Wireless communication method and device of intelligent door lock

Abstract

The invention discloses a wireless communication method and device of an intelligent door lock, and relates to the field of door lock wireless communication, the device comprises a switching decision module, a safety matching module, a link control module, a power control adjustment module and a response module, a fractional order random channel evolution model is constructed, and a metreger-Leifler function is adopted for analysis, so that a wireless communication mode of the intelligent door lock is established. If the threshold value is exceeded, a frequency band switching signal is sent out; a frequency band priority score matching encryption measure is calculated through a frequency band priority score function; an eight-dimensional state vector of a link is obtained; a topological boundary forming a stable state is analyzed; the method comprises the following steps: establishing a transmitting power adjustment model, performing power response speed adjustment on the transmitting power adjustment model through a fractional order adaptive control function, performing frequency band calculation and switching, obtaining frequency values of frequency band switching, and taking measures corresponding to different frequency values, thereby ensuring safety and wireless communication stability.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication for door locks, and specifically to a wireless communication method and device for intelligent door locks. Background Art

[0002] With the continuous popularization of Internet of Things technology, intelligent door locks have now become the core entrance to home security. However, their security and wireless communication stability face severe challenges. Traditional wireless communication control for door locks often relies on manual intervention and cannot meet the current wireless communication requirements for intelligent door locks. Therefore, the wireless communication method and device for intelligent door locks came into being.

[0003] Most of the existing wireless communication methods and devices for intelligent door locks use traditional integer-order models to describe the dynamic characteristics of the channel, making it difficult to accurately depict the historical cumulative effect of random interference from devices such as Wi-Fi and Bluetooth and the non-local dependence of channel quality, resulting in a lagging or over-reactive frequency band switching strategy. At the same time, existing wireless communications generally adopt encryption strategies and power control with fixed parameters and cannot adapt to dynamic change scenarios such as device energy consumption, network congestion, and door lock operation frequency, leading to frequent security vulnerabilities, abnormal power consumption, or communication interruptions. Meanwhile, in the existing complex electromagnetic environment, the quality of the wireless channel is affected by various factors such as co-channel interference and multipath fading and exhibits non-Gaussian noise characteristics. Existing Gaussian white noise models and integer-order differential equations are difficult to capture fractional-order dynamic behavior and its long memory. In addition, the stability analysis of the existing wireless communication for link states (such as multi-dimensional parameters such as spectrum occupancy rate, delay, and bit error rate) mostly relies on single indicators or threshold comparisons, lacking geometric analysis and systematic modeling of the high-dimensional state space, further resulting in insufficient anomaly detection accuracy. In terms of power control, the existing fixed-weight transmit power adjustment strategy for the wireless communication of intelligent door locks cannot respond in real time to network congestion and changes in user activity patterns, easily causing power waste or deterioration of communication quality.

[0004] To address the above defects, a technical solution is provided now. Summary of the Invention

[0005] To solve the technical problems raised in the above background art, the present invention is proposed. Embodiments of the present invention provide a wireless communication method and device for intelligent door locks.

[0006] The object of the present invention can be achieved by the following technical solutions: A wireless communication method for an intelligent door lock, characterized by comprising the following steps: Step S100: Construct a fractional-order random channel evolution model, analyze the model using the Mittag-Leffler function, generate sample paths using Monte Carlo simulation, obtain the channel quality variance through the variance function, and if it exceeds the threshold, send a frequency band switching signal and select the highest scoring frequency band for switching; Step S200: Calculate the frequency band priority score using the frequency band priority scoring function for the frequency band after switching and the non - issued frequency band switching signal, match the corresponding encryption measures, issue a warning for the corresponding frequency band priority score, and send the frequency band switching signal back to Step S100 for frequency band calculation and switching; Step S300: Obtain the eight - dimensional state vector of the link, calculate and analyze it using the Gaussian - type fuzzy membership function to obtain the fuzzy precise conversion index, limit the parameters in the safe area using the dynamic barrier function, fuse through the hybrid constraint function, map the eight - dimensional state parameters of the link into a simplex, retain the simplex with a diameter not exceeding the threshold for analysis to form the topological boundary of the stable state, map the topological boundary of the stable state to the sliding mode surface through the sliding mode control function to obtain the sliding mode surface function value of the link. If the absolute value of the sliding mode surface function value is greater than the allowable deviation tolerance, return to Step S100 for frequency band calculation and switching; Step S400: Obtain the door lock power control parameters, establish a transmit power adjustment model, adjust the power response speed of the transmit power adjustment model through the fractional - order adaptive control function, monitor the absolute power error during power adjustment, and return to Step S100 for frequency band calculation and switching for the corresponding state; Step S500: Obtain the frequency values of the frequency band switching in the above steps, and different frequency values correspond to different measures.

[0007] Further, the step of analyzing the switching of the selected highest - scoring frequency band: (1) Construct a fractional - order stochastic channel evolution model to obtain the variation law of the door lock channel quality over time. The fractional - order stochastic channel evolution model includes: D ɑ A(t)=γ×A(t)+σ×D ζ W(t), where D ɑ and D ζ are the Caputo - type fractional - order derivatives respectively, A(t) represents the channel quality at time t, γ represents the state decay coefficient, σ is the noise intensity, and W(t) is the standard Brownian motion; (2) Analytically analyze the fractional - order stochastic channel evolution model through the Mittag - Leffler function to obtain the probability distribution of the channel quality, where the Mittag - Leffler function: , where E ɑ，β (k) is the Mittag - Leffler function, h is the summation index, ɑ and β are function parameters, k is the input variable, Γ(*) is the gamma function. Substitute it into the fractional - order stochastic channel evolution model for analysis to obtain the model solution form: A(t)=A(0)×E ɑ，1 (γ×t ɑ )+σ×∫(t - s) ɑ-1 E ɑ，ɑ(γ × (t - s) ɑ ) D ζ W(s)ds, where A(0) represents the channel quality at the initial moment, s is the integration variable, and based on the sliding window data, the numerical values of the parameters ɑ, ζ, γ, and σ are estimated using the fractional least squares method. By minimizing the objective function , where represents the squared norm of the vector, measuring the degree of difference, i is the summation index variable, Φ1 is the regularization parameter, the parameter ζ is implicit in the calculation of A(t), and the parameters ɑ, γ, and σ jointly participate in the optimization adjustment of the objective function. By iterative solution, the objective function is minimized to obtain the numerical values of ɑ, ζ, γ, and σ. A large number of sample paths are generated using Monte Carlo simulation, and the distribution of A(t) at different time points is statistically analyzed , where P(A, t) is the probability density function of the channel quality being A at time t, δ(*) is the Dirac function. When A = , then δ(A - = 1, otherwise it is equal to 0; (3) The channel quality variance Var(A(t)) is obtained through the variance function = ∫A 2 × P(A, t)dA - u(t) 2 , where A represents the value of the channel quality, u(t) is the mean of the channel quality A(t). If the channel quality variance is greater than the set threshold kl, a frequency band switching signal is sent, and no other operations are performed. The fractional-order random channel evolution model is analyzed for frequency band stability through the frequency band priority scoring function, and the frequency band with the highest selected channel score is obtained and switched. The frequency band priority scoring function is , where f j is the serial number of the frequency band, A hist (s) is the historical quality data of the frequency band, T is the time window, and S(f j ) is the priority score of the frequency band.

[0008] Further, the corresponding encryption measure analysis step: Calculate the frequency band priority scores for after the frequency band switching and when no frequency band switching signal is sent using the frequency band priority scoring function in step S100. If the frequency band priority score is within the set comparison interval ll1, the corresponding encryption measure one is used. If the frequency band priority score is within the set comparison interval ll2, the corresponding encryption measure two is used. If the frequency band priority score is within the set comparison interval ll3, a warning is issued, a frequency band switching signal is sent, and the frequency band with the highest calculated channel frequency band priority score in step S100 is obtained and switched.

[0009] Further, the sliding mode surface function value analysis step of the link: Map the eight-dimensional state parameter space of the link to a simplicial complex, set a distance threshold tr1, only retain the simplices with a diameter not exceeding the threshold tr1, merge all the simplices that meet the conditions to form a connected region, and the union of the connected regions forms the topological boundary of the stable state in the link parameter space; Map the obtained topological boundary of the stable state and the mixed constraint function B(x) to a sliding mode surface through a sliding mode control function. The sliding mode control function: , where d i represents the sliding mode coefficient, which is used to adjust the influence weight of each parameter dimension on the sliding mode surface. x i，ref represents the reference stable value, χ represents the control switching speed, sign(*) is the sign function. When B(x) exceeds the threshold B th Force the state trajectory to converge along the sliding mode surface. f(x) represents the sliding mode surface function value of the link. If |f(x)| > jd1, then trigger the frequency band switching signal, where jd1 is the allowable deviation tolerance.

[0010] Furthermore, the analysis steps of the mixed constraint function B(x): Obtain the bit error rate, delay, number of retransmissions, signal-to-noise ratio, delay jitter, spectrum occupancy rate, phase noise, and Doppler frequency shift of the multi-dimensional link, mark them as the state parameters of the multi-dimensional link, and construct an eight-dimensional state vector: x = [x1, x2,..., x i T , where T represents the transpose of the matrix, i represents the serial number of the eight-dimensional state vector. Design a Gaussian-type fuzzy membership function for each parameter. The membership degree u 低BE (x i ) = exp(-(x i - z low ) 2 ) / (2η 2 ), where exp represents the exponential function with the natural constant e as the base, z low represents the ideal low bit error rate center value, η is the standard deviation. If the stability levels are unstable, medium, and stable, the corresponding center values are 0.25, 0.65, and 0.95 respectively. Obtain the membership degrees of condition 1 and condition 2 in the stability fuzzy rules of each multi-dimensional link, and take their minimum value and mark it as the activation intensity; calculate the activation intensity matrix h mn of the nth rule in all samples, where m is the sample index, calculate the entropy value of each rule , according to the formula , obtain the weight w n of each rule, where M is the number of samples, N is the number of rules, and use the center method formula to obtain the fuzzy precise conversion index W(x); ​By means of the dynamic barrier function G(x), the state parameters of the multi-dimensional link are forced within the safe area. The specific dynamic barrier function is as follows: , where x i，max and x i，min represent the maximum and minimum values allowed for the state parameters of the multi-dimensional link, and ε is an extremely small constant to prevent the denominator from being zero. The dynamic barrier function G(x) and the fuzzy exact conversion exponent W(x) are combined through the hybrid constraint function B(x) = G(x) × (1 + Ω × W(x)), where Ω is the fuzzy weight coefficient, to obtain the hybrid constraint function B(x).

[0011] Furthermore, the steps for analyzing the stability fuzzy rules of the multi-dimensional link are as follows: The state parameters of the multi-dimensional link are subjected to data standardization processing as input parameters, and the link stability level is used as the output target. The CART algorithm is used to train a decision tree, with the information gain ratio as the splitting criterion. Each path from the root to the leaf in the trained decision tree is transformed into a clear rule in the form of "IF-THEN". Fuzzy sets are defined for each input parameter, and based on the above Gaussian-type fuzzy membership function, the thresholds in the decision tree are mapped to the intersection points of the fuzzy subsets. The conditions of the clear rules are replaced with combinations of fuzzy subsets to obtain fuzzy rules. The membership degrees of the real-time input parameters in each fuzzy subset are calculated, and the activation strength of each fuzzy rule is calculated according to the logical relationship of the rule premise conditions. Finally, the stability level corresponding to the rule with the maximum activation strength is selected as the output. If the strengths of multiple rules are the same, the average level is taken to obtain the stability fuzzy rules of the multi-dimensional link.

[0012] Furthermore, the steps for analyzing the absolute power error when monitoring and adjusting the power are as follows: Obtain the user activity patterns of the door lock, select the operation frequency, the standard deviation of the operation time interval, and the night operation in the user activity patterns to construct a three-dimensional input space. Standardize each index, and use the fuzzy K-means clustering algorithm to obtain the membership degree vectors of each sample for the three types of activity patterns of low, medium, and high through iterative optimization of the objective function. Calculate the three-dimensional feature weights by the entropy weight method, calculate the entropy values of each feature, and convert them into weights wj, where j represents the serial number of the index. Calculate through the weighted centroid method to obtain the user activity clustering model value y; And obtain the network congestion index, device moving speed, remaining battery power ratio of the door lock, and the user activity aggregation mode value y to form the door lock power control parameters, and establish a transmit power adjustment model, specifically including: Ptx(t) = Pbas × (ω1(t) / (1 + Nconge) + ω2(t) × Bcuf + ω3(t) × vrove / vstand + ω4(t) × y), where Ptx(t) represents the transmit power adjustment value at time t, Pbas represents the reference power, ω1, ω2, w3, and ω4 represent the weights of the door lock power control parameters, and the weights are adjusted in real time through reinforcement learning, Nconge represents the network congestion index, Bcuf represents the remaining battery power ratio, vrove represents the device moving speed, and vstand represents the device moving standard speed; Adjust the power response speed of the transmit power adjustment model through the fractional-order adaptive control function. The fractional-order adaptive control function includes: D ν Ptx(t) = κ × (Ppred - Ptx(t)) + ϵ0 × D u ×W(t), where D ν , D u represents the Caputo fractional derivative, Ppred represents the target transmit power, κ represents the convergence rate, ϵ0 represents the environmental disturbance intensity, W(t) is a noise term simulating burst interference, and the absolute error between the target power and the actual power ep(t) = |Ppred - Ptx(t)| is monitored in real time. If the absolute error ep(t) is greater than the set threshold eth for a duration greater than tfau, it is determined that the power adjustment fails, the power adjustment is stopped, and then return to step S100 for frequency band calculation and switching.

[0013] Furthermore, the real-time weight adjustment analysis step: Collect the historical data of the door lock power control parameter values and the corresponding channel quality variance Var(A(t)). Process the door lock power control parameter values using the Z-score standardization method, denoted as the input vector O(t), construct a data set, use the input vector as the input, divide the training set and the test set in a ratio of 8:2, train a random forest model with the training set, use the bootstrap sampling method and randomly select features when constructing the decision tree, calculate the reduction in channel quality variance for different feature splits when splitting nodes, and accumulate the contribution measures of the variance reduction of each feature after the entire tree is constructed. Calculate the mean square error function MSE = 1 / |Dtest| × ∑(O(t), Var(A(t))) ∈ Dtest(Var(Apred(t)) - Var(A(t))) 2, where |Dtest| represents the number of samples in the test set Dtest, and Var(Apred(t)) is the predicted value. If the mean squared error MSE exceeds the threshold, the depth of the tree is adjusted until the mean squared error value is less than the set threshold and then ends. The real-time input door lock power control parameters are substituted into the trained model, and the contribution measure is normalized to obtain the real-time contribution value fi(t) of each factor to the channel quality variance. The weight is adjusted in real time through the weight adjustment rule wi(t + 1) = wi(t) + η0×(Var target - Var(A(t))), where i represents the serial number of the door lock power control parameter, η0 represents the learning rate, and Vartarget represents the standard target channel quality variance.

[0014] Furthermore, the different frequency values correspond to different measure analysis steps: Obtain the number of frequency band switches of the door lock per unit time in the above steps, and mark it as the frequency value of the frequency band switch. If the frequency value of the frequency band switch is between the threshold values tv1 and tv2, then divide the threshold kl of the channel quality variance set in step S100, the deviation tolerance jd1 allowed in step S300, and the threshold eth of the absolute error ep(t) set in step S400 by the correction coefficient θ1. If the frequency value of the frequency band switch is greater than the threshold value tv2, then issue a warning and push an alarm message containing specific abnormal data to the administrator terminal.

[0015] As a preferred embodiment of the present invention, a wireless communication device for an intelligent door lock includes a switching decision module, a security matching module, a link control module, a power control adjustment module, and a response module. The switching decision module constructs a fractional-order stochastic channel evolution model, analyzes the model using the Mittag-Leffler function, generates sample paths using Monte Carlo simulation, obtains the probability distribution of the channel quality, and obtains the channel quality variance through the variance function. If it exceeds the threshold, a frequency band switching signal is issued, and the highest scoring frequency band is selected for switching using the frequency band priority scoring function. The security matching module calculates the frequency band priority score using the frequency band priority scoring function for the frequency band after switching and for the case where no frequency band switching signal is issued, and matches the corresponding encryption measures. If the frequency band priority score is within the set comparison interval ll3, then issue a warning and issue a frequency band switching signal to return to the switching decision module for frequency band calculation and switching. The link control module obtains the eight-dimensional state vector of the link, calculates and analyzes it using the Gaussian fuzzy membership function to obtain the fuzzy-precise conversion index, limits the parameters in the safe area using the dynamic barrier function, fuses them through the hybrid constraint function, maps the eight-dimensional state parameters of the link into a simplicial complex, retains the simplices with diameters not exceeding the threshold for analysis to form the topological boundary of the stable state, maps the topological boundary of the stable state to the sliding mode surface through the sliding mode control function to obtain the sliding mode surface function value of the link. If the absolute value of the sliding mode surface function value is greater than the allowable deviation tolerance, it returns to the handover decision module for frequency band calculation and handover; The power control adjustment module obtains the door lock power control parameters, establishes a transmit power adjustment model, adjusts the power response speed of the transmit power adjustment model through the fractional-order adaptive control function, monitors the absolute power error during power adjustment, and returns the corresponding status to the decision module for frequency band calculation and handover; The response module obtains the frequency values for frequency band handover, and different frequency values correspond to different measures.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. The present invention constructs a fractional-order stochastic channel evolution model, analyzes the model using the Mittag-Leffler function, generates sample paths through Monte Carlo simulation, obtains the channel quality variance through the variance function, emits a frequency band switching signal if it exceeds the threshold, selects the highest-scoring frequency band for switching, calculates the frequency band priority score for the frequency band after switching and for the case where no frequency band switching signal is emitted using the frequency band priority scoring function, matches the corresponding encryption measures, issues a warning for the corresponding frequency band priority score, and emits a frequency band switching signal to return for frequency band calculation and switching. The eight-dimensional state vector of the link is obtained and analyzed using the Gaussian-type fuzzy membership function to obtain the fuzzy-precise conversion index. The parameters are limited to the safe area using the dynamic barrier function, and fusion is performed through the hybrid constraint function. The eight-dimensional state parameters of the link are mapped to a simplex, and the simplex with a diameter not exceeding the threshold is retained for analysis to form the topological boundary of the stable state. The topological boundary of the stable state is mapped to the sliding mode surface through the sliding mode control function to obtain the sliding mode surface function value of the link. If the absolute value of the sliding mode surface function value is greater than the allowed deviation tolerance, it returns for frequency band calculation and switching, obtains the door lock power control parameter, establishes a transmit power adjustment model, and adjusts the power response speed of the transmit power adjustment model through the fractional-order adaptive control function, monitors the absolute power error during power adjustment, and returns to the corresponding state for frequency band calculation and switching. It can accurately describe the historical cumulative effect of random interference of devices such as Wi-Fi and Bluetooth and the non-local dependence of channel quality, can adapt to dynamic change scenarios such as device energy consumption, network congestion, and door lock operation frequency, can exhibit non-Gaussian noise characteristics under the influence of various factors such as co-channel interference and multipath fading, capture the fractional-order dynamic behavior and its long memory, can perform geometric analysis and systematic modeling of the high-dimensional state space, and can dynamically adjust the weight of the transmit power to respond in real time to network congestion and changes in user activity patterns.

[0017] 2. The present invention obtains the frequency values of frequency band switching, where different frequency values correspond to different measures. By constructing a multi-layer closed-loop architecture of channel modeling - link control - power adjustment - frequency monitoring, it solves the deficiencies of traditional solutions in historical dependence modeling, dynamic interference adaptation, and cross-layer collaborative optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for description in the embodiments. The following drawings are not deliberately drawn to scale in actual size, and the focus is on showing the gist of the present invention.

[0019] Figure 1 It is the conceptual flowchart of the present invention; Figure 2 It is the method flowchart of the present invention; Figure 3 It is the system block diagram of the present invention. Detailed implementation mode

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts also belong to the scope of protection of the present invention.

[0021] As Figure 1 , Figure 2 shown, a wireless communication method for an intelligent door lock includes the following steps: Step S100: Construct a fractional-order stochastic channel evolution model, analyze the model using the Mittag-Leffler function, generate sample paths using Monte Carlo simulation, obtain the probability distribution of channel quality, obtain the channel quality variance through the variance function, if it exceeds the threshold, send a frequency band switching signal, and select the highest scoring frequency band for switching using the frequency band priority scoring function; Step S200: Calculate the frequency band priority score using the frequency band priority scoring function for the frequency band after switching and the case where no frequency band switching signal is sent, and match the corresponding encryption measures. If the frequency band priority score is within the set comparison interval ll3, then send a warning and send a frequency band switching signal to return to step S100 for frequency band calculation and switching; Step S300: Obtain the eight-dimensional state vector of the link, calculate and analyze it using the Gaussian-type fuzzy membership function to obtain the fuzzy exact conversion index, use the dynamic barrier function to limit the parameters in the safe area, fuse through the mixed constraint function, map the eight-dimensional state parameters of the link into a simplex, retain the simplex with a diameter not exceeding the threshold for analysis to form the topological boundary of the stable state, map the topological boundary of the stable state to the sliding mode surface through the sliding mode control function to obtain the sliding mode surface function value of the link. If the absolute value of the sliding mode surface function value is greater than the allowable deviation tolerance, then return to step S100 for frequency band calculation and switching; Step S400: Obtain the door lock power control parameters, establish a transmit power adjustment model, and adjust the power response speed of the transmit power adjustment model through the fractional-order adaptive control function, monitor the absolute power error during power adjustment, and return to step S100 for frequency band calculation and switching for the corresponding state; Step S500: Obtain the frequency values of the above-mentioned frequency band switching, and different frequency values correspond to different measures.

[0022] Specifically, the analysis of the frequency band switching in step S100 is as follows: (1) Construct a fractional-order stochastic channel evolution model to describe the change law of the door lock channel quality over time. The fractional-order stochastic channel evolution model includes: D ɑ A(t)=γ×A(t)+σ×Dζ W(t), where D ɑ and D ζ are Caputo fractional derivatives respectively, where 0 < ɑ, ζ < 1, A(t) represents the channel quality at time t, γ represents the state attenuation coefficient, σ is the noise intensity, and W(t) is a standard Brownian motion, simulating random interference in the environment, specifically the change in signal strength caused by the random occupancy of devices such as Bluetooth and Wi-Fi; (2) Analytically analyze the fractional-order stochastic channel evolution model through the Mittag-Leffler function to obtain the probability distribution of the channel quality, where the Mittag-Leffler function: , where E ɑ，β (k) is the Mittag-Leffler function, h is the summation index, ranging from 0 to ∞, representing the number of terms in the series expansion, ɑ and β are function parameters, k is the input variable, Γ(*) is the gamma function, which is the extension of the factorial in the real number domain. Substitute it into the fractional-order stochastic channel evolution model for analysis to obtain the model solution form: , where A(0) represents the channel quality at the initial moment, s is the integration variable, representing historical time points. Based on the sliding window data, use the fractional least squares method to estimate the values of the parameters ɑ, ζ, γ, and σ. By minimizing the objective function , where represents the squared norm of the vector, measuring the degree of difference. i is the summation index variable, used to traverse the I data points within the sliding window. Φ1 is the regularization parameter. The parameter ζ is implicitly involved in the calculation of A(t), and the parameters ɑ, γ, and σ jointly participate in the optimization adjustment of the objective function. By iterative solution, the objective function is minimized to obtain the values of ɑ, ζ, γ, and σ. Use Monte Carlo simulation to generate a large number of sample paths and statistically analyze the distribution of A(t) at different time points, , where P(A, t) is the probability density function of the channel quality being A at time t, δ(*) is the Dirac function. When A = , then δ(A - = 1, otherwise it is equal to 0, used to screen out the situation where the channel quality is at the i-th simulation, (3) Obtain the channel quality variance Var(A(t)) = ∫A 2 × P(A, t)dA - u(t) 2 , where A represents the value of the channel quality, u(t) is the mean of the channel quality A(t). If the channel quality variance is greater than the set threshold kl, then send a frequency band switching signal, and perform no other operations. And perform a frequency band stability analysis on the fractional-order stochastic channel evolution model through the frequency band priority scoring function to obtain the frequency band with the highest selected score of the channel and perform the switching, where the frequency band priority scoring function: , where f j is the serial number of the frequency band, A hist (s) is the historical quality data of the frequency band, T is the time window, and S(f j ) frequency band is the priority score; Specifically, the fractional-order stochastic channel evolution model comprehensively simulates the historical dependence, natural attenuation, and random interference in the environment of channel quality by combining fractional-order derivatives, state attenuation coefficients, and standard Brownian motion, thus more accurately describing the variation law of channel quality over time. In this model, D ɑ and D ζ characterize the historical dependence characteristics of channel quality changes, and the fractional-order stochastic channel evolution model is analyzed through the Mittag-Leffler function. The Mittag-Leffler function plays a bridging role here. It transforms the abstract fractional-order derivative model into a specific channel quality expression, enabling the complex fractional-order stochastic model to be analyzed, and thus obtaining the specific form of channel quality changing over time. E ɑ，1 is multiplied by the initial channel quality to describe the evolution law of the initial channel quality over time t, reflecting the channel quality in the case of no noise. E ɑ，ɑ appears in the integral term and combines with the historical noise term to describe the cumulative impact of historical random interference on the current channel quality, characterizing the historical dependence of the noise impact under fractional-order calculus. The three together support the theoretical framework for channel quality analysis and frequency band decision-making. Based on the model analysis results, optimization operations such as frequency band switching can be triggered. Stable frequency bands are screened through the frequency band priority scoring function to improve the reliability and performance of the wireless communication system.

[0023] Specifically, the corresponding encryption measures for step S200 are analyzed as follows: Calculate the frequency band priority score for the frequency band after switching and for the case where no frequency band switching signal is sent using the frequency band priority scoring function in step S100. If the frequency band priority score is within the set comparison interval ll1, then AES-128 encryption is used, and the key is updated every 30 minutes. If the frequency band priority score is within the set comparison interval ll2, then AES-256 encryption is used, and the key is updated every 10 minutes. If the frequency band priority score is within the set comparison interval ll3, then a warning is issued, and a frequency band switching signal is sent, and the frequency band with the highest channel frequency band priority score calculated in step S100 is switched.

[0024] Specifically, the analysis of the sliding mode surface function value in step S300 is as follows: Obtain the bit error rate, delay, number of retransmissions, signal-to-noise ratio, delay jitter, spectrum occupancy rate, phase noise, and Doppler frequency shift of the multi-dimensional link, mark them as the state parameters of the multi-dimensional link, and construct an eight-dimensional state vector: x = [x1, x2,..., x i ​T , where T represents the transpose of the matrix, i represents the serial number of the eight-dimensional state vector, which takes the value [1, 8]. A Gaussian fuzzy membership function is designed for each parameter, and the membership u 低BE (x i ) = exp (- (x i -z low ) 2 ) / (2η 2 ), where exp represents an exponential function with the natural constant e as the base, z low represents the ideal low bit error rate center value, η is the standard deviation, which is used to control the fuzzy range. Using the stability fuzzy rule of the multi-dimensional link, if the stability level is unstable, medium and stable, the corresponding center values ​​are 0.25, 0.65 and 0.95 respectively. The membership of condition 1 and condition 2 in each rule is obtained, and the minimum value is marked as the activation intensity; the activation intensity matrix h of the nth rule in all samples is calculated mn , where m is the sample index, calculate the entropy value of each rule , according to the formula , get the weight w of each rule n , where M is the number of samples and N is the number of rules. Using the center method formula, the fuzzy exact conversion index W(x) is obtained. The center method formula is: W(x)=∑ (w n ×Activation Strength n ×Output center value n ) / ∑ Activation Strength n ; Through the dynamic barrier function G(x), the state parameters of the multi-dimensional link are forced to be within the safe area. The specific dynamic barrier function is: , where x i，max and x i，min represents the maximum and minimum values ​​allowed by the state parameters of the multidimensional link. ε is a very small constant to prevent the denominator from being zero. The dynamic barrier function G(x) is accurately converted to The index W(x) is obtained by the hybrid constraint function B(x)=G(x)×(1+Ω×W(x)), where Ω is the fuzzy weight coefficient. The hybrid constraint function combines the rigid protection of the barrier function with the flexible evaluation of the fuzzy precise conversion index, and can adjust the constraint tightness according to the real-time link status.

[0025] The eight-dimensional state parameter space of the link is mapped to a simplicial complex, and the distance threshold tr1 is set. Only simplexes with diameters not exceeding the threshold tr1 are retained. All simplexes that meet the conditions are merged to form a connected area. The union of the connected areas forms the topological boundary of the stable state in the link parameter space. The above specific formula is The topological boundary of the stable state = ⋃{σ ∈ SimplicialComplex ∣ diam(ψ) ≤ tr1}, where ψ represents a simplex in the simplicial complex, diam(ψ) represents the maximum distance between any two points within the simplex, tr1 represents the distance threshold that controls noise filtering and structure preservation, ⋃ represents the union operation that combines all qualified simplices, and Simplicial Complex represents the set of simplicial complexes constructed from the point set of the parameter space.

[0026] The obtained topological boundary of the stable state is mapped to a sliding mode surface through a sliding mode control function: , where d i represents the sliding mode coefficient, which is used to adjust the influence weight of each parameter dimension on the sliding mode surface, x i，ref represents the reference stable value, which is the ideal state of each parameter within the stability boundary, χ represents the control switching speed, which determines the response rate of the system to return from the deviation state to the sliding mode surface, sign(*) is the sign function, and when B(x) exceeds the threshold B th forces the state trajectory to converge along the sliding mode surface, f(x) represents the sliding mode surface function value of the link. If |f(x)| > jd1, then a frequency band switching signal is triggered, where jd1 is the allowable deviation tolerance.

[0027] Specifically, the analysis steps of the stability fuzzy rules for the multi-dimensional link are as follows: Perform data standardization processing on the state parameters of the multi-dimensional link as input parameters, and use the link stability level as the output target. Specifically, 1 = unstable, 2 = medium, and 3 = stable. Use the CART algorithm to train a decision tree with the information gain ratio as the splitting criterion. Convert each path from the root to the leaf in the trained decision tree into a clear rule in the form of "IF-THEN", such as IF the bit error rate < 5% and the delay < 20ms → stability = 3. Define fuzzy sets for each input parameter and map the thresholds in the decision tree to the intersection points of the fuzzy subsets according to the above Gaussian-type fuzzy membership function. Replace the conditions of the clear rules with combinations of fuzzy subsets to obtain fuzzy rules, such as IF the bit error rate is a1 and the delay is a2 → stability = 3. Calculate the membership degrees of the real-time input parameters in each fuzzy subset, calculate the activation strength of each fuzzy rule according to the logical relationship of the rule premise conditions, and finally select the stability level corresponding to the rule with the maximum activation strength as the output. If the strengths of multiple rules are the same, take the average level to obtain the stability fuzzy rules for the multi-dimensional link; Specifically, the analysis of adjusting the power response speed in step S400 is as follows: Obtain the user activity patterns of the door lock. Select the operation frequency, standard deviation of operation time intervals, and proportion of night operations in the user activity patterns and mark them as rn, σt, and f respectively. Construct a three-dimensional input space, perform standardization processing on each index, and use the fuzzy K-means clustering algorithm to iteratively optimize the objective function , where m represents the number of clustering categories, corresponding to low, medium, and high activity patterns respectively, represents the membership degree of the i-th sample to the m-th category, with a value range of [0, 1], and k represents the fuzzy coefficient, which is used to control the fuzziness of clustering, represents the Euclidean distance between the i-th sample and the center of the m-th category, J represents the objective function of the clustering effect, I represents the number of samples, and obtain the membership degree vectors of each sample to the low, medium, and high activity patterns , , . Calculate the three-dimensional feature weights by the entropy weight method. First, calculate the entropy values of each feature and convert them into weights wj, j = 1, 2, 3, where j represents the serial number of the index. The specific weight calculation operation has been discussed above and will not be repeated here. Calculate through the weighted centroid method to obtain the user activity clustering mode value y. The weighted centroid method is , where a0, a1, and a2 are the quantization reference values of the three types of patterns, and y ∈ [0, 1]; And obtain the network congestion index, device movement speed, and remaining battery percentage of the door lock, and combine them with the user activity clustering mode value y to form the door lock power control parameters, and establish a transmit power adjustment model. Specifically, it includes: Ptx(t) = Pbas × (ω1(t) / (1 + Nconge) + ω2(t) × Bcuf + ω3(t) × vrove / vstand + ω4(t) × y), where Ptx(t) represents the transmit power adjustment value at time t, Pbas represents the reference power, ω1, ω2, w3, and ω4 represent the weights of the door lock power control parameters and are adjusted in real time through reinforcement learning, Nconge represents the network congestion index, Bcuf represents the remaining battery percentage, vrove represents the device movement speed, and vstand represents the device movement standard speed, The specific steps for real-time adjustment of the weights are as follows: Collect the historical data of the lock power control parameter values and the corresponding channel quality variance Var(A(t)). Process the lock power control parameter values using the Z-score standardization method, denoted as the input vector O(t). Construct a dataset. Take the input vector as the input, and divide the training set and the test set in a ratio of 8:2. Use the training set to train a random forest model. When constructing the decision tree, use the bootstrap sampling method and randomly select features. When splitting nodes, calculate the reduction in channel quality variance for different feature splits. After constructing the entire tree, accumulate the variance reduction of each feature to obtain the contribution measure. Calculate the mean squared error function MSE = 1 / |Dtest|×∑(O(t), Var(A(t)))∈Dtest(Var(A_pred(t)) - Var(A(t))) 2 , where |Dtest| represents the number of samples in the test set Dtest, Var(A_pred(t)) is the predicted value. If the mean squared error MSE exceeds the threshold, then adjust the depth of the tree until the mean squared error value is less than the set threshold and end. Substitute the real-time input lock power control parameters into the trained model, normalize the contribution measure to obtain the real-time contribution value fi(t) of each factor to the channel quality variance. Through the weight adjustment rule wi(t + 1) = wi(t) + η0×(Var_target - Var(A(t))), where i represents the serial number of the lock power control parameter, taking values of 1, 2, 3, 4, η0 represents the learning rate, controlling the step size of weight adjustment, and Var_target represents the standard channel quality variance; Adjust the power response speed of the transmit power adjustment model through the fractional-order adaptive control function. The fractional-order adaptive control function includes: D ν P_tx(t) = κ×(P_pred - P_tx(t)) + ϵ0×D u ×W(t), where D ν , D u represents the Caputo fractional derivative, used to describe the non-integer-order dynamic characteristics of power changes. P_pred represents the target transmit power, κ represents the convergence rate, determining the speed at which the power adjusts to the target value, ϵ0 represents the environmental disturbance intensity, reflecting the influence degree of external interference on power adjustment, and W(t) is a noise term simulating burst interference, making the model more suitable for the actual complex environment. And continuously monitor the absolute error e_p(t) = |P_pred - P_tx(t)| between the target power and the actual power. If the absolute error e_p(t) is greater than the set threshold e_th for a duration greater than t_fau, then it is determined that the power adjustment fails, stop the power adjustment, and return to step S100 for frequency band calculation and switching.

[0028] Specifically, the analysis of different measures in step S500 is as follows: Obtain the number of frequency band switches of the door lock in the above steps per unit time, and mark it as the frequency value of frequency band switching. If the frequency value of frequency band switching is less than the threshold value tv1, no corresponding operation is performed. If the frequency value of frequency band switching is between the threshold value tv1 and the threshold value tv2 (including the values at both endpoints), then divide the threshold kl set for the channel quality variance in step S100, the deviation tolerance jd1 allowed in step S300, and the threshold eth set for the absolute error ep(t) in step S400 by the correction coefficient θ1, where θ1 is a value greater than 1, specifically 1.2, to reduce the sensitivity threshold of the system trigger condition and make the detection mechanisms in each link more stringent, so as to identify potential problems and optimize the switching strategy. If the frequency value of frequency band switching is greater than the threshold value tv2, issue a warning and push an alarm message containing specific abnormal data to the administrator terminal to prompt manual intervention to investigate the root cause of high-frequency switching.

[0029] Specifically, through a multi-layer technical architecture of fractional-order modeling → fuzzy logic analysis → topological boundary recognition → adaptive power control → switching frequency closed-loop, it breaks through the limitations of a single model and fixed parameters in traditional wireless communication systems, supports the robustness of the system in practical scenarios such as channel fluctuations, multi-device interference, and power limitations, provides a reusable methodology for the wireless communication optimization of Internet of Things devices, and adopts a closed-loop mechanism of trigger signal series connection, status information sharing, and parameter dynamic correction to integrate independent channel modeling, link control, and power adjustment modules into an organic whole.

[0030] As Figure 3 shown, a wireless communication method device for an intelligent door lock includes a handover decision module, a security matching module, a link control module, a power control adjustment module, and a response module. The handover decision module constructs a fractional-order stochastic channel evolution model, analyzes the model using the Mittag-Leffler function, generates sample paths using Monte Carlo simulation, obtains the probability distribution of channel quality, and obtains the channel quality variance through the variance function. If it exceeds the threshold, it issues a frequency band switching signal and selects the highest scoring frequency band for switching using the frequency band priority scoring function. The security matching module calculates the frequency band priority score using the frequency band priority scoring function for both after frequency band switching and when no frequency band switching signal is issued, and matches the corresponding encryption measures. If the frequency band priority score is within the set comparison interval ll3, it issues a warning and issues a frequency band switching signal to return to the handover decision module for frequency band calculation and switching. The link control module obtains the eight-dimensional state vector of the link, calculates and analyzes it using the Gaussian fuzzy membership function to obtain the fuzzy-precise conversion index, limits the parameters in the safe area using the dynamic barrier function, fuses them through the hybrid constraint function, maps the eight-dimensional state parameters of the link into a simplicial complex, retains the simplices with diameters not exceeding the threshold for analysis to form the topological boundary of the stable state, maps the topological boundary of the stable state to the sliding mode surface through the sliding mode control function to obtain the sliding mode surface function value of the link. If the absolute value of the sliding mode surface function value is greater than the allowable deviation tolerance, it returns to the handover decision module for frequency band calculation and handover; The power control adjustment module obtains the door lock power control parameters, establishes a transmission power adjustment model, adjusts the power response speed of the transmission power adjustment model through the fractional-order adaptive control function, monitors the absolute power error during power adjustment, and returns the corresponding status to the decision module for frequency band calculation and handover; The response module obtains the frequency values of the frequency band handover, and different frequency values correspond to different measures.

[0031] The above is a description of the present invention and should not be regarded as a limitation thereof. Although several exemplary embodiments of the present invention have been described, those skilled in the art will readily understand that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of the present invention. Therefore, all such modifications are intended to be included within the scope of the present invention as defined by the claims. It should be understood that the above is a description of the present invention and should not be considered limited to the specific embodiments disclosed, and modifications to the disclosed embodiments as well as other embodiments are intended to be included within the scope of the appended claims. The present invention is defined by the claims and their equivalents.

Claims

1. A wireless communication method for an intelligent door lock, characterized in that, It includes the following steps: Step S100: Construct a fractional-order stochastic channel evolution model, analyze the model using the Mittag-Leffler function, generate sample paths by Monte Carlo simulation, obtain the channel quality variance through the variance function, if it exceeds the threshold, send a frequency band switching signal, and select the highest scoring frequency band for switching; Step S200: Calculate the frequency band priority score for the frequency band after switching and the case where no frequency band switching signal is sent using the frequency band priority scoring function, match the corresponding encryption measures, issue a warning for the corresponding frequency band priority score, and send a frequency band switching signal to return to Step S100 for frequency band calculation and switching; Step S300: Obtain the eight-dimensional state vector of the link, calculate and analyze it using the Gaussian-type fuzzy membership function to obtain the fuzzy exact conversion index, use the dynamic barrier function to limit the parameters in the safe area, fuse through the mixed constraint function, map the eight-dimensional state parameters of the link into a simplex, retain the simplex with a diameter not exceeding the threshold, analyze the formed stable state topological boundary, map the stable state topological boundary to the sliding mode surface through the sliding mode control function to obtain the sliding mode surface function value of the link, if the absolute value of the sliding mode surface function value is greater than the allowed deviation tolerance, return to Step S100 for frequency band calculation and switching; Step S400: Obtain the door lock power control parameter, establish a transmit power adjustment model, and adjust the power response speed of the transmit power adjustment model through the fractional-order adaptive control function, monitor the absolute power error during power adjustment, and return to Step S100 for frequency band calculation and switching for the corresponding state; Step S500: Obtain the frequency values of the frequency band switching in the above steps, and different frequency values correspond to different measures.

2. The wireless communication method of an intelligent door lock according to claim 1, characterized in that, The analysis step of selecting the highest scoring frequency band for switching: Construct a fractional-order stochastic channel evolution model to obtain the variation law of the door lock channel quality over time. The fractional-order stochastic channel evolution model includes: D ɑ A(t)=γ×A(t)+σ×D ζ W(t), where D ɑ and D ζ are the Caputo-type fractional-order derivatives respectively. A(t) represents the channel quality at time t, γ represents the state decay coefficient, σ is the noise intensity, and W(t) is the standard Brownian motion; Analyze the fractional-order stochastic channel evolution model through the Mittag-Leffler function to obtain the probability distribution of the channel quality, where the Mittag-Leffler function: , where E ɑ，β (k) is the Mittag-Leffler function, h is the summation index, ɑ and β are function parameters, k is the input variable, Γ(*) is the gamma function. Substituting it into the fractional-order stochastic channel evolution model for analysis, the form of the model solution is obtained as follows: A(t)=A(0)×E ɑ，1 (γ×t ɑ )+σ×∫(t - s) ɑ-1 E ɑ，ɑ (γ×(t - s) ɑ )D ζ W(s)ds, where A(0) represents the channel quality at the initial moment, s is the integration variable, and the values of parameters ɑ, ζ, γ, and σ are estimated using the fractional least squares method based on the sliding window data. By minimizing the objective function , where represents the squared norm of the vector, measuring the degree of difference, i is the summation index variable, Φ1 is the regularization parameter, the parameter ζ is implicit in the calculation of A(t), and the parameters ɑ, γ, and σ jointly participate in the optimization adjustment of the objective function. By iterative solution, the objective function reaches the minimum value, and the values of ɑ, ζ, γ, and σ are obtained. A large number of sample paths are generated using Monte Carlo simulation, and the distribution of A(t) at different time points is statistically analyzed , where P(A, t) is the probability density function of the channel quality being A at time t, δ(*)is the Dirac function. When A = , then δ(A - = 1, otherwise it is equal to 0; The channel quality variance Var(A(t)) is obtained through the variance function: Var(A(t)) = ∫A 2 ×P(A, t)dA - u(t) 2 , where A represents the value of the channel quality, u(t) is the mean of the channel quality A(t). If the channel quality variance is greater than the set threshold kl, a frequency band switching signal is sent, and no other operations are performed. The fractional-order stochastic channel evolution model is analyzed for frequency band stability through the frequency band priority scoring function, and the frequency band with the highest channel selection score is obtained and switched. The frequency band priority scoring function is as follows: , where f j is the serial number of the frequency band, A hist (s) is the historical quality data of the frequency band, T is the time window, S(f j ) frequency band is the priority score.

3. A wireless communication method for an intelligent door lock according to claim 1, characterized in that The analysis step of matching the corresponding encryption measures: Calculate the frequency band priority score for the frequency band after switching and the case where no frequency band switching signal is sent using the frequency band priority scoring function in Step S100. If the frequency band priority score is within the set comparison interval ll1, the corresponding encryption measure one is used. If the frequency band priority score is within the set comparison interval ll2, the corresponding encryption measure two is used. If the frequency band priority score is within the set comparison interval ll3, issue a warning and send a frequency band switching signal, and perform the calculation of the channel frequency band with the highest priority score in Step S100 and switch.

4. A wireless communication method for an intelligent door lock according to claim 1, characterized in that, The analysis step of the sliding mode surface function value of the link: Map the eight-dimensional state parameter space of the link into a simplex, set the distance threshold tr1, only retain the simplex with a diameter not exceeding the threshold tr1, merge all the simplexes that meet the conditions to form a connected region, and the union of the connected regions forms the topological boundary of the stable state in the link parameter space; Map the obtained topological boundary of the stable state and the mixed constraint function B(x) to the sliding mode surface through the sliding mode control function, and the sliding mode control function: , where d i represents the sliding mode coefficient, which is used to adjust the influence weight of each parameter dimension on the sliding mode surface. x i，ref represents the reference stable value, χ represents the control switching speed, sign(*) is the sign function. When B(x) exceeds the threshold B th forces the state trajectory to converge along the sliding mode surface. f(x) represents the sliding mode surface function value of the link. If |f(x)| > jd1, a frequency band switching signal is triggered, where jd1 is the allowable deviation tolerance.

5. The wireless communication method of an intelligent door lock according to claim 4, characterized in that, The analysis step of the mixed constraint function B(x): Obtain the bit error rate, delay, number of retransmissions, signal-to-noise ratio, delay jitter, spectrum occupancy rate, phase noise, and Doppler frequency shift of the multi-dimensional link, mark them as the state parameters of the multi-dimensional link, and construct an eight-dimensional state vector: x = [x1, x2, …, x i T , where T represents the transpose of the matrix, i represents the serial number of the eight-dimensional state vector, design a Gaussian-type fuzzy membership function for each parameter, and the membership degree u 低BE (x i ) = exp(-(x i - z low ) 2 ) / (2η 2 ), where exp represents the exponential function with the natural constant e as the base, z low represents the ideal low bit error rate central value, η is the standard deviation. If the stability levels are unstable, medium, and stable, the corresponding central values are 0.25, 0.65, and 0.95 respectively. Obtain the membership degrees of condition 1 and condition 2 in the stability fuzzy rules of each multi-dimensional link, and take their minimum value as the activation intensity;​ Calculate the activation intensity matrix h of the nth rule in all samples mn , where m is the sample index, and calculate the entropy value of each rule , according to the formula , obtain the weight w of each rule n , where M is the number of samples and N is the number of rules. Using the center method formula, obtain the fuzzy-precise conversion index W(x); Through the dynamic barrier function G(x), the state parameters of the multi-dimensional link are forced within the safe area. The specific dynamic barrier function: , where x i，max and x i，min Represents the maximum and minimum values ​​allowed by the state parameters of the multidimensional link. ε is a very small constant to prevent the denominator from being zero. The dynamic barrier function G(x) is combined with the fuzzy precise conversion index W(x) to obtain the hybrid constraint function B(x)=G(x)×(1+Ω×W(x)), where Ω is the fuzzy weight coefficient.

6. The wireless communication method of an intelligent door lock according to claim 5, characterized in that, The steps for analyzing the stability fuzzy rules of the multi-dimensional link: Perform data standardization processing on the state parameters of the multi-dimensional link as input parameters, and use the link stability level as the output target. Train a decision tree using the CART algorithm, with the information gain ratio as the splitting criterion. Convert each path from the root to the leaf in the trained decision tree into a clear rule in the form of "IF-THEN". Define fuzzy sets for each input parameter and, based on the above Gaussian fuzzy membership function, map the thresholds in the decision tree to the intersection points of the fuzzy subsets. Replace the conditions of the clear rules with combinations of fuzzy subsets to obtain fuzzy rules. Calculate the membership degrees of the real-time input parameters in each fuzzy subset, calculate the activation strength of each fuzzy rule according to the logical relationship of the rule premise conditions, and finally select the stability level corresponding to the rule with the maximum activation strength as the output. If the strengths of multiple rules are the same, take the average level to obtain the stability fuzzy rules of the multi-dimensional link.

7. The wireless communication method of an intelligent door lock according to claim 1, characterized in that The steps for analyzing the absolute power error when monitoring and adjusting the power: Obtain the user activity patterns of the door lock, select the operation frequency, the standard deviation of the operation time interval, and the night operation in the user activity patterns to construct a three-dimensional input space. Perform standardization processing on each index, adopt the fuzzy K-means clustering algorithm, and through iterative optimization of the objective function, obtain the membership degree vectors of each sample for the three types of activity patterns of low, medium, and high. Calculate the three-dimensional feature weights through the entropy weight method, calculate the entropy values of each feature, and convert them into weights wj, where j represents the serial number of the index. Calculate through the weighted centroid method to obtain the user activity clustering value y; And obtain the network congestion index, device movement speed, and remaining battery percentage of the door lock, and form the door lock power control parameters together with the user activity clustering value y. Establish a transmit power adjustment model, specifically including: Ptx(t)=Pbas×(ω1(t) / (1 + Nconge)+ω2(t)×Bcuf + ω3(t)×vrove / vstand + ω4(t)×y), where Ptx(t) represents the transmit power adjustment value at time t, Pbas represents the reference power, ω1, ω2, w3, and ω4 represent the weights of the door lock power control parameters, and the weights are adjusted in real time through reinforcement learning. Nconge represents the network congestion index, Bcuf represents the remaining battery percentage, vrove represents the device movement speed, and vstand represents the device movement standard speed; Adjust the power response speed of the transmit power adjustment model through the fractional-order adaptive control function. The fractional-order adaptive control function includes: D ν Ptx(t)=κ×(Ppred - Ptx(t)) + ϵ0×D u ×W(t), where D ν and D u represents the Caputo fractional-order derivative, Ppred represents the target transmit power, κ represents the convergence rate, ϵ0 represents the environmental disturbance intensity, W(t) is the noise term simulating burst interference. The absolute error between the target power and the actual power is monitored in real time, ep(t)=|Ppred - Ptx(t)|. If the absolute error ep(t) is greater than the set threshold eth for a duration greater than tfau, it is determined that the power adjustment fails, the power adjustment is stopped, and then return to step S100 for frequency band calculation and switching.

8. The wireless communication method of an intelligent door lock according to claim 7, characterized in that, The steps for analyzing the real-time weight adjustment: Collect the historical data of the lock power control parameter values and the corresponding channel quality variance Var(A(t)), process the lock power control parameter values using the Z-score normalization method, denoted as the input vector O(t), construct a data set, use the input vector as the input, divide the training set and the test set in a ratio of 8:2, train a random forest model with the training set, use the bootstrap sampling method and randomly select features when constructing the decision tree, calculate the reduction in channel quality variance for different feature splits when splitting nodes, and after the entire tree is constructed, accumulate the variance reduction of each feature to obtain the contribution measure. Calculate the mean squared error function MSE = 1 / |Dtest| × ∑(O(t), Var(A(t))) ∈ Dtest (Var(A_pred(t)) - Var(A(t))) 2 , where |Dtest| represents the number of samples in the test set Dtest, Var(A_pred(t)) is the predicted value. If the mean squared error MSE exceeds the threshold, then adjust the depth of the tree until the mean squared error value is less than the set threshold and end. Substitute the real-time input lock power control parameters into the trained model, normalize the contribution measure to obtain the real-time contribution value fi(t) of each factor to the channel quality variance, and adjust the weight in real time through the weight adjustment rule wi(t + 1) = wi(t) + η0 × (Var_target - Var(A(t))), where i represents the serial number of the lock power control parameter, η0 represents the learning rate, and Var_target represents the standard target channel quality variance 9. The wireless communication method of an intelligent door lock according to claim 1, characterized in that, The steps for analyzing different measures corresponding to different frequency values: Obtain the number of frequency band switches of the door lock per unit time in the above steps, and mark it as the frequency value of frequency band switching. If the frequency value of frequency band switching is between the threshold values tv1 and tv2, then divide the threshold kl set for the channel quality variance in step S100, the deviation tolerance jd1 allowed in step S300, and the threshold eth set for the absolute error ep(t) in step S400 by the correction coefficient θ1. If the frequency value of frequency band switching is greater than the threshold value tv2, then issue a warning and push an alarm message containing specific abnormal data to the administrator terminal.

10. A wireless communication device for an intelligent door lock, characterized in that A wireless communication method for an intelligent door lock implementing any one of claims 1-9, including a handover decision module, a security matching module, a link control module, a power control adjustment module, and a response module; The handover decision module constructs a fractional-order stochastic channel evolution model, analyzes the model using the Mittag-Leffler function, generates sample paths using Monte Carlo simulation to obtain the probability distribution of channel quality, obtains the channel quality variance through a variance function, issues a frequency band switching signal if it exceeds the threshold, and selects the highest scoring frequency band for switching using a frequency band priority scoring function; The security matching module calculates the frequency band priority score using the frequency band priority scoring function for both after the frequency band switch and when no frequency band switching signal is issued, and matches the corresponding encryption measures. If the frequency band priority score is within the set comparison interval, then issue a warning and send the frequency band switching signal back to the handover decision module for frequency band calculation and switching; The link control module obtains the eight-dimensional state vector of the link, calculates and analyzes it using a Gaussian-type fuzzy membership function to obtain a fuzzy exact conversion index, limits the parameters in the safe area using a dynamic barrier function, fuses them through a mixed constraint function, maps the eight-dimensional state parameters of the link into a simplex, retains the simplex with a diameter not exceeding the threshold for analysis to form the topological boundary of the stable state, maps the topological boundary of the stable state to a sliding mode surface through a sliding mode control function to obtain the sliding mode surface function value of the link. If the absolute value of the sliding mode surface function value is greater than the allowed deviation tolerance, then return to the handover decision module for frequency band calculation and switching; The power control adjustment module obtains the door lock power control parameters, establishes a transmit power adjustment model, adjusts the power response speed of the transmit power adjustment model through a fractional-order adaptive control function, monitors the absolute power error during power adjustment, and returns the corresponding state to the decision module for frequency band calculation and switching; The response module obtains the frequency value of the frequency band switch, and different frequency values correspond to different measures.

Images