Intelligent security method and system based on Internet of Things

Abstract

The invention relates to the technical field of intelligent security and protection, in particular to an intelligent security and protection method and system based on the Internet of Things, and the method comprises the following steps: obtaining a bioelectrical impedance signal generated when a user touches an access control sensing area, identifying a current response condition of a differentiated frequency band, matching stored user data, analyzing the continuity of an electrical impedance parameter, and screening abnormal fluctuation data. And the fluctuation frequency and the variation amplitude in a short time are identified, and a bioelectrical impedance matching value is obtained. According to the invention, through comprehensive analysis of the electrical impedance parameters and the gait data, the individual identity can be identified and verified more accurately, the protection capability for unauthorized access is improved, through careful monitoring of physiological and behavioral characteristics, the security is significantly improved, and the stability and reliability in various environments can be adjusted and improved according to environmental changes. The capability of preventing potential security threats is enhanced for risk assessment of abnormal behaviors, and generation of security vulnerabilities is effectively prevented by monitoring behavior deviation in the identity verification process in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent security, and particularly to an intelligent security method and system based on the Internet of Things. Background Art

[0002] The technical field of intelligent security involves using computer networks, sensing devices, automatic control technologies, and data analysis to enhance the effectiveness and response speed of security systems. This field utilizes a variety of Internet-connected sensors and devices to monitor and manage security threats, including intrusion detection, access control, and video surveillance. Intelligent security systems rely on Internet of Things technologies to enable remote control and data exchange of devices for real-time monitoring and event response, and also utilize data analysis to optimize security strategies and predict potential risks.

[0003] Among them, the intelligent security method of the Internet of Things refers to the coordinated work of various sensors and actuators connected through a network to improve the overall performance and intelligence level of the security system. The patent theme covers the use of specific sensing technologies to detect environmental changes, unauthorized activities, or security-related events, and transmit information through the network to a central processing system. The central system executes predetermined response measures based on the received data, such as alarms or access control, to ensure the timeliness and accuracy of security measures. This patent theme mainly focuses on the integration of Internet of Things devices and data interaction protocols, as well as how to effectively manage and control devices to provide continuous and reliable security services.

[0004] The prior art relies on traditional sensors and basic network connections and cannot perform complex data analysis and real-time adaptive adjustment. This results in the system being unable to accurately identify or adapt to new security threats in complex or changing environments. Just relying on a basic intrusion detection system cannot effectively distinguish when faced with carefully planned deception or forged identity credentials. The data processing in the prior art is relatively simple and cannot comprehensively consider the diverse behavior patterns of users, which leads to false alarms or missed alarms in actual operation, reducing the overall security and user experience of the system. Due to the lack of the ability to deeply analyze abnormal behaviors, when there are atypical access attempts or environmental changes, the system cannot provide the necessary responses, putting the security system at risk of being breached. Summary of the Invention

[0005] To solve the technical problems existing in the prior art, namely, the inability to perform complex data analysis and real-time adaptive adjustment, which results in the system being unable to accurately identify or adapt to new security threats in complex or changing environments. Relying solely on a basic intrusion detection system cannot effectively distinguish when faced with carefully planned deceptive behaviors or forged identity credentials. The data processing in the prior art is relatively simple and cannot comprehensively consider the diverse behavior patterns of users, which leads to false positives or false negatives in actual operation, reducing the overall security of the system and the user experience. Due to the lack of the ability to deeply analyze abnormal behaviors, when there are atypical access attempts or environmental changes, the system cannot provide the necessary responses, putting the security system at risk of being breached. Embodiments of the present invention provide an intelligent security method and system based on the Internet of Things. The technical solutions are as follows:

[0006] On the one hand, an intelligent security method based on the Internet of Things is provided, and the method includes:

[0007] S1: Obtain the bioelectrical impedance signal of the user touching the access control sensing area, identify the current response situation in different frequency bands, match the stored user data, analyze the continuity of the impedance parameters, screen out abnormally fluctuating data, and identify the fluctuation frequency and change amplitude within a short time to obtain a bioelectrical impedance matching value;

[0008] S2: Invoke the bioelectrical impedance matching value, monitor the user gait data when the user approaches the access control area, calculate the length change rate of adjacent steps, perform time series segmentation on the joint angle change, calculate the step frequency offset rate and perform gait behavior feature matching to obtain a gait matching value;

[0009] S3: Calculate the weight ratio of the gait matching value and the bioelectrical impedance matching value, identify the matching feature with the priority in weight ratio, and analyze the stability of the matching deviation to obtain a multi-modal identity authentication result;

[0010] S4: Utilize the multi-modal identity authentication result to analyze the change situation of gait features within a short time, monitor the step length mutation, gait angle distortion and step frequency change situation, calculate the abnormal behavior offset rate, compare it with the normal behavior data to obtain an abnormal behavior risk assessment value, and perform security warning according to the priority degree of the risk assessment value.

[0011] As a further solution of the present invention, the bioelectrical impedance matching value includes impedance stability, conductance fluctuation amplitude, and dielectric abnormal frequency; the gait matching value includes step length change rate, joint movement period, and step frequency stability; the multi-modal identity authentication result includes feature weight ratio, matching threshold deviation, and environmental adaptability; the abnormal behavior risk assessment value includes step length abnormality rate, joint distortion degree, and step frequency mutation index.

[0012] As a further solution of the present invention, the steps for obtaining the bioelectrical impedance matching value are specifically as follows:

[0013] S101: Obtain the bioelectrical impedance signal of the user touching the access control sensing area, transmit a multi-frequency weak current to the skin, record the conductivity, permittivity and impedance distribution, identify the combined characteristics of conductivity and permittivity in multiple frequency bands according to the response conditions of different frequency bands, and obtain the characteristic value of the current response frequency band;

[0014] S102: Based on the characteristic value of the current response frequency band, call the preset user data to perform a response comparison for each frequency band, compare the combined values of conductivity and permittivity in multiple frequency bands numerically, identify the response conditions and change amplitudes of the corresponding frequency bands, and screen the data with the frequency band response difference within the comparison range to obtain the corresponding interval of impedance characteristics;

[0015] S103: According to the corresponding interval of the impedance characteristics, continuously extract the impedance change curve within the frequency band, identify the impedance fluctuation points in a short time, count the fluctuation frequency and change amplitude, and obtain the bioelectrical impedance matching value.

[0016] As a further solution of the present invention, the steps for obtaining the gait matching value are specifically as follows:

[0017] S201: Call the bioelectrical impedance matching value, monitor the user's gait when approaching the access control area, record the position point sequence and timestamp information in consecutive steps, extract the displacement data and time interval between each step, calculate the ratio of the length change of adjacent steps, and obtain the step length change rate sequence;

[0018] S202: According to the step length change rate sequence, analyze the periodic boundary points between consecutive steps, identify the time nodes at the start and end of the steps, and combine the joint angle change values recorded on the step time axis to perform segmentation in the time axis direction and divide the angle change data segments within the complete period to obtain the joint angle time series interval;

[0019] S203: Based on the joint angle time series interval, extract the step frequency value within each gait cycle, identify the change amplitude of the step frequency between adjacent cycles and the difference between the average cycle step frequency, combine the time span to judge the offset direction and offset rate, calculate the step frequency offset value, and obtain the step frequency offset recognition result;

[0020] The formula for calculating the step frequency offset value is:

[0021]

[0022] where Δf z represents the step frequency offset value of the z-th gait cycle, f z represents the step frequency value of the z-th gait cycle, f z-1represents the step frequency value of the (z - 1)-th gait cycle, T z represents the time span of the z-th gait cycle, T z-1 represents the time span of the (z - 1)-th gait cycle represents the average value of the step frequency values from the (z - 3)-th to the z-th gait cycles;

[0023] S204: According to the step frequency offset recognition result, match the step frequency pattern and the joint angle change rhythm in each cycle, screen the data segments that meet the step frequency offset rate threshold range and the angle change rhythm continuity condition, conduct gait behavior pattern comparison, and obtain the gait matching value.

[0024] As a further solution of the present invention, the step of obtaining the multi-modal identity authentication result is specifically as follows:

[0025] S301: Call the gait matching value and the bioelectrical impedance matching value, respectively extract the numerical sequences in the same time period, identify the interval distribution range, identify the response weights under the joint input condition, and obtain the matching value weight ratio;

[0026] S302: According to the matching value weight ratio, extract multiple groups of matching value change data under different environmental parameter conditions, statistically calculate the numerical trend slope of the weight-dominated matching value under different parameter conditions, judge the fluctuation interval of the change curve, and obtain the environmental adaptability matching interval;

[0027] S303: Based on the environmental adaptability matching interval, identify the matching features with a weight ratio exceeding the set threshold, calculate the deviation change rate and judge the stability of the fluctuation range, and perform corresponding verification with the identity verification record in the access control interaction response to obtain the multi-modal identity authentication result.

[0028] As a further solution of the present invention, the step of obtaining the abnormal behavior risk assessment value is specifically as follows:

[0029] S401: Call the multi-modal identity authentication result, extract the gait feature data in the corresponding time period, identify the step length, gait angle and step frequency sequence in continuous steps, monitor the abnormal step length mutation points, angle distortion points and step frequency mutation points in a short time, calculate the occurrence proportion of multiple types of mutation events per unit time, and obtain the abnormal behavior offset rate;

[0030] S402: Based on the abnormal behavior offset rate, extract the standard fluctuation intervals of the step length, angle and step frequency under normal behavior, conduct segmented judgment on the distribution of the real-time offset rate value in multiple dimensions, calculate the offset distribution intensity characteristic value, identify the offset segments exceeding the standard interval, and calculate the proportion interval of the overall offset segment in the full sample to obtain the abnormal behavior risk assessment value.

[0031] As a further solution of the present invention, the formula for calculating the offset distribution intensity eigenvalue is as follows:

[0032]

[0033] Wherein, RS represents the offset distribution intensity eigenvalue, n r represents the number of offset data in the r-th segment, O rq represents the q-th real-time offset rate observation value in the r-th segment, μ rq represents the mean value corresponding to the q-th data in the corresponding dimension, σ rq represents the standard deviation corresponding to the q-th data in the corresponding dimension, Δ rq represents the offset difference in dimension between the q-th data and the data at the previous moment, ε is a positive constant, w rq represents the critical weight of the q-th data in the r-th segment.

[0034] As a further solution of the present invention, the method further includes step S5:

[0035] S5: Compare the abnormal behavior risk assessment value with the safety threshold, screen risk behaviors for permission restriction, judge whether the multi-modal identity authentication result matching value meets the authorization requirements, screen the data whose matching value does not exceed the set range and authorize passage. If the deviation of the matching value exceeds the matching range, trigger secondary verification to obtain the access control permission instruction;

[0036] The access control permission instruction includes an authorization passage instruction, a permission restriction instruction, and a secondary verification trigger condition.

[0037] As a further solution of the present invention, the steps for obtaining the access control permission instruction are specifically as follows:

[0038] S501: Invoke the abnormal behavior risk assessment value, compare it with the set risk behavior safety threshold, judge whether the risk assessment value exceeds the defined interval, screen the data segments where the gait mutation rate and the angle offset rate exceed the defined interval, mark them as behavior segments to be restricted, and obtain the risk behavior screening result;

[0039] S502: Based on the risk behavior screening result, judge whether the matching value in the multi-modal identity authentication result is within the authorization requirement range, extract the combined value under the corresponding weights of the gait matching value and the impedance matching value, screen the data whose matching value does not exceed the set range, and obtain the initialization authorization passage data;

[0040] S503: According to the initialization authorization passage data, judge whether the deviation of the matching value is within the matching range limit. If there is a phenomenon that the combined matching value deviates beyond the set range in both the stride frequency and impedance dimensions, mark it as requiring supplementary verification, trigger secondary verification, and obtain the access control permission instruction.

[0041] On the other hand, the Internet of Things-based intelligent security system is used to execute the above-mentioned Internet of Things-based intelligent security method, and the system includes:

[0042] The impedance identification module acquires the skin contact points, contact duration data, and current response values of the user touching the access control sensing area, analyzes the current response changes in different frequency bands, calculates the change amplitude and change frequency between multi-band responses, screens the abnormal fluctuation data sections within a short time, and obtains the bioimpedance matching value;

[0043] The gait recognition module, based on the bioimpedance matching value, detects the movement amplitude of the lower limb joints, single-step length, and walking cycle interval of the user during the process of approaching the access control area, divides the angle change nodes in continuous time periods, evaluates the joint change trajectories within multiple cycle segments, and obtains the gait matching value;

[0044] The identity fusion module calls the gait matching value and the bioimpedance matching value, detects the change fluctuation trend under the air humidity value, temperature value, and contact pressure data in the real-time access control area, statistically calculates the change value of the weight ratio under different environment combinations, and determines the real-time weight dominant term, and obtains the multi-modal identity authentication result;

[0045] The behavior risk monitoring module, according to the multi-modal identity authentication result, monitors the continuous step length value, joint torsion angle interval, and step frequency change amount of the real-time contact user within a short time, screens the abnormal offset degree corresponding to the abnormal fluctuation point, and obtains the abnormal behavior risk assessment value;

[0046] The permission determination module compares the abnormal behavior risk assessment value with the set risk safety threshold, screens the risk behaviors according to whether the deviation rate exceeds the threshold, and makes an interval judgment on the real-time matching rate and the access control permission standard value under the condition that the risk condition is not triggered, and judges whether it belongs to the authorized range, and obtains the access control permission instruction.

[0047] The beneficial effects brought by the technical solution provided by the embodiment of the present invention at least include:

[0048] By acquiring the bioimpedance signal of the user touching the access control sensing area and analyzing its continuity and fluctuation, the accuracy and security of identity authentication are enhanced. Through the comprehensive analysis of impedance parameters and gait data, the individual identity can be more accurately identified and verified, and the protection ability against unauthorized access is improved. By carefully monitoring physiological and behavioral characteristics, the security is significantly improved. By dynamically adjusting the matching threshold, it can be adjusted according to environmental changes, and the stability and reliability in diverse environments are enhanced. During the recognition process, the risk assessment of abnormal behaviors further enhances the ability to prevent potential security threats, and by real-time monitoring the behavior deviation during the identity verification process, the generation of security vulnerabilities is effectively prevented. Brief Description of the Drawings

[0049] Figure 1 is a schematic diagram of the workflow of the present invention;

[0050] Figure 2 is a system flowchart of the present invention. Detailed implementation manners

[0051] The technical solutions in the present invention will be described below with reference to the accompanying drawings.

[0052] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as an "example" in the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, the use of the word "example" is intended to present concepts in a specific manner. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one of the two.

[0053] To make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0054] Please refer to Figure 1 , the embodiments of the present invention provide an intelligent security method based on the Internet of Things. The processing flow of this method may include the following steps:

[0055] S1: Obtain the bioelectrical impedance signal of the user touching the access control sensing area, transmit a multi-frequency weak current to the skin, record the conductivity, permittivity and impedance distribution, identify the current response situation in the differential frequency band, match the stored user data, analyze the continuity of the impedance parameters, screen the abnormally fluctuating data, and identify the fluctuation frequency and change amplitude within a short time to obtain the bioelectrical impedance matching value;

[0056] S2: Call the bioelectrical impedance matching value, monitor the user gait data when the user approaches the access control area, calculate the length change rate of adjacent steps according to the step length stability, analyze the gait cycle and perform time series segmentation on the joint angle change, calculate the step frequency deviation rate and perform gait behavior feature matching to obtain the gait matching value;

[0057] S3: Calculate the weight ratio of the gait matching value and the bioelectrical impedance matching value, adjust the matching threshold according to the change trend of the matching data under different environmental parameters, identify the matching feature with the priority of the weight ratio, and analyze the stability of the matching deviation, and perform verification in combination with the access control interaction situation to obtain the multi-modal identity authentication result;

[0058] S4: Analyze the changes in gait characteristics within a short period using the multi-modal identity authentication results, monitor the step length mutation, gait angle distortion, and step frequency changes, calculate the abnormal behavior deviation rate, compare it with the normal behavior data, and obtain the abnormal behavior risk assessment value;

[0059] S5: Compare the abnormal behavior risk assessment value with the safety threshold, screen for risk behaviors and impose permission restrictions, determine whether the matching value of the multi-modal identity authentication result meets the authorization requirements, screen the data whose matching value does not exceed the set range and authorize passage. If the deviation of the matching value exceeds the matching range, trigger secondary verification to obtain the access control permission instruction;

[0060] The bioelectrical impedance matching value includes impedance stability, conductance fluctuation amplitude, and dielectric anomaly frequency. The gait matching value includes step length change rate, joint movement period, and step frequency stability. The multi-modal identity authentication result includes characteristic weight ratio, matching threshold deviation, and environmental adaptability. The abnormal behavior risk assessment value includes step length abnormality rate, joint distortion degree, and step frequency mutation index. The access control permission instruction includes authorization passage instruction, permission restriction instruction, and secondary verification trigger condition.

[0061] The specific steps for obtaining the bioelectrical impedance matching value are as follows:

[0062] S101: Obtain the bioelectrical impedance signal of the user touching the access control sensing area, transmit multi-frequency weak current to the skin, record the conductivity, permittivity, and impedance distribution, and identify the combined characteristics of conductivity and permittivity in multiple frequency bands according to the response conditions of different frequency bands to obtain the current response frequency band characteristic value;

[0063] By presetting an array module including multiple skin-attached electrodes in the access control sensing area, after the user's finger touches, the system immediately transmits multi-frequency weak current to the electrode array through the controller. The transmitted frequency range can be set from low frequency 10 kHz to high frequency 1 MHz, covering the commonly used bioelectrical impedance detection frequency bands. The step size is generally selected as 10 kHz according to the set resolution. In actual operation, the system collects the voltage generated between the skin and the electrode at different frequencies, records the corresponding response voltage curve through the signal processing unit, and gradually constructs multi-dimensional parameters such as conductivity and capacitance characteristics through frequency scanning. During the signal acquisition process, to ensure the stability and reliability of the biological signal, it is also necessary to simultaneously monitor environmental parameters such as external temperature and finger humidity for dynamic error correction. Taking the daily use scenario as an example, when the user touches the metal access control device in the cold and dry winter, the system will automatically identify the contact quality through the embedded sensor and adjust the micro-current transmission duration or increase the sampling frequency to improve data stability. During the whole process, the response signal data of each frequency band will be uniformly encapsulated into a structured record to obtain the current response frequency band characteristic value.

[0064] S102: Based on the eigenvalue of the current response frequency band, call the preset user data to perform a frequency-band-by-frequency-band response comparison, numerically compare the combined values of conductivity and permittivity for multiple frequency bands, identify the response conditions and change amplitudes of the corresponding frequency bands, screen the data with the response difference within the comparison range, and obtain the corresponding interval of impedance characteristics;

[0065] The system will automatically retrieve the preset impedance data corresponding to the user's identity and compare the currently collected eigenvalue with it band by band. This process is based on a pre-established user database that contains the combined values of conductivity and permittivity of different users under different conditions and different frequency bands. The system uses a set data comparison strategy to match and compare each set of parameters in the current frequency band with the values in the corresponding frequency band in the database, and makes a response matching determination based on the difference and change trend. During the comparison process, the absolute difference and relative change percentage between the current value and the preset value will be evaluated first, and different threshold intervals will be set to divide the response status. For example, when the historical conductivity of a user at a frequency of 50 kHz is 0.0020 and the current value is 0.0021, the difference is within the allowable range, and the system determines it as a normal response. Further, if multiple consecutive frequency bands are within a similar range, the frequency band is designated as a characteristic response interval. In the actual application of the access control system, if the high-frequency band parameters in each user's entry and exit records are within a stable range, this interval will be considered by the system as the frequency band with the most stable recognition accuracy and will be preferentially used for identity matching analysis to improve the overall data credibility and consistency, and obtain the corresponding interval of impedance characteristics.

[0066] S103: According to the corresponding interval of impedance characteristics, continuously extract the impedance change curve within the frequency band, identify the impedance fluctuation points in a short time, count the fluctuation frequency and change amplitude, and obtain the bio-impedance matching value;

[0067] The system starts a high-density scanning program to carefully collect each frequency point within the interval. The frequency step can be set to 1 kHz, and the coverage interval is, for example, from 60 kHz to 70 kHz, that is, the collected frequency points are 60 kHz, 61 kHz to 70 kHz, a total of 11 points. Each frequency point is sampled 5 times to reduce the error caused by random fluctuations, and the average value of the 5 sampling results is used as the stable impedance value of that frequency. The impedance means of all frequency points are arranged in ascending order of frequency to form a set of continuous impedance curve data. Then, start to detect the impedance change trend and identify the impedance fluctuation points. The identification of impedance fluctuation points is judged based on the change rate, and its calculation formula is:

[0068]

[0069] where δZ is the impedance change rate, Z i is the average impedance value of the i-th frequency point, Z i+1is the average impedance value at the (i + 1)-th frequency point, f i is the frequency value at the i-th frequency point, f i+1 is the frequency value at the (i + 1)-th frequency point;

[0070] Set the frequency point f i = 63 kHz, f i+1 = 64 kHz, and their corresponding impedance values are Z i = 4800 Ω, Z i+1 = 5100 Ω, then:

[0071]

[0072] Set the impedance fluctuation threshold in the system to 200 Ω / kHz. If the change rate exceeds this threshold, the current frequency point is regarded as a fluctuation point. After traversing the entire frequency band, record all frequency points that meet this condition. For example, if a total of 3 fluctuation points are found and there are 11 points in the frequency band, then the fluctuation frequency is:

[0073]

[0074] At the same time, count the maximum and minimum impedance values within the fluctuation interval. For example, if the highest value is 5300 Ω and the lowest value is 4600 Ω, then the fluctuation amplitude is:

[0075] 5300 - 4600 = 700;

[0076] The system performs a fusion calculation on the fluctuation frequency and the fluctuation amplitude to obtain the matching value of the bioelectrical impedance, which is often expressed by the weighted average method:

[0077] MA = w1·F + w2·A;

[0078] where MA is the impedance matching value, F is the fluctuation frequency,

[0079] A: the fluctuation amplitude, w1 and w2 are weight coefficients, with a value range of 0 to 1, and w1 + w2 = 1;

[0080] For example, if w1 = 0.6 and w2 = 0.4, then:

[0081] MA = 0.6·0.27 + 0.4·700 = 0.162 + 280 = 280.162;

[0082] This matching value is used to quantify the comprehensive situation of impedance fluctuations within this frequency band, providing a reference for subsequent identity recognition and signal classification

[0083] The specific steps for obtaining the gait matching value are as follows:

[0084] S201: Invoke the bioelectrical impedance matching value, monitor the user's gait when approaching the access control area, record the sequence of position points and timestamp information in consecutive steps, extract the displacement data and time interval between each step, calculate the ratio of the length change between adjacent steps, and obtain the step length change rate sequence;

[0085] The system enters the user behavior perception mode. At this time, the access control system will continuously monitor the dynamic behavior of users within a certain range in front of the door, obtain the user's movement trajectory through non-contact sensing technologies such as ground pressure sensors, infrared arrays, or lidar. The system will detect the changes in plantar force or body movement coordinates during the process of the user approaching the access control area, and form a continuous sequence of position points in combination with timestamp information. When the user slowly approaches from more than 2 meters away from the access control, the system records the two-dimensional spatial coordinates corresponding to each time the user's foot lands and marks the landing time point. By calculating the spatial distance between adjacent position changes, the step length data for each step is obtained; combined with the timestamp interval, the actual time used between steps is extracted to form "step time" data pairs. Further, the system calculates the step length change ratio between every two steps. For example, if the previous step is 0.65 meters and the next step is 0.70 meters, the change ratio is the ratio of the two, and a complete step length change rate sequence is constructed. During the whole process, the system stores and continuously records multiple sets of step information in real time for analyzing the user's step stability and individual movement characteristics and obtaining the step length change rate sequence.

[0086] S202: According to the step length change rate sequence, analyze the periodic boundary points between consecutive steps, identify the start and end time nodes of the steps, and combine the joint angle change values recorded on the step time axis to perform segmentation in the time axis direction and divide the angle change data segments within the complete period to obtain the joint angle time series interval;

[0087] Further analyze the periodic boundary between the user's consecutive steps, judge the start and end points of each independent step. When identifying the periodic boundary points, the system uses the turning points, minimum values, or sudden change points in the step length change rate curve as references for period division. When the step length suddenly decreases after continuous increase, this turning point is the estimated step termination sign. Combining with the timestamp information, the system locates the start time and end time of each step cycle and synchronously extracts the corresponding joint angle change data on the time axis. The data can be recorded in real time through an inertial measurement unit (IMU), including the angle changes of major movement nodes such as the knee joint and ankle joint. The system will divide the entire angle data into several segments according to the step cycle. Set cycle 1 to correspond to the angle change sequence from 1.2 seconds to 1.8 seconds. Each data segment is arranged in the order of the time axis, retaining the angle change trend information, reflecting the specific angle activity range and rhythm characteristics of the user during each step movement, and providing a structural input for subsequent step frequency feature recognition to form the joint angle time series interval.

[0088] S203: Extract the step frequency values within each gait cycle based on the joint angle time series interval, identify the difference between the step frequency change amplitude and the average cycle step frequency in adjacent cycles, combine the time span to determine the offset direction and offset rate, calculate the step frequency offset value, and obtain the step frequency offset recognition result;

[0089] The formula for calculating the step frequency offset value is:

[0090]

[0091] where, Δf z represents the step frequency offset value of the z-th gait cycle, f z represents the step frequency value of the z-th gait cycle, f z-1 represents the step frequency value of the (z - 1)-th gait cycle, T z represents the time span of the z-th gait cycle, T z-1 represents the time span of the (z - 1)-th gait cycle, represents the average value of the step frequency values from the (z - 3)-th to the z-th gait cycles;

[0092] Parameter meaning and formula calculation derivation process:

[0093] The step frequency values f z and f z-1 are the step frequency information collected by the lower limb inertial measurement unit IMU, calculated from the number of steps per unit time during each step movement. The sampling frequency is set to 100Hz during the detection process, and the step frequency data is collected for each gait cycle. The zero velocity update method is used to identify the start and end points of the cycle. In actual collection, the step frequency of the z-th cycle is f z = 1.82Hz, and the step frequency of the (z - 1)-th cycle is f z-1 = 1.67Hz;

[0094] The gait cycle time spans T z and T z-1 are obtained by taking the reciprocal of the step frequency, respectively:

[0095]

[0096] The average step frequency value is calculated from the step frequencies in the last four cycles. Using the same method with the IMU, the step frequencies of the first two cycles are f z-2 = 1.76Hz, f z-3 = 1.71Hz, then:

[0097]

[0098] Substitute all the obtained parameters into the formula:

[0099]

[0100] The items in the calculation are as follows: Step frequency difference:

[0101] 1.82 - 1.67 = 0.15;

[0102] Sum of time:

[0103] 0.549 + 0.599 = 1.148;

[0104] Absolute value of time difference:

[0105] |0.549 - 0.599| = 0.05;

[0106] Absolute value of the difference between step frequency and the mean:

[0107] |1.82 - 1.74| = 0.08;

[0108] Calculation of the square root part:

[0109]

[0110] Denominator part:

[0111] 2 · 0.3606 = 0.7212;

[0112] Numerator part:

[0113] 0.15 · 1.148 = 0.1722;

[0114] Substitute into the formula for calculation:

[0115]

[0116] This result indicates that in the z-th cycle, relative to the previous cycle, there is a composite difference index of approximately 0.2387 in the step frequency under the influence of the time span change and the individual average step frequency level. The value reflects the amplitude trend of the step frequency fluctuation between cycles, provides a quantitative input for the subsequent calculation of the step frequency deviation rate, and this value participates in the derivation of the step frequency deviation direction and rate as an intermediate variable parameter.

[0117] S204: According to the step frequency deviation recognition result, match the step frequency pattern and the joint angle change rhythm in each cycle, screen the data segments that meet the step frequency deviation rate threshold range and the angle change rhythm continuity condition, conduct a gait behavior pattern comparison, and obtain the gait matching value;

[0118] The step frequency patterns extracted in each period are standardized and matched, and the rhythm of joint angle changes in the corresponding period is synchronously analyzed. To ensure the accuracy of matching, the system filters out the periods where the step frequency offset value is within the set threshold range, and sets the stable range of the offset rate to ±0.05Hz / s. All periods with step frequency offset rates within this interval are regarded as candidate periods. The system further verifies whether the joint angle changes in the period have rhythm continuity. The judgment methods include the periodic repetition of angle changes, the balance degree of rising and falling, and the proportional relationship between the maximum and minimum amplitudes. If the angle change waveform shows the same-amplitude sine change in three consecutive periods, it can be determined as a continuous rhythm pattern. The system organizes all period segments that meet the above two conditions into candidate data segments, compares them with the gait behavior data of registered users, constructs a comprehensive comparison vector based on the time-domain structure, angle amplitude sequence, and step frequency difference combination, and identifies whether it is the target user. The whole process can be applied to automatic identity behavior recognition in access control areas, especially helping to improve the accuracy in non-contact authentication scenarios, and outputs the gait matching value.

[0119] The specific steps for obtaining the multi-modal identity authentication result are as follows:

[0120] S301: Call the gait matching value and the bioelectrical impedance matching value, respectively extract the numerical sequences within the same time period, identify the interval distribution range, identify the response weights of the two under the combined input conditions, and obtain the matching value weight ratio;

[0121] After calling the gait matching value and the bioelectrical impedance matching value, the continuous value sequences of the two within the same time window are synchronously extracted to construct a matching value comparison matrix. Select a 10-second time period and record the data at a sampling frequency of once per second. Then, a total of 10 groups of gait matching values and impedance matching values can be extracted. After extraction, numerical variance analysis is performed on the two groups of data respectively to statistically analyze their fluctuation degrees. Set that when a certain matching value sequence shows a significant concentration or dispersion trend within this time period, the system can initially judge its stability, and then further divide its interval distribution range, that is, divide the numerical values into multiple equally spaced sections (such as each section is 0.05), count the frequency distribution of the data within each interval, obtain its interval density curve, and compare it with the data structure of another matching value. During the comparison process, the system will analyze which matching value is more sensitive to the change of the data structure. When the numerical jitter caused by environmental noise disturbance mainly concentrates on the gait matching value, while the impedance matching value remains relatively stable, the system will calculate the response weights of the two under the current conditions according to their change amplitudes and interval coverage rates, and mark it as the matching value fusion weight benchmark in the current scenario, which is used to set the proportion of each modal data in the multi-modal fusion process, and obtain the matching value weight ratio.

[0122] S302: Extract multiple sets of matching value change data under different environmental parameter conditions according to the matching value weight ratio, calculate the numerical trend slope of the weight-dominated matching value under different parameter conditions, determine the fluctuation range of the change curve, and obtain the environmental adaptability matching range.

[0123] Expandable adaptation evaluation will be carried out under multiple different environmental parameter conditions. For example, when external variables such as temperature, humidity, and light are set to different levels (such as high, medium, and low), the output of the matching value will be tracked and extracted. Under each set of environmental parameters, the system extracts a combination sequence of gait matching values and bioelectrical impedance matching values within a certain period of time. Priority is given to analyzing the matching value dominated by the weight. If the weight ratio of the gait matching value reaches 70% under a certain environmental condition, then this value is used as the dominant matching item and is key-tracked. The system analyzes the trend changes of the dominant matching value under different environmental variables, calculates its upward or downward trend in the numerical time series, and represents the trend direction and intensity in the form of a slope. Check whether there are large fluctuations or continuous fluctuations in the trend change curve, that is, determine whether the difference range of the change curve within the sliding window exceeds the set fluctuation threshold. The set threshold is 0.1. If there are continuous changes exceeding this amplitude within a 5-second window, it is marked as an unstable section. The system records the stable and unstable section intervals, forms the response curve structure of the dominant matching value under different environmental variables, aggregates and extracts all interval segments with low fluctuation amplitude and stable slope, and outputs the environmental adaptability matching range.

[0124] S303: Based on the environmental adaptability matching range, identify the matching features with a weight ratio exceeding the set threshold, calculate the deviation change rate and determine the stability of the fluctuation range, and perform corresponding verification with the identity verification records in the access control interaction response to obtain the multi-modal identity authentication result.

[0125] The joint features of the two-modal matching values in this range will be further screened to identify the parts where the weight ratio of the matching value in all data segments exceeds the set threshold. Set the weight threshold to 65%, then only retain the data segments where the weight of the gait matching value or the impedance matching value is greater than this value as the candidate matching features, and monitor the fluctuation deviation of the matching features. By calculating the deviation change rate between consecutive sampled values, its calculation formula is:

[0126]

[0127] where, Δ m p is the deviation change rate of the multi-modal matching value, v j and v j+1 are the matching values at two adjacent time points, t j and t j+1 are the corresponding timestamps.

[0128] Set the dominant matching value to 0.82 at 3 seconds and 0.88 at 4 seconds, then:

[0129]

[0130] Traverse the sequence of matching values in a sliding window manner, extract all the change rate data, and judge the upper and lower bounds of its fluctuation range. If the maximum change rate within several consecutive windows does not exceed the threshold (such as 0.1), it is considered that the fluctuation is stable. Compare the stable matching data segments one by one with the authentication results in the access control interaction record, including verification success and failure labels, verification time, behavioral characteristics, etc., establish the corresponding relationship between the current data segment and the identity. The system uses the combined comparison vector to perform fusion judgment on the multi-modal features of each data segment, mark the matching success and the corresponding user ID, realize cross-modal and cross-environment high-robustness identity verification, and output the multi-modal identity authentication result.

[0131] The specific steps for obtaining the abnormal behavior risk assessment value are as follows:

[0132] S401: Invoke the multi-modal identity authentication result, extract the gait feature data within the corresponding time period, identify the step length, gait angle, and step frequency sequence in consecutive steps, monitor the abnormal step length mutation points, angle distortion points, and step frequency mutation points within a short period of time, calculate the occurrence proportion of various mutation events per unit time, and obtain the abnormal behavior deviation rate.

[0133] The time window covered during the automatic positioning authentication process will be extracted, and the gait feature data within this time period will be obtained. The system will synchronously integrate the step length, gait angle, and step frequency data that match the authentication event in the acquisition device (such as an inertial measurement unit, a pressure sensing array, or a video structured recognition system) according to the timestamp information. For each complete gait cycle, the system extracts three types of characteristic parameters: the step length is used to reflect the degree of body center of gravity displacement, the gait angle records the rotation range of joints such as the knee and ankle during movement, and the step frequency represents the number of steps repeated per unit time. During the recognition stage, the system calculates the change amplitude of the parameters within any time period through the sliding window analysis of the continuous sequence. When it is detected that the difference between a certain step length and the previous and subsequent steps exceeds the threshold, it is marked as a step length mutation point; when the joint angle suddenly shows an abnormal deviation (such as a mutation exceeding a certain angle range), it is recorded as an angle distortion point; if the continuous step frequency change exceeds the specified difference, it is regarded as a step frequency mutation point. The system counts the occurrence times of the above three types of mutation events within a unit time window (such as every 5 seconds), classifies and labels the types of various mutation points, and calculates the proportion between the total number of mutation events and the total number of steps to obtain the abnormal behavior deviation rate.

[0134] S402: Based on the abnormal behavior deviation rate, extract the standard fluctuation ranges of step length, angle, and step frequency under normal behavior, make piecewise judgments on the distribution of real-time deviation rate values in multiple dimensions, calculate the deviation distribution intensity eigenvalue, identify the deviation segments that exceed the standard range, calculate the proportion range of the overall deviation segment in the full sample, and obtain the abnormal behavior risk assessment value;

[0135] The formula for calculating the deviation distribution intensity eigenvalue is as follows:

[0136]

[0137] Among them, RS represents the deviation distribution intensity eigenvalue, n r represents the number of deviation data in the r-th segment, O rq represents the q-th real-time deviation rate observation value in the r-th segment, μ rq represents the mean value corresponding to the q-th data under the corresponding dimension, σ rq represents the standard deviation corresponding to the q-th data under the corresponding dimension, Δ rq represents the deviation difference of the q-th data from the previous moment data in the dimension, ε is a positive constant, w rq represents the key weight of the q-th data in the r-th segment;

[0138] Parameter meaning and formula calculation derivation process:

[0139] In the behavior anomaly detection scenario, it is set that for a certain detected object in the 3rd behavior segment (i.e., r = 3), a total of n3 = 5 deviation rate observation data points are recorded in this segment, and the monitoring dimensions include three items: step length, step frequency, and gait angle;

[0140] The observation data is from an embedded inertial measurement unit (IMU) device, the data frequency is 100Hz, the value window period is 1 second, and the following 5 groups of data are obtained from it:

[0141] The 1st to 5th deviation rate observation values in the 1st dimension (step length):

[0142] O 31 = 0.12, O 32 = 0.15, O 33 = 0.20, O 34 = 0.18, O 35 = 0.22;

[0143] The corresponding mean value under normal behavior (extracted from the mean model of large sample normal behavior training data in the early stage, the step length mean value range is 0.13 - 0.18 meters):

[0144] μ 31 = 0.14, μ 32 = 0.14, μ 33= 0.15, μ 34 = 0.16, μ 35 = 0.16;

[0145] Corresponding standard deviation (reference sample standard fluctuation range, step standard deviation stable at 0.02 - 0.04 meters):

[0146] σ 31 = 0.03, σ 32 = 0.03, σ 33 = 0.03, σ 34 = 0.03, σ 35 = 0.03;

[0147] The offset difference between the front and back time points is calculated by continuous sampling, and the quantization method is the current sampling value minus the value of the previous sampling period:

[0148] Δ 31 = 0.01, Δ 32 = -0.02, Δ 33 = 0.03, Δ 34 = -0.01, Δ 35 = 0.02;

[0149] Weight parameter w rq Influence weight of the reference step length on behavior stability, the value range is 0.3 - 0.5, dynamically adjusted according to the gait stability score result, and the values here are as follows:

[0150] w 31 = 0.40, w 32 = 0.42, w 33 = 0.38, w 34 = 0.41, w 35 = 0.39;

[0151] Set the stability coefficient ε = 0.0001 to avoid division by zero operation;

[0152] Substitute into the calculation process as follows:

[0153] The first item:

[0154]

[0155] The second item:

[0156]

[0157] The third item:

[0158]

[0159] The fourth item:

[0160]

[0161] Item 5:

[0162]

[0163] Sum and take the average:

[0164]

[0165] The result shows that the eigenvalue of the step - length dimension offset distribution intensity in the 3rd segment is 0.0768. Combining with the downstream standard interval boundary determination logic, the corresponding threshold is 0.065. Since the current segment intensity value has exceeded the normal behavior fluctuation range, it is regarded as an offset segment and enters the subsequent proportion calculation step.

[0166] The steps for obtaining the access control permission instruction are specifically as follows:

[0167] S501: Invoke the abnormal behavior risk assessment value, compare it with the set risk - behavior safety threshold, determine whether the risk assessment value exceeds the defined interval, screen the data segments where the gait mutation rate and the angle offset rate exceed the defined interval, mark them as behavior segments to be restricted, and obtain the risk - behavior screening result;

[0168] Invoke the preset security control rule library to make a comparison and determination on this value. The system will load the risk - behavior safety threshold set for the current access control scenario. This threshold can be dynamically set according to different access control levels. It is set to 10% in high - security areas (such as data center computer rooms) and 25% in ordinary office areas. The system numerically compares the current abnormal behavior risk assessment value with this threshold. If the detected value exceeds the safety threshold, it is preliminarily determined that there is a potential abnormal risk in the user's behavior. The system further extracts key abnormal data segments from the original gait sequence, focuses on analyzing whether two types of indicators, the gait mutation rate and the angle offset rate, simultaneously exceed the preset threshold. For example, if the step - length mutation rate exceeds 15% and the angle offset rate exceeds ±30, if it is found that both indicators exceed the standard in a certain time period, that segment will be marked as a behavior segment to be restricted and added to the screening buffer area. All the screened behavior segments will be uniformly output to form a complete risk - behavior screening result, which serves as the pre - condition basis for the access control logic to obtain the risk - behavior screening result.

[0169] S502: Based on the risk - behavior screening result, determine whether the matching value in the multi - modal identity authentication result is within the authorized requirement range, extract the combined value under the corresponding weights of the gait matching value and the impedance matching value, and screen the data where the matching value does not exceed the set range to obtain the initialized authorized access data;

[0170] Perform matching value verification on the current multi-modal identity authentication result. The system determines whether the gait matching value and the impedance matching value in the current authentication record are within the legal range set by authorization. This range is set based on the standard physiological parameters and behavioral data model of the user during registration and is a dynamic range. For example, the gait matching value needs to be between [0.75–0.95], and the impedance matching value needs to be between [0.70–0.92]. The system extracts the combined matching value of the two modalities under the weight combination from the current authentication data. According to the previously calculated matching value fusion weights (such as gait 0.6, impedance 0.4), the fusion result is calculated in sequence and then the interval judgment is performed. If the matching value combination is within the authorized range, the system marks it as compliant data and automatically incorporates this segment of data into the initialized authorized access data. Conversely, if the fusion value deviates from the set interval, this segment of data will not be included in the subsequent access judgment process. All the confirmed compliant fusion matching data will be output to determine whether to trigger supplementary verification, further control the final access permission status, and obtain the initialized authorized access data.

[0171] S503: According to the initialized authorized access data, determine whether the matching value deviation is within the matching range limit. If there is a phenomenon that the combined matching value deviates beyond the set range in the two dimensions of step frequency and impedance, it is marked as requiring supplementary verification to trigger secondary verification and obtain the access control permission instruction.

[0172] Further error judgment will be performed on the combined matching value of the two dimensions of step frequency and impedance. This process is based on the previously generated gait matching value and impedance matching value, and combines their fusion weights to perform boundary inspection on the combined matching result, and weighted integration of the two matching values of the current sample is performed. Set the current gait matching value to 0.78, the impedance matching value to 0.82, and the fusion weight is set to gait 0.6, impedance 0.4. The system combines the two to calculate the combined matching value of 0.796. Compare this combined result with the matching range limit set by the system. If the current combined matching value is lower than the lowest authorized threshold (such as 0.80), it is determined that there is a deviation beyond the matching range. The system will separately check the deviations in the step frequency dimension and the impedance dimension. If any dimension deviates from its respective reference interval (such as the step frequency change exceeds ±0.05Hz, the impedance exceeds ±0.03), and this deviation occurs in the combined matching of the current time period, it is marked as a combined anomaly. For such combined deviation situations, the system does not directly issue an access instruction but automatically triggers a supplementary verification mechanism, sets a prompt to require the user to perform secondary identity verification, or invokes an alternative biometric recognition method such as a face or finger vein recognition module. Finally, the system will decide whether to grant access control permission based on the processing result of the secondary verification. If the supplementary verification is passed, an authorization instruction will be issued. If it still fails, the system will reject the current access request and archive the behavior record to obtain the access control permission instruction.

[0173] Please refer to Figure 2 , an intelligent security system based on the Internet of Things, the system includes:

[0174] The impedance recognition module obtains the skin contact points, contact duration data and current response values of the user touching the access control induction area, analyzes the current response changes in different frequency bands, calculates the change amplitude and change frequency between multi-band responses, screens the abnormal fluctuation data sections within a short time, and obtains the bio-impedance matching value;

[0175] The gait recognition module, based on the bio-impedance matching value, detects the movement amplitude of the lower limb joints, single-step length and walking cycle interval of the user during the process of approaching the access control area, divides the angle change nodes in continuous time periods, evaluates the joint change trajectories in multiple cycle segments, and obtains the gait matching value;

[0176] The identity fusion module calls the gait matching value and the bio-impedance matching value, detects the change fluctuation trend under the air humidity value, temperature value and contact pressure data in the real-time access control area, counts the change value of the weight ratio under different environment combinations and judges the real-time weight dominant item, and obtains the multi-modal identity authentication result;

[0177] The behavior risk monitoring module, according to the multi-modal identity authentication result, monitors the continuous step length value, joint torsion angle interval and step frequency change amount of the real-time contact user within a short time, screens the abnormal offset degree corresponding to the abnormal fluctuation point, and obtains the abnormal behavior risk assessment value;

[0178] The permission determination module compares the abnormal behavior risk assessment value with the set risk safety threshold, screens the risk behaviors according to whether the deviation rate exceeds the threshold, and judges whether it belongs to the authorized range by judging the interval between the real-time matching rate and the access control permission standard value under the condition that the risk condition is not triggered, and obtains the access control permission instruction.

[0179] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. An intelligent security method based on the Internet of Things, characterized in that, It includes the following steps: S1: Obtain the bioelectrical impedance signal of the user touching the access control sensing area, identify the current response situation in different frequency bands, match the stored user data, analyze the continuity of impedance parameters, screen out abnormally fluctuating data, and identify the fluctuation frequency and change amplitude within a short time to obtain the bioelectrical impedance matching value; S2: Call the bioelectrical impedance matching value, monitor the user gait data when the user approaches the access control area, calculate the length change rate of adjacent steps, perform time series segmentation on the joint angle change, calculate the step frequency offset rate and perform gait behavior feature matching to obtain the gait matching value; S3: Calculate the weight ratio of the gait matching value and the bioelectrical impedance matching value, identify the matching feature with the priority of weight ratio, and analyze the stability of the matching deviation to obtain the multi-modal identity authentication result; S4: Utilize the multi-modal identity authentication result to analyze the change situation of gait features within a short time, monitor the step length mutation, gait angle distortion and step frequency change situation, calculate the abnormal behavior offset rate, compare it with the normal behavior data to obtain the abnormal behavior risk assessment value, and conduct security warning according to the priority of the risk assessment value.

2. The intelligent security method based on the Internet of Things according to claim 1, wherein, The bioelectrical impedance matching value includes impedance stability, conductance fluctuation amplitude, and dielectric abnormal frequency. The gait matching value includes step length change rate, joint movement period, and step frequency stability. The multi-modal identity authentication result includes feature weight ratio, matching threshold deviation, and environmental adaptability. The abnormal behavior risk assessment value includes step length abnormality rate, joint distortion degree, and step frequency mutation index.

3. The intelligent security method based on the Internet of Things according to claim 1, characterized in that The specific steps for obtaining the bioelectrical impedance matching value are as follows: S101: Obtain the bioelectrical impedance signal of the user touching the access control sensing area, transmit multi-frequency weak current to the skin, record the conductivity, permittivity, and impedance distribution, and identify the combined characteristics of conductivity and permittivity in multiple frequency bands according to the response situation in different frequency bands to obtain the current response frequency band characteristic value; S102: Based on the current response frequency band characteristic value, call the preset user data for step-by-step frequency band response comparison, conduct numerical comparison on the combined values of conductivity and permittivity in multiple frequency bands, identify the response situation and change amplitude of the corresponding frequency band, and screen out the data with the frequency band response difference within the comparison range to obtain the impedance characteristic corresponding interval; S103: According to the impedance characteristic corresponding interval, continuously extract the impedance change curve within the frequency band, identify the impedance fluctuation points within a short time, count the fluctuation frequency and change amplitude to obtain the bioelectrical impedance matching value.

4. The intelligent security method based on the Internet of Things according to claim 1, wherein, The specific steps for obtaining the gait matching value are as follows: S201: Call the bioelectrical impedance matching value, monitor the user gait situation when the user approaches the access control area, record the position point sequence and timestamp information in consecutive steps, extract the displacement data and time interval between each step, and calculate the length change ratio of adjacent steps to obtain the step length change rate sequence; S202: Parse the periodic boundary points between consecutive steps according to the step length change rate sequence, identify the time nodes of the start and end of the steps, and segment in the time axis direction by combining the joint angle change values recorded on the step time axis to divide the angle change data segments within a complete period and obtain the joint angle time series interval; S203: Based on the joint angle time series interval, extract the step frequency values within each gait cycle, identify the change amplitude of the step frequency between adjacent cycles and the difference from the average cycle step frequency, combine the time span to judge the offset direction and offset rate, calculate the step frequency offset value, and obtain the step frequency offset recognition result; The formula for calculating the step frequency offset value is: Among them, Δf z represents the step frequency offset value of the z-th gait cycle, f z represents the step frequency value of the z-th gait cycle, f z-1 represents the step frequency value of the (z - 1)-th gait cycle, T z represents the time span of the z-th gait cycle, T z-1 represents the time span of the (z - 1)-th gait cycle, represents the average value of the step frequency values from the (z - 3)-th to the z-th gait cycles; S204: According to the step frequency offset recognition result, match the step frequency pattern and the joint angle change rhythm within each cycle, screen the data segments that meet the step frequency offset rate threshold range and the condition of the continuity of the angle change rhythm, and perform gait behavior pattern comparison to obtain the gait matching value.

5. The intelligent security method based on the Internet of Things according to claim 1, characterized in that The specific steps for obtaining the multi-modal identity authentication result are as follows: S301: Call the gait matching value and the bioelectrical impedance matching value, extract the numerical sequences within the same time period respectively, identify the interval distribution range, and identify the response weights of the two under the joint input condition to obtain the matching value weight ratio; S302: According to the matching value weight ratio, extract multiple groups of matching value change data under different environmental parameter conditions, statistically analyze the numerical trend slope of the weight-dominated matching value under different parameter conditions, judge the fluctuation interval of the change curve, and obtain the environmental adaptability matching interval; S303: Based on the environmental adaptability matching interval, identify the matching features with a weight ratio exceeding the set threshold, calculate the deviation change rate and judge the stability of the fluctuation range, and perform corresponding verification with the identity verification records in the access control interaction response to obtain the multi-modal identity authentication result.

6. The intelligent security method based on the Internet of Things according to claim 1, wherein The specific steps for obtaining the abnormal behavior risk assessment value are as follows: S401: Call the multi-modal identity authentication result, extract the gait feature data within the corresponding time period, identify the step length, gait angle and step frequency sequence in consecutive steps, monitor the abnormal step length mutation points, angle distortion points and step frequency mutation points in a short time, and calculate the occurrence ratio of multiple types of mutation events per unit time to obtain the abnormal behavior offset rate; S402: Based on the abnormal behavior offset rate, extract the standard fluctuation intervals of the step length, angle and step frequency under normal behavior, perform segmented judgment on the distribution of the real-time offset rate value in multiple dimensions, calculate the offset distribution intensity feature value, identify the offset segments exceeding the standard interval, and calculate the proportion interval of the overall offset segment in the full sample to obtain the abnormal behavior risk assessment value.

7. The intelligent security method based on the Internet of Things according to claim 1, wherein, The formula for calculating the offset distribution intensity feature value is as follows: Among them, RS represents the offset distribution intensity eigenvalue, n r represents the number of offset data in the r-th segment, O rq represents the q-th real-time offset rate observation value in the r-th segment, μ rq represents the mean value corresponding to the q-th data in the corresponding dimension, σ rq represents the standard deviation corresponding to the q-th data in the corresponding dimension, Δ rq represents the offset difference in the dimension between the q-th data and the data at the previous moment, ε is a positive constant, w rq represents the key weight of the q-th data in the r-th segment.

8. The intelligent security method based on the Internet of Things according to claim 1, characterized in that The method further includes step S5: S5: Compare the abnormal behavior risk assessment value with the safety threshold, screen the risk behaviors for permission restriction, judge whether the matching value of the multi-modal identity authentication result meets the authorization requirements, screen the data with the matching value not exceeding the set range and authorize passage, and if the deviation of the matching value exceeds the matching range, trigger secondary verification to obtain the access control permission instruction; The access control permission instruction includes an authorized access instruction, a permission restriction instruction, and a secondary verification trigger condition.

9. The intelligent security method based on the Internet of Things according to claim 1, characterized in that, The specific steps for obtaining the access control permission instruction are as follows: S501: Invoke the abnormal behavior risk assessment value, compare it with the set risk behavior safety threshold, determine whether the risk assessment value exceeds the defined range, screen the data segments where the gait mutation rate and angle deviation rate exceed the defined range, mark them as behavior segments to be restricted, and obtain the risk behavior screening result; S502: Based on the risk behavior screening result, determine whether the matching value in the multimodal identity authentication result is within the authorized requirement range, extract the combined value under the corresponding weights of the gait matching value and the impedance matching value, screen the data where the matching value does not exceed the set range, and obtain the initialized authorized access data; S503: According to the initialized authorized access data, determine whether the matching value deviation is within the matching range limit. If there is a phenomenon that the combined matching value deviates beyond the set range in both the stride frequency and impedance dimensions, mark it as requiring supplementary verification, trigger secondary verification, and obtain the access control permission instruction.

10. The intelligent security system based on the Internet of Things is characterized in that, According to the Internet of Things-based intelligent security method according to any one of claims 1-9, the system includes: The impedance identification module obtains the skin contact point, contact duration data, and current response value of the user contacting the access control sensing area, analyzes the current response changes in different frequency bands, calculates the change amplitude and change frequency between multi-band responses, screens the data segments with abnormal fluctuations in a short time, and obtains the bio-impedance matching value; The gait recognition module, based on the bio-impedance matching value, detects the movement amplitude of the lower limb joints, single-step length, and walking cycle interval of the user during the process of approaching the access control area, divides the angle change nodes in continuous time periods, and evaluates the joint change trajectory in multiple cycle segments to obtain the gait matching value; The identity fusion module invokes the gait matching value and the bio-impedance matching value, detects the change fluctuation trend under the air humidity value, temperature value, and contact pressure data in the real-time access control area, statistically analyzes the change value of the weight ratio under different environmental combinations, and determines the real-time weight dominant term to obtain the multimodal identity authentication result; The behavior risk monitoring module, according to the multimodal identity authentication result, monitors the continuous step length value, joint torsion angle interval, and stride frequency change amount of the user in real-time contact within a short time, screens the abnormal deviation degree corresponding to the abnormal fluctuation points, and obtains the abnormal behavior risk assessment value; The permission determination module compares the abnormal behavior risk assessment value with the set risk safety threshold, screens the risk behaviors according to whether the deviation rate exceeds the threshold, and judges whether it belongs to the authorized range by judging the interval between the real-time matching rate and the access control permission authorization standard value under the condition that the risk condition is not triggered, and obtains the access control permission instruction.

Images