A method for estimating intrusion detection boundary based on received signal strength

Abstract

The present invention discloses an intrusion detection boundary estimation method based on received signal strength, which belongs to the field of intrusion detection technology; it includes: defining the intrusion detection boundary; collecting the corresponding wireless signal data when a person intrudes in an indoor environment, and extracting the RSS raw data therefrom; analyzing the RSS data, establishing a motion disturbance RSS model, and deriving the blocking area and the reflection area in the motion disturbance RSS model; deriving the RSS intrusion detection boundary model according to the motion disturbance RSS model, and performing intrusion detection boundary detection using the RSS intrusion detection boundary model. The present invention derives the intrusion detection boundary model by analyzing the impact of human intrusion on the wireless channel; the intrusion detection boundary model accurately describes the intrinsic relationship between human intrusion and channel change; at the same time, it can measure and even model the boundary position of the intruder, make preparations in advance to eliminate hidden dangers, and improve the accuracy and security of intrusion detection boundary estimation.

Description

Technical Field

[0001] The invention belongs to the technical field of intrusion detection, and in particular relates to an intrusion detection boundary estimation method based on received signal strength. Background Art

[0002] In recent years, with the rapid development of wireless technology, intrusion detection has attracted more and more attention and has great potential in many applications such as border protection, smart home, and elderly healthcare. Various intrusion detection techniques have been proposed, such as video-based, infrared-based, sound-based, and sensor-based methods. Through real-time images, video-based methods may cause serious privacy issues for home users. Although passive infrared methods achieve better privacy protection, they have limited detection angle / range. Although sound-based methods can support NLoS scenarios, the detection accuracy is easily affected by environmental noise. Sensor-based methods attach sensors to the body, which may be inconvenient in some cases. Therefore, there is a strong desire to adopt ubiquitous, low-cost, and privacy-first passive intrusion detection methods.

[0003] WiFi-based intrusion detection methods can meet the above requirements. So far, many received WiFi signal measurements, such as received signal strength (RSS) and channel state information (CSI), have been used to passively detect moving targets by identifying any changes. Compared with RSS, CSI contains the amplitude and phase of each subcarrier, is more sensitive to the path changes of WiFi signals, and achieves a higher human motion detection rate. Most of the existing intrusion detection methods are based on CSI, but CSI data acquisition requires specific equipment, has poor universality, and is costly. While RSS may have lower accuracy and sensitivity in some aspects, in many application scenarios, especially when real-time requirements are high and hardware resources are limited, choosing RSS is a more practical and feasible choice.

[0004] Although the current RSS-based intrusion detection systems can obtain accurate results, we find that these works lack in-depth analysis and evaluation of the intrusion detection boundary. Most existing works ignore the problem of evaluating the maximum distance at which the receiver can detect human intrusion under the condition that the intrusion is detectable.

[0005] In summary, how to solve the problem of estimating the maximum distance of human intrusion detection by the intrusion detection boundary model, thereby improving the accuracy of the intrusion detection boundary model, is a technical problem that the present invention aims to solve. Summary of the invention

[0006] The purpose of the present invention is to provide an intrusion detection boundary estimation method based on received signal strength to solve the problems raised in the above background technology.

[0007] The object of the present invention is achieved by: a method for estimating the intrusion detection boundary based on received signal strength, characterized in that the method comprises the following steps:

[0008] Step S1: define the intrusion detection boundary;

[0009] Step S2: Collecting wireless signal data corresponding to human intrusion in the indoor environment, and extracting RSS raw data therefrom;

[0010] Step S3: Analyze the collected RSS data, establish a motion disturbance RSS model, and derive the blocking area and reflection area in the motion disturbance RSS model;

[0011] Step S4: derive an RSS intrusion detection boundary model based on the motion disturbance RSS model, and use the RSS intrusion detection boundary model to estimate the intrusion detection boundary.

[0012] Preferably, in step S1, the intrusion detection boundary is defined, and the specific operations are:

[0013] Assuming that the LoS distance of the transceiver nodes is constant and the indoor environment remains unchanged, the farthest position of human motion detected by the RSS-based human intrusion detection method is the perception boundary;

[0014] When the RSS feature extracted by the human motion detection method is greater than the threshold, the farthest position where the human motion is detected is recorded as the intrusion detection boundary. The intrusion detection boundary IDB is expressed as:

[0015] (1);

[0016] in, is the extracted RSS feature, is the threshold value, is the distance between the intruder and the transceiver.

[0017] Preferably, in step S3, the collected RSS data is analyzed, a motion disturbance RSS model is established, and the blocking area and the reflection area in the motion disturbance RSS model are derived. The specific operations are:

[0018] Step S3-1: Establishing a motion disturbance RSS model;

[0019] In an indoor environment, the signal reaching the receiver is the superposition of signals from multiple paths. It is known that RSS depends on the amplitude and phase of the multipath signal at the receiver; in terms of frequency and time The channel frequency response of the signal at is expressed as follows:

[0020] (2);

[0021] in, is the number of all paths, is the carrier frequency, For time, is an imaginary number, is the wavelength of the signal, Indicates The complex number of the amplitude attenuation and initial phase of each path, For the The distance of the path, To measure noise;

[0022] Transmission power Represented, then RSS is represented as:

[0023] (3);

[0024] Combining formula (2) and (3), we get:

[0025] (4);

[0026] in, represents the shortest path. If there is a LoS path, the shortest path is the LoS path; otherwise, the shortest path is the shortest non-line-of-sight path. and Related, expressed as:

[0027] (5);

[0028] in, is the product of the antenna gains at the transmitter and receiver, is the path loss factor, For the The initial phase of each path; is an imaginary number;

[0029] Assuming there is only one intruder in the indoor environment, when the intruder appears around the transceiver, Usually the superposition of static and dynamic components, in addition Also affected by noise, Written as:

[0030] (6);

[0031] in, is the impulse noise in amplitude, is the random phase offset; It is a static component, which remains unchanged in a short time. is the dynamic component; in the case of invasion, Corresponding to static signals, including LoS path signals and signals reflected by static objects such as walls and floors; It is a dynamic signal, which is reflected by the moving objects of the intruder and changes with human movement;

[0032] When there is no one in the indoor environment, the static signal reaching the receiver is the superposition of signals from all static paths; in an environment with rich multipaths, if LoS propagation exists, the shortest path is directly from the transmitter to the receiver; other paths include reflection paths, on which the transmitted signal is reflected one or more times on various reflectors; each reflection will produce a corresponding reflection loss, and the value of the reflection loss is related to the material of the reflecting surface, It is expressed as follows:

[0033] (7);

[0034] in, is the set of all static paths, Indicates The complex number of the amplitude decay and initial phase of the path, Indicates The distance of the path; Indicates The number of reflections along a path, Indicates The first path The reflection area of ​​the secondary reflection; Indicates The first The reflection loss coefficient per unit reflection area of ​​the secondary reflection, is the measurement noise in a static environment;

[0035] Multiple reflections greatly weaken the signal reaching the receiver. When only one reflection path is considered, formula (7) is abbreviated as:

[0036] (8);

[0037] in, For the The reflection area of ​​a path reflected once, For the The reflection loss coefficient per unit reflection area for a single reflection on a path;

[0038] When someone appears in an indoor environment, part of the static path is blocked by the human body, and the RSS will be disturbed; Corresponding to the signal reflected and blocked by the human body, it is written as:

[0039] (9);

[0040] in, is the area reflected by the human body, The area blocked by the human body, is the unit area reflection coefficient of the human body, is the reflection coefficient per unit area blocked by the human body; and are complex numbers, representing the amplitude attenuation and initial phase of the path reflected by the human body and the blocked path respectively; and They represent the distances of the human body reflection path and the occlusion path respectively; For measuring noise in environments with people moving;

[0041] Step S3-2: Based on the model derived in step S3-1, derive the blocking area in the model;

[0042] Step S3-3: Based on the model derived in step S3-1, derive the reflection area in the model.

[0043] Preferably, the static path signal in the indoor environment in step S3-1 remains unchanged in a short time. Combining formulas (3), (6) and (9), RSS is written as:

[0044] (10);

[0045] in, Indicates the area blocked by the human body.

[0046] Preferably, in step S3-2, based on the model derived in step S3-1, the blocking area in the model is derived, and the specific operation is as follows:

[0047] The area blocked by the human body varies with the direction and position of the human body, and the area blocked by the human body is approximated by calculating the projection area on the wall;

[0048] Knowing the length of the room is , the width is , the transmitters are located at a distance of and The distance between the transmitter and the receiver is , the connecting line between the transmitter and the receiver is parallel to the south wall;

[0049] Establish a two-dimensional coordinate system, regard the emitter as the origin, and specify the east direction as positive. Axis direction, south is positive Axis direction, The axis is parallel to the south wall;

[0050] The human body is modeled as a cuboid, with length, width and height represented as , and , is the connecting line between the transmitter and the lower right corner of the wall; The angle with the south wall is , the angle between the front of the human body and the south wall is , the clockwise direction of the angle is defined as the positive direction; the person is at point When viewed from above, the shape of a person is a , width is The four vertices of a rectangle are represented as , , and , , , and They represent the upper left vertex, upper right vertex, lower right vertex, and lower left vertex of the top view of the human body respectively;

[0051] Calculate the coordinates of the four vertices, calculate the slope between the emitter and these vertices, and determine the boundary of the blocked area; the range blocked by the human body is determined by considering the maximum slope and minimum slope Determine that the two slopes correspond to the left and right boundaries of the blocked range, respectively;

[0052] When a person is moving, the transmission signal blocked by the human body is only projected on the south wall, or only on the east wall or on both walls; according to the proportional characteristics of the triangle, the blocked area projected on the wall is expressed as follows:

[0053] (11);

[0054] in, (12);

[0055] (13);

[0056] , , ; , , , ;

[0057] when hour, ,at this time , the blocking area is only projected on the east wall;

[0058] when hour, ,at this time , the blocking area is only projected on the south wall.

[0059] Preferably, in step S3-3, based on the model derived in step S3-1, the reflection area in the model is derived, and the specific operation is as follows:

[0060] The human body is regarded as a rectangle, and the four vertices are marked as , , and ,point Represents the upper left vertex of the human body top view, point Represents the upper right vertex of the human body top view, point Represents the lower right vertex of the human body top view, point represents the lower left vertex of the top view of the human body, , , , They are the center points of the front, back, left, and right surfaces of the human body;

[0061] The determination of the effective reflection surface depends on whether the line between the center point of a specific surface of the human body and the transceiver passes through the human body;

[0062] definition The right side of the human body. is the line between the transmitter and the midpoint of the right surface of the human body, is the line connecting the transmitter and the midpoint of the right surface of the human body, for , A connecting line between two points, for , The connecting line between two points, for , A connecting line between two points;

[0063] if or With line segment , or If they intersect, The surface cannot be considered a valid reflective surface. If it is not, The surface is regarded as an effective reflecting surface;

[0064] Determine the intersection of two line segments by computing the vector cross product:

[0065] point and In line segment On both sides, vector and In vector The cross product results of the two sides are of opposite signs; point and In line segment On both sides, vector and In vector The two sides of the cross product have opposite signs; when these two conditions are met at the same time, the line segment and line segments intersect;

[0066] definition This represents the result of the intersection of two line segments:

[0067] (14);

[0068] use Indicates whether a specific face of the human body is considered a reflective face. For the back face of the human body:

[0069] (15);

[0070] in, is a transceiver and The set of lines between the center points of the faces, is the set of the remaining three faces; if , which represents a line segment and With line segment , and If they do not intersect, The surface is a reflective surface; if ,but The surface is not a reflective surface;

[0071] The effective reflection area of ​​the human body is expressed as:

[0072] (16).

[0073] Preferably, in step S4, the RSS intrusion detection boundary model is derived according to the motion disturbance RSS model, and the intrusion detection boundary is estimated using the RSS intrusion detection boundary model. The specific operations are:

[0074] Based on the intrusion detection boundary When the RSS feature extracted by the human motion detection method is greater than the threshold, the farthest position where the human motion is detected is recorded as the intrusion detection boundary. In order to obtain the IDB, the maximum amplitude improved FFT method is used. The specific operation is as follows:

[0075] When there is no intruder, the measured RSS value is recorded as ; When an intruder appears, the measured RSS value is recorded as ;

[0076] First, perform FFT on the RSS data measured in the static environment and the intrusion environment respectively:

[0077] (17);

[0078] (18);

[0079] in, , represents the first Output results, represents the first Output results, is the number of signal samples;

[0080] ;

[0081] Then extract the maximum amplitude except the 0 frequency component: , ;

[0082] When the intruder is away from the WiFi device, its impact on the wireless channel is reduced. Gradually decrease, when When it becomes small enough, it is impossible to sense the intruder's intrusion; at this time, the threshold is set to .

[0083] Preferably, in step S4, the RSS intrusion detection boundary model is derived according to the motion disturbance RSS model, and the RSS intrusion detection boundary model is derived by using an improved wavelet transform method based on scale projection, and the specific operation is as follows:

[0084] First, DWT is performed on the RSS data measured in the static environment and the intrusion environment, and the detail coefficients are extracted:

[0085] (19);

[0086] (20);

[0087] in, and They represent the decomposition level of DWT in static environment and intrusion environment respectively. Represents two sets of discrete orthogonal functions; after obtaining the detail coefficients, calculate the corresponding variance , As intruders move away from WiFi devices, their impact on the wireless channel decreases. Gradually decrease; when When it becomes small enough, it is impossible to detect the intruder's intrusion; the threshold is set to .

[0088] Compared with the prior art, the present invention has the following improvements and advantages:

[0089] 1. By analyzing the impact of human intrusion on wireless channels, the general model of intrusion disturbance RSS is derived. On the basis of the derivation of the general model of intrusion disturbance RSS, the intrusion detection boundary model is derived; the intrusion detection boundary model accurately describes the intrinsic relationship between human intrusion and channel changes; at the same time, the boundary position of the intruder can be measured and even modeled, the advance degree of intrusion detection can be known, preparations can be made in advance to eliminate hidden dangers, and the accuracy and security of intrusion detection boundary estimation can be improved.

[0090] 2. Collect the corresponding wireless signal data when people intrude through WiFi, extract RSS raw data, reflect the multipath propagation effect of wireless signals, and perform fine-grained perception of the environment; realize user contactless behavior perception with low cost and privacy protection; at the same time, it is very important to help deploy WiFi devices and achieve wide detection coverage. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 Flow chart of the method of the present invention.

[0092] Figure 2 This is a structural diagram of the signal multipath propagation scenario of the present invention.

[0093] Figure 3 This is a signal propagation diagram for an unmanned situation in an example of the present invention.

[0094] Figure 4 This is a signal propagation diagram for the presence of people in an example of the present invention.

[0095] Figure 5 This is a top view of the signal transmitted by the present invention being blocked by the human body and projected onto two walls.

[0096] Figure 6 This is a top view in which the signal transmitted by the present invention is blocked by the human body and is only projected on the south wall.

[0097] Figure 7 This is a top view showing that the signal transmitted by the present invention is blocked by the human body and is only projected onto the east wall.

[0098] Figure 8 It is a side view of the signal transmitted by the present invention being blocked by a human body and projected onto the south wall.

[0099] Fig. 9 It is a side view of the signal transmitted by the present invention being blocked by a human body and projected onto the east wall.

[0100] Fig.10 It is a reflection surface diagram of the human body in different positions and orientations in the invention.

[0101] Fig.11 It is a top view of the connection between the center points of the four sides of the human body and the transmitter and receiver in the present invention.

[0102] Fig.12 It is a top view of the connection between the center point of the right surface of the human body and the transmitter in the present invention.

[0103] Fig.13 It is a diagram showing the effect of the improved FFT method based on maximum amplitude of the present invention.

[0104] Fig.14 It is the effect diagram of the improved wavelet transform method based on scale projection of the present invention. DETAILED DESCRIPTION

[0105] The present invention is further summarized below with reference to the accompanying drawings.

[0106] like Figure 1 As shown, a method for estimating an intrusion detection boundary based on received signal strength comprises the following steps:

[0107] Step S1: define the intrusion detection boundary;

[0108] In the WiFi-based sensing system, the sensing range is mainly affected by signal attenuation and propagation characteristics; assuming that the LoS distance of the transceiver node is constant and the indoor environment remains unchanged, the farthest position of human movement detected by the RSS-based human intrusion detection method is the sensing boundary;

[0109] When the RSS feature extracted by the human motion detection method is greater than the threshold, the farthest position where the human motion is detected is recorded as the intrusion detection boundary. The intrusion detection boundary IDB is expressed as:

[0110] (1);

[0111] in, is the extracted RSS feature, is the threshold value, is the distance between the intruder and the transceiver.

[0112] Step S2: Collecting wireless signal data corresponding to human intrusion in the indoor environment, and extracting RSS raw data therefrom;

[0113] Use one transmitting antenna and one receiving antenna to send and receive RSS data, specifically:

[0114] Commercial WiFi devices equipped with Intel 5300 NICs were used to collect RSS data. Each device had an external antenna. The frequency was set to 5.24 GHz, the bandwidth was set to 20 MHz, and the transmitter sent 200 packets per second at a transmit power of 15 dBm. The two devices used Monitor mode for wireless communication.

[0115] Step S3: Analyze the collected RSS data, establish a motion disturbance RSS model, and derive the blocking area and reflection area in the motion disturbance RSS model. The specific operations are as follows:

[0116] Step S3-1: Establishing a motion disturbance RSS model;

[0117] like Figure 2 As shown, in an indoor environment, the signal reaching the receiver is the superposition of signals from multiple paths. It is known that RSS depends on the amplitude and phase of the multipath signal at the receiver; in frequency and time The channel frequency response of the signal at is expressed as follows:

[0118] (2);

[0119] in, is the number of all paths, is the carrier frequency, For time, is an imaginary number, is the wavelength of the signal, Indicates The complex number of the amplitude attenuation and initial phase of each path, For the The distance of the path, To measure noise;

[0120] Transmission power Represented, then RSS is represented as:

[0121] (3);

[0122] Combining formula (2) and (3), we get:

[0123] (4);

[0124] in, represents the shortest path. If there is a LoS path, the shortest path is the LoS path; otherwise, the shortest path is the shortest non-line-of-sight path. and Related, expressed as:

[0125] (5);

[0126] in, is the product of the antenna gains at the transmitter and receiver, is the path loss factor, For the The initial phase of each path; is an imaginary number;

[0127] Assuming there is only one intruder in the indoor environment, when the intruder appears around the transceiver, Usually the superposition of static and dynamic components, in addition Also affected by noise, Written as:

[0128] (6);

[0129] in, is the impulse noise in amplitude, is the random phase offset; It is a static component, which remains unchanged in a short time. is the dynamic component; in the case of invasion, Corresponding to static signals, including LoS path signals and signals reflected by static objects such as walls and floors; It is a dynamic signal, which is reflected by the moving objects of the intruder and changes with human movement;

[0130] like Figure 3 As shown in the figure, when there is no one in the indoor environment, the static signal reaching the receiver is the superposition of signals from all static paths; in an environment with rich multipaths, if LoS propagation exists, the shortest path is directly from the transmitter to the receiver; other paths include reflection paths, on which the transmitted signal is reflected one or more times on various reflectors; each reflection will produce a corresponding reflection loss, and the value of the reflection loss is related to the material of the reflecting surface. It is expressed as follows:

[0131] (7);

[0132] in, is the set of all static paths, Indicates The complex number of the amplitude decay and initial phase of the path, Indicates The distance of the path; Indicates The number of reflections along a path, Indicates The first path The reflection area of ​​the secondary reflection; Indicates The first The reflection loss coefficient per unit reflection area of ​​the secondary reflection, is the measurement noise in a static environment;

[0133] Multiple reflections greatly weaken the signal reaching the receiver. When only one reflection path is considered, formula (7) is abbreviated as:

[0134] (8);

[0135] in, For the The reflection area of ​​a path reflected once, For the The reflection loss coefficient per unit reflection area for a single reflection on a path;

[0136] like Figure 4 As shown, when someone appears in the indoor environment, part of the static path is blocked by the human body and the RSS will be disturbed; Corresponding to the signal reflected and blocked by the human body, it is written as:

[0137] (9);

[0138] in, is the area reflected by the human body, is the area blocked by the human body, is the unit area reflection coefficient of the human body, is the reflection coefficient per unit area blocked by the human body; and are complex numbers, representing the amplitude attenuation and initial phase of the path reflected by the human body and the blocked path respectively; and They represent the distances of the human body reflection path and the occlusion path respectively; For measuring noise in environments with people moving;

[0139] The static path signal in the indoor environment remains unchanged in a short time. Combining formulas (3), (6) and (9), RSS is written as:

[0140] (10);

[0141] in, Indicates the area blocked by the human body.

[0142] Step S3-2: Based on the model derived in step S3-1, derive the blocking area in the model. The specific operations are as follows:

[0143] The area blocked by the human body varies with the direction and position of the human body, and the area blocked by the human body is approximated by calculating the projection area on the wall;

[0144] Knowing the length of the room is , the width is , the transmitters are located at a distance of and The distance between the transmitter and the receiver is , the connecting line between the transmitter and the receiver is parallel to the south wall;

[0145] Establish a two-dimensional coordinate system, regard the emitter as the origin, and specify the east direction as positive. Axis direction, south is positive Axis direction, The axis is parallel to the south wall;

[0146] The human body is modeled as a cuboid, with length, width and height represented as , and , is the connection line between the transmitter and the lower right corner of the wall; Figure 5 , 6 , 7, 8, 9, the line segment The angle with the south wall is , the angle between the front of the human body and the south wall is , the clockwise direction of the angle is defined as the positive direction; the person is at point When viewed from above, the shape of a person is a , width is The four vertices of a rectangle are represented as , , and , , , and They represent the upper left vertex, upper right vertex, lower right vertex, and lower left vertex of the top view of the human body respectively;

[0147] Calculate the coordinates of the four vertices, calculate the slope between the emitter and these vertices, and determine the boundary of the blocked area; the range blocked by the human body is determined by considering the maximum slope and minimum slope Determine that the two slopes correspond to the left and right boundaries of the blocked range, respectively;

[0148] like Figure 5 , 6 As shown in , 7, when a person is moving, the transmission signal blocked by the human body is projected only on the south wall, or only on the east wall or on both walls; the projection heights of the south wall and the east wall are as follows Figure 8 , 9 As shown; according to the triangle ratio characteristics, the occlusion area projected on the wall is expressed as follows:

[0149] (11);

[0150] in, (12);

[0151] (13);

[0152] , , ; , , , ;

[0153] when hour, ,at this time , the blocking area is only projected on the east wall;

[0154] when hour, ,at this time , the blocking area is only projected on the south wall, such as Figure 7 shown.

[0155] Step S3-3: Based on the model derived in step S3-1, derive the reflection area in the model. The specific operation is as follows:

[0156] like Fig.10 As shown in (a) and (b), when a person is in the same orientation but in different positions, the reflection area changes. Similarly, when a person is in the same position but in different orientations, the reflection area also changes, such as Fig.10 As shown in (b) and (c), the direction and position of the human body are constantly changing during the movement.

[0157] like Fig.11 As shown, Fig.11A top view showing the center points of the four surfaces of the human body and the lines between the transmitter and receiver;

[0158] The human body is regarded as a rectangle, and the four vertices are marked as , , and ,point Represents the upper left vertex of the human body top view, point Represents the upper right vertex of the human body top view, point Represents the lower right vertex of the human body top view, point represents the lower left vertex of the top view of the human body, , , , They are the center points of the front, back, left, and right surfaces of the human body;

[0159] According to the law of reflection and refraction of electromagnetic waves, when electromagnetic waves reach the surface of the human body, they will be reflected. , and , electromagnetic waves need to penetrate the human body to reach the receiver.

[0160] Since about 70% of the human body is composed of water, electromagnetic waves will be significantly attenuated in water. Therefore, the determination of the effective reflection surface depends on whether the line between the center point of a specific surface of the human body and the transceiver passes through the human body;

[0161] definition The right side of the human body. is the line between the transmitter and the midpoint of the right surface of the human body, is the line connecting the transmitter and the midpoint of the right surface of the human body, for , A connecting line between two points, for , The connecting line between two points, for , A connecting line between two points;

[0162] if or With line segment , or If they intersect, The surface cannot be considered a valid reflective surface. If it is not, The surface is regarded as an effective reflecting surface;

[0163] Determine the intersection of two line segments by computing the vector cross product:

[0164] By line segment and line segments For example, Fig.12 As shown, point and In line segment On both sides, vector and In vector The cross product results of the two sides are of opposite signs; point and In line segment On both sides, vector and In vector The two sides of the cross product have opposite signs; when these two conditions are met at the same time, the line segment and line segments intersect;

[0165] definition This represents the result of the intersection of two line segments:

[0166] (14);

[0167] use Indicates whether a specific face of the human body is considered a reflective face. For the back face of the human body:

[0168] (15);

[0169] in, is a transceiver and The set of lines between the center points of the faces, is the set of the remaining three faces; if , which represents a line segment and With line segment , and If they do not intersect, The surface is a reflective surface; if ,but The surface is not a reflective surface;

[0170] The effective reflection area of ​​the human body is expressed as:

[0171] (16).

[0172] Step S4: derive an RSS intrusion detection boundary model based on the motion disturbance RSS model, and use the RSS intrusion detection boundary model to perform intrusion detection boundary detection. The specific operations are as follows:

[0173] Based on the intrusion detection boundary When the RSS feature extracted by the human motion detection method is greater than the threshold, the farthest position where the human motion is detected is recorded as the intrusion detection boundary. In order to obtain the IDB, the maximum amplitude improved FFT method is used. The specific operation is as follows:

[0174] When there is no intruder, the measured RSS value is recorded as ; When an intruder appears, the measured RSS value is recorded as ;

[0175] First, perform FFT on the RSS data measured in the static environment and the intrusion environment respectively:

[0176] (17);

[0177] (18);

[0178] in, , represents the first Output results, represents the first Output results, is the number of signal samples;

[0179] ;

[0180] Then extract the maximum amplitude except the 0 frequency component: , ;

[0181] When the intruder is away from the WiFi device, its impact on the wireless channel is reduced. Gradually decrease, when When it becomes small enough, it is impossible to sense the intruder's intrusion; at this time, the threshold is set to .

[0182] The improved wavelet transform method based on scale projection is used to derive the RSS intrusion detection boundary model. The specific operations are as follows:

[0183] First, DWT is performed on the RSS data measured in the static environment and the intrusion environment, and the detail coefficients are extracted:

[0184] (19);

[0185] (20);

[0186] in, and They represent the decomposition level of DWT in static environment and intrusion environment respectively. Represents two sets of discrete orthogonal functions; after obtaining the detail coefficients, calculate the corresponding variance , As intruders move away from WiFi devices, their impact on the wireless channel decreases. Gradually decrease; when When it becomes small enough, it is impossible to detect the intruder's intrusion; the threshold is set to .

[0187] The receiver and transmitter are fixed in position. The intruder gradually moves away from the WiFi device. The intruder moves at different distances from the WiFi link. During this period, the receiver collects RSS data; obtain , and constantly check whether the constraints are met. When the test result does not meet the constraints, the corresponding It is the intrusion detection boundary.

[0188] In order to prove the feasibility and effectiveness of the invention, we conducted real experiments in corridors and conference rooms:

[0189] The corridor is 3 meters wide and 15 meters long, and the conference room is 6 meters wide and 8 meters long. We reduced the distance between the volunteers and the transceiver from 10 meters (corridor) or 8 meters (conference room) to 1 meter. At each 1-meter position, the volunteers walked back and forth for 8 seconds. In this experiment, we used commercial WiFi devices equipped with Intel 5300 NICs to collect RSS data. Each device has an external antenna. The frequency is set to 5.24 GHz, the bandwidth is set to 20 MHz, and the transmitter sends 200 packets per second at a transmit power of 15dBm. To evaluate the intrusion detection range of the system, we placed a series of tags on the ground to mark the distance from the WiFi link.

[0190] like Fig.13 As shown, Fig.13 The detection performance of the improved FFT method based on maximum amplitude is shown; we can observe that the FFT amplitude gradually decreases as the distance increases. When the distance extends to 5 meters, the FFT amplitude is lower than the threshold RSS th1, and the intrusion detection boundary is established at 4 meters.

[0191] like Fig.14 As shown, Fig.14 The detection performance of the improved wavelet transform method based on scale projection is shown. We can observe that the variance of the detail coefficients gradually decreases with the increase of distance. When the distance extends to 6 meters, the variance of the detail coefficients is below the threshold, and the intrusion detection boundary is established at 5 meters.

[0192] The above description is only an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A signal strength-based intrusion detection boundary estimation method, characterized in that: The method comprises the following steps: Step S1: define the intrusion detection boundary; Define the intrusion detection boundary. The specific operations are as follows: Assuming that the LoS distance of the transceiver nodes is constant and the indoor environment remains unchanged, the farthest position of human motion detected by the RSS-based human intrusion detection method is the perception boundary; When the RSS feature extracted by the human motion detection method is greater than the threshold, the farthest position where the human motion is detected is recorded as the intrusion detection boundary. The intrusion detection boundary IDB is expressed as: IDB=max(d TP |RSS ft >RSS th ) (1); Among them, RSS ft RSS is the extracted RSS feature. th is the threshold, d TP is the distance between the intruder and the transceiver; Step S2: Collecting wireless signal data corresponding to human intrusion in the indoor environment, and extracting RSS raw data therefrom; Step S3: Analyze the collected RSS data, establish a motion disturbance RSS model, and derive the blocking area and reflection area in the motion disturbance RSS model; Analyze the collected RSS data, establish the motion disturbance RSS model, and derive the blocking area and reflection area in the motion disturbance RSS model. The specific operations are as follows: Step S3-1: Establishing a motion disturbance RSS model; In an indoor environment, the signal reaching the receiver is the superposition of signals from multiple paths. It is known that RSS depends on the amplitude and phase of the multipath signal at the receiver; the channel frequency response of the signal at frequency f and time t is expressed as follows: Where L is the number of all paths, f is the carrier frequency, t is time, j is an imaginary number, λ is the wavelength of the signal, and a l (f, t) represents the complex number of the amplitude attenuation and initial phase of the lth path, d l (t) is the distance of the lth path, ε(f,t) is the measurement noise; Transmission power is P t Represented, then RSS is represented as: RSS(f,t)=|H(f,t)| 2 ×P t (3); Combining formula (2) and (3), we get: Where L = 1 indicates the shortest path. If there is a LoS path, the shortest path is the LoS path; otherwise, the shortest path is the shortest non-line-of-sight path. l (f,t) and d l (t) is related to, expressed as: Among them, G l is the product of the antenna gains at the transmitter and receiver, n is the path loss factor, is the initial phase of the lth path; j is an imaginary number; Assuming there is only one intruder in the indoor environment, when the intruder appears around the transceiver, H(f, t) is usually the superposition of static and dynamic components. In addition, H(f, t) is also affected by noise. H(f, t) is written as: Among them, A niose is the impulse noise in amplitude, is the random phase offset; H s (f) is the static component, which remains unchanged in a short time. d (f, t) is the dynamic component; in the case of invasion, H s (f) corresponds to static signals, including LoS path signals and signals reflected by static objects such as walls and floors; H d (f, t) is a dynamic signal, which is reflected by the moving object of the intruder and changes with human movement; When there is no one in the indoor environment, the static signal reaching the receiver is the superposition of signals from all static paths; in an environment with rich multipaths, if LoS propagation exists, the shortest path is directly from the transmitter to the receiver; other paths include reflection paths, on which the transmitted signal is reflected one or more times on various reflectors; each reflection will produce a corresponding reflection loss, and the value of the reflection loss is related to the material of the reflecting surface, H s (f) is expressed as follows: Among them, P s is the set of all static paths, a k (f) represents the complex number of the amplitude attenuation and initial phase of the kth path, d k represents the distance of the kth path; N K represents the number of reflections of the kth path, represents the reflection area of ​​the nth reflection of the kth path; Represents the reflection loss coefficient per unit reflection area of ​​the nth reflection of the kth path, ε s (f) is the measurement noise in a static environment; Multiple reflections greatly weaken the signal reaching the receiver. When only one reflection path is considered, formula (7) is shortened to: Among them, S k is the reflection area of ​​the kth path once reflected, Γ k is the reflection loss coefficient per unit reflection area of ​​the first reflection of the kth path; When someone appears in the indoor environment, part of the static path is blocked by the human body, and the RSS will be disturbed; d (f, t) corresponds to the signal reflected and blocked by the human body and is written as: Among them, S mo (t) is the area reflected by the human body, S bl (t) is the area blocked by the human body, Γ mo is the unit area reflection coefficient of the human body, Γ bl A is the unit area reflection coefficient blocked by the human body; mo (f,t) and a bl (f, t) are complex numbers, representing the amplitude attenuation and initial phase of the path reflected by the human body and the blocked path, respectively; d mo (t) and d bl (t) represents the distance of the human body reflection path and the occlusion path respectively; ε d (f, t) is the measurement noise in a human motion environment; Step S3-2: Based on the model derived in step S3-1, derive the blocking area in the model; Step S3-3: based on the model derived in step S3-1, deriving the reflection area in the model; Step S4: deriving an RSS intrusion detection boundary model according to the motion disturbance RSS model, and using the RSS intrusion detection boundary model to estimate the intrusion detection boundary; The RSS intrusion detection boundary model is derived based on the motion disturbance RSS model, and the intrusion detection boundary is estimated using the RSS intrusion detection boundary model. The specific operations are as follows: According to the intrusion detection boundary IDB=max(d TP |RSS ft >RSS th ), when the RSS feature extracted by the human motion detection method is greater than the threshold, the farthest position where the human motion is detected is recorded as the intrusion detection boundary; in order to obtain the IDB, the maximum amplitude improved FFT method is used, and the specific operation is as follows: When there is no intruder, the measured RSS value is recorded as rss0(t); when an intruder appears, the measured RSS value is recorded as rss d (t); First, perform FFT on the RSS data measured in the static environment and the intrusion environment respectively: in, FFT s(m) Represents the mth output result after the fast Fourier transform, FFT d(n) Represents the nth output result after fast Fourier transform, where N is the number of signal samples; Then extract the maximum amplitude except the 0 frequency component: M s =max(|FFT s[1:N-1] |), M d =max(|FFT d[1:N-1] |); When the intruder is away from the WiFi device, its impact on the wireless channel decreases, M d Gradually decreases, when M d When it becomes small enough, it is impossible to perceive the intruder's intrusion; at this time, the threshold RSS th Set to M s .

2. The method for estimating intrusion detection boundaries based on signal strength according to claim 1, characterized in that: In step S3-1, the static path signal in the indoor environment remains unchanged in a short time. Combining formulas (3), (6) and (9), RSS is written as: Among them, S bl (t) represents the area blocked by the human body.

3. The method for estimating intrusion detection boundaries based on signal strength according to claim 1, characterized in that: In step S3-2, based on the model derived in step S3-1, the blocking area in the model is derived. The specific operation is: The area blocked by the human body varies with the direction and position of the human body, and the area blocked by the human body is approximated by calculating the projection area on the wall; Knowing that the length of the room is M, the width is N, the transmitter is located at distances a and b from the east and south walls respectively, the distance between the transmitter and the receiver is c, and the connecting line between the transmitter and the receiver is parallel to the south wall; Establish a two-dimensional coordinate system, regard the transmitter as the origin, specify the east direction as the positive x-axis direction, the south direction as the positive y-axis direction, and the x-axis is parallel to the south wall; The human body is modeled as a cuboid, with the length, width and height represented as l, w and h respectively, and TW as the connecting line between the transmitter and the lower right corner of the wall; the angle between the line segment TW and the south wall is β, and the angle between the front of the human body and the south wall is θ, and the clockwise direction of the angle is defined as the positive direction; The person is at point P(x, y). When viewed from above, the person is a rectangle with a length of l and a width of w. The four vertices are represented by A, B, C, and D. A, B, C, and D represent the upper left vertex, upper right vertex, lower right vertex, and lower left vertex of the top view of the human body, respectively. The coordinates of the four vertices are calculated, the slopes between the emitter and these vertices are calculated, and the boundaries of the blocked area are determined; the range blocked by the human body is determined by considering the maximum slope k max and minimum slope k min Determine that the two slopes correspond to the left and right boundaries of the blocked range, respectively; When a person is moving, the transmission signal blocked by the human body is projected only on the south wall, or only on the east wall or on both walls; according to the proportional characteristics of the triangle, the blocked area projected on the wall is expressed as follows: S bl =S bl1 +S bl2 (11); in, when When S bl1 =0, then S bl =S bl2 , the blocking area is only projected on the east wall; When min(k max ,tan β )-k min When ≤0, S bl2 =0, then S bl =S bl1 , the blocking area is only projected on the south wall.

4. The method for estimating intrusion detection boundaries based on signal strength according to claim 1, characterized in that: In step S3-3, based on the model derived in step S3-1, the reflection area in the model is derived. The specific operation is as follows: The human body is regarded as a rectangle, and the four vertices are marked as E, F, G and H. Point E represents the upper left vertex of the human body in the top view, point F represents the upper right vertex of the human body in the top view, point G represents the lower right vertex of the human body in the top view, and point H represents the lower left vertex of the human body in the top view. f , S b , S l , S r They are the center points of the front, back, left, and right surfaces of the human body; The determination of the effective reflection surface depends on whether the line between the center point of a specific surface of the human body and the transceiver passes through the human body; Define FG surface as the right surface of the human body, TS r RS is the line connecting the transmitter and the midpoint of the right surface of the human body. r is the line connecting the receiver and the midpoint of the right surface of the human body, EF is the line connecting the two points E and F, EH is the line connecting the two points E and H, and GH is the line connecting the two points G and H; If TS r or RS r If it intersects with line segment EF, EH or GH, then the FG surface cannot be regarded as a valid reflection surface. If not, the FG surface is regarded as a valid reflection surface; Determine the intersection of two line segments by computing the vector cross product: Points E and F are on line segment TS r On both sides, vector and In vector The cross product results of the two sides are of opposite signs; points T and S r On both sides of line segment EF, vector and In vector The two sides of the cross product have opposite signs; when these two conditions are met at the same time, the line segment TS r Intersects line segment EF; Definition I represents the result of the intersection of two line segments: Use J to indicate whether a specific face of the human body is considered a reflective face. For the back face of the human body: Among them, Center is the set of lines between the transceiver and the center point of the FG surface, and Surface is the set of the other three surfaces; if J(FG) = 0, it means that the line segment TS r and RS r If it does not intersect with line segments EF, EH and GH, then the FG surface is the reflection surface; if J(FG)≠0, then the FG surface is not the reflection surface; The effective reflection area of ​​the human body is expressed as:

5. The method for estimating intrusion detection boundaries based on signal strength according to claim 1, characterized in that: In step S4, the RSS intrusion detection boundary model is derived from the motion disturbance RSS model, and the RSS intrusion detection boundary model is derived using an improved wavelet transform method based on scale projection. The specific operations are as follows: First, DWT is performed on the RSS data measured in the static environment and the intrusion environment, and the detail coefficients are extracted: Where l and u represent the decomposition level of DWT in static environment and intrusion environment respectively, h ′ Represents two sets of discrete orthogonal functions; after obtaining the detail coefficients, calculate the corresponding variance V s =var(DWT s ), V d =var(DWT d ); As intruders move away from WiFi devices, their impact on the wireless channel decreases, V d Gradually decrease; when V d When it becomes small enough, it is impossible to detect the intruder's intrusion; the threshold RSS th is set to V s .

Images