier is corrected to obtain the phase φ of the Nth subcarrier after phase correction at time t N-correct (t): f N-correct (t)=φ N (t)-Df N (t) For each subcarrier’s CSI data, calculate the ratio of the Nth subcarrier amplitude at time t to the reference signal amplitude: Among them A N (t) represents the amplitude of the Nth subcarrier at time t, A re (t) represents the amplitude of the reference subcarrier at time t; According to the amplitude ratio, the amplitude of the Nth subcarrier at time t is corrected:

5. The method for detecting personnel intrusion into a power substation based on wireless channel state information according to claim 1, characterized in that: The dynamic matching algorithm is as follows: Collect CSI signal templates for areas without personnel activities and areas with personnel activities in advance, and obtain real-time CSI signals. The specific structures of the CSI signal template T and the real-time CSI signal S are as follows: T=[t1,t2,...,t n ] Each t j is a vector containing the amplitude and phase of 51 subcarriers; S=[s1,s2,....,s m ] Each of these s i is a vector containing the amplitude and phase of 51 subcarriers; A threshold is set to calculate the optimal matching total similarity between the real-time CSI signal and the CSI signal template indicating no human activity in the area. If the optimal matching total similarity is higher than the threshold, it is considered that there is no human activity. If the optimal matching total similarity is lower than the threshold, it is considered that there may be human activity. The real-time CSI signal is compared with the CSI signal templates of each area with human activity for the best matching total similarity, and the area with the highest best matching total similarity among all area templates is selected as the area of ​​human activity.

6. The method for detecting personnel intrusion into a power substation based on wireless channel state information according to claim 5, characterized in that: The steps for calculating the total similarity of the CSI optimal match are as follows: For each subcarrier k (k = 1, 2, ..., 51), the Euclidean distance of its amplitude and phase is expressed as: in, and is the amplitude and phase of the kth subcarrier of the ith signal in the real-time CSI signal S, and is the amplitude and phase of the kth subcarrier of the jth signal in the template CSI signal T; In order to obtain the total similarity between two signal points, the similarities of all subcarriers are summed and averaged. The total similarity calculation formula is as follows: Then, the dynamic matching algorithm DTW is used to obtain the optimal matching total similarity. An m×n similarity matrix D is defined, where D[i][j] represents the optimal matching total similarity between the first i signals of the real-time CSI signal S and the first j signals of the template signal T. The recursive formula is as follows: Start from the lower right corner of the matrix D[m][n] and backtrack to find the optimal matching path. According to the recursive formula, select the path direction corresponding to the minimum value. Finally, the optimal matching total similarity D[i][j] between the CSI signal and the CSI signal template is obtained in real time by calculation.

7. The method for detecting personnel intrusion into a power substation based on wireless channel state information according to claim 1, characterized in that: The linkage composite verification is specifically as follows: the power station intrusion detection system automatically obtains the access control status signal and equipment operating parameters. When a personnel activity signal is detected, if the access control status is open, it is judged as normal personnel activity; if the access control status is closed, it is judged as personnel intrusion activity. In addition, whether there is abnormal fluctuation in the equipment operating parameters is combined to indicate whether the personnel intrusion activity has an impact on the electrical equipment.

8. A power station personnel intrusion detection system based on wireless channel status information, characterized in that: It includes a processor, a memory and a computer program stored in the memory. When the processor executes the computer program, it specifically performs the steps in the power station personnel intrusion detection method based on wireless channel status information as described in any one of claims 1 to 7.

Images