Communication control method for smart energy unit security access module

Abstract

The invention discloses a communication control method for a smart energy unit security access module. The communication control method comprises the following steps: constructing a wireless network fingerprint containing a main base station ID, signal strength and adjacent cell information; environment segmentation judgment is carried out based on network topology change, and accurate identification and classification of a network environment are realized; a multi-stage monitoring strategy and a differentiated validity period management mechanism are adopted, and the monitoring strategy is dynamically adjusted according to the environmental stability, the data priority and the energy state; a small sample fast adaptation algorithm is utilized, and communication parameters in a new environment are quickly optimized through similar environment parameter migration. The problem of communication failure caused by sudden change of network topology in a complex electromagnetic environment is solved, the communication success rate is improved, and energy consumption is reduced.

Description

Technical Field

[0001] The present invention belongs to the field of smart energy management and Internet of Things applications, and in particular to a communication control method for a smart energy unit secure access module. Background Art

[0002] With the rapid development of smart grids and distributed energy systems, smart energy units, as data collection and control terminals at the end of power systems, play a vital role in energy management, load control, and fault monitoring. These devices typically exchange data with cloud platforms via 4G networks to ensure real-time monitoring and intelligent scheduling of power systems. However, due to the complexity and variability of deployment environments, such as electromagnetic interference generated by power equipment and signal fluctuations in remote locations, smart energy units face severe challenges in communication efficiency and reliability. Efficient and reliable communication control methods are urgently needed to ensure the real-time and integrity of data transmission.

[0003] Current research on communication control for smart energy units focuses primarily on simple retry mechanisms and basic signal strength monitoring. Typical approaches include: fixed-interval retry strategies, which repeat attempts at fixed intervals after a communication failure; simple signal strength threshold methods, which determine whether to transmit data based solely on the RSSI value; historical statistical time window selection, which analyzes periods of the day with good signal strength for centralized transmission; and simple data caching mechanisms, which store data in a local cache for subsequent transmission when communication fails. Additionally, some studies have employed fixed-period network status monitoring and simple linear prediction models to estimate future signal conditions. However, these methods are mostly based on a single signal strength metric and lack comprehensive awareness of complex network environments.

[0004] However, existing technologies have obvious shortcomings when facing the complex network environments unique to smart energy units. First, the existing network environment modeling methods use continuous time series models, which cannot effectively deal with sudden changes in network topology caused by base station switching, the activation of electromagnetic interference sources, etc., making the prediction model completely ineffective when the environment suddenly changes. Secondly, in new environments or data-sparse conditions, traditional methods require a large number of samples to establish an accurate communication parameter model, causing smart energy units to undergo a long parameter adjustment process after environmental changes, significantly reducing the communication success rate. These specific problems are particularly prominent in harsh electromagnetic environments and energy-constrained scenarios, seriously restricting the application effect of smart energy units in complex environments. Summary of the Invention

[0005] The purpose of the invention is to provide a communication control method for a smart energy unit secure access module, in order to solve at least one technical problem existing in the prior art.

[0006] Technical solution: A communication control method for a smart energy unit secure access module, comprising: Obtain the currently connected base station information and adjacent cell information, and construct the network environment fingerprint and similarity function; Based on the similarity function and the current and historical network environment fingerprints, the degree of environmental change is calculated, the type of environmental change is determined, and environmental parameters are generated. This is combined with the priority of the data to be transmitted to determine the monitoring level, obtain a multi-dimensional monitoring indicator set, and calculate the comprehensive signal quality index. Based on environmental parameters, comprehensive signal quality index and priority of data to be transmitted, the transmission decision score is calculated, the transmission decision result is generated and the transmission is executed to obtain the transmission execution result.

[0007] Beneficial effects: The present invention can distinguish between noise changes caused by signal fluctuations and topology changes caused by actual position movement, realize accurate quantification and classification of network environment changes, and improve communication reliability and stability; realize rapid adaptation to new environments under conditions of very little or even zero sample data, improve communication success rate and reduce energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A flowchart of the steps of a communication control method for a smart energy unit secure access module provided in an embodiment of the present application.

[0009] Figure 2 A flowchart of the steps for constructing a network environment fingerprint and similarity function provided in an embodiment of the present application.

[0010] Figure 3 A flowchart of the steps for generating environmental parameters provided in an embodiment of the present application.

[0011] Figure 4 A flowchart of the steps for calculating the comprehensive signal quality index provided in an embodiment of the present application. DETAILED DESCRIPTION

[0012] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0013] It should be noted that to clearly illustrate the steps of this application, serial numbers are assigned to each step in the specification. These serial numbers are for illustrative purposes only and do not limit the order in which the steps must be executed. In actual operation, depending on the technical requirements of the specific implementation scenario, the steps may be executed in a different order than shown in the specification, and in some cases, parallel processing between steps may be implemented.

[0014] During the research process, it was found that the existing indicator monitoring strategy adopts full-data collection in a fixed period, which consumes a large amount of precious electricity and communication resources in remote deployment scenarios with limited energy. Simply extending the monitoring interval will lead to indicator expiration and decision-making errors, and it is impossible to achieve an effective balance between resource consumption and communication reliability.

[0015] like Figure 1 As shown, a communication control method for a smart energy unit secure access module is proposed, comprising the following steps: S1. Obtain the information of the currently connected base station and adjacent cells, and construct a network environment fingerprint and similarity function; Specifically, the currently connected base station information includes the base station ID, signal strength (RSSI), and signal quality (SINR). In addition to connecting to the current base station, the device also detects other nearby base stations, thereby collecting information about neighboring cells. This information is combined with the current base station and neighboring cells to form a network environment fingerprint, similar to the characteristic identifier of a geographic location. This means that different locations have different network environment fingerprints. These different network fingerprints are compared and a similarity function is calculated.

[0016] S2. Based on the similarity function and the current and historical network environment fingerprints, calculate the degree of environmental change, determine the type of environmental change, and generate environmental parameters; Specifically, a similarity function is used to compare the current network environment fingerprint with historical fingerprints to observe the differences between the two. If the difference is small, the environment is relatively stable; if the difference is large, it indicates that the network environment has changed significantly. Environmental parameters include the current environment ID and the environment stability score. The current environment ID is like a label for the current network environment, which helps identify it. The environment stability score reflects the reliability of the current environment. For example, high stability indicates a strong signal with little fluctuation, suitable for long-term connections; low stability indicates significant signal fluctuations, which may require adjustment of the connection strategy.

[0017] S3. Combine environmental parameters with the priority of the data to be transmitted to determine the monitoring level, obtain a multi-dimensional monitoring indicator set, and calculate a comprehensive signal quality index; Specifically, the system determines how much network monitoring is needed based on the stability of the environment and the priority of the data. For example, in a stable network environment, the monitoring level can be lower to reduce resource consumption; in a complex or fluctuating environment, the monitoring level can be increased to obtain more network quality data.

[0018] S4. Based on the environmental parameters, the comprehensive signal quality index and the priority of the data to be transmitted, a transmission decision score is calculated, a transmission decision result is generated, and the transmission is executed to obtain a transmission execution result.

[0019] Specifically, the transmission decision score is used to evaluate whether the current network is suitable for transmitting specific data. Based on the calculated score, the system decides: whether to transmit data immediately (such as high-priority data); whether to optimize the transmission method (select different encoding methods, reduce the rate, etc.); and whether to delay transmission (if the environment is unstable, wait for better conditions).

[0020] like Figure 2 As shown, according to one aspect of the present application, constructing a network environment fingerprint and a similarity function includes: S11. Obtain the ID, signal strength, and quality of the currently connected base station to form a basic network parameter set. Unlike traditional methods, this embodiment not only collects signal strength values but also records base station identification information, laying the foundation for subsequent environmental change detection. S12. Reading a neighboring network parameter set, including a list of neighboring cells perceivable by the device and a list of their corresponding signal strengths; using this information as a supplementary feature to distinguish different network environments. This embodiment goes beyond the limitations of traditional methods that focus only on connected base stations and introduces a network topology perspective; S13. Construct a network environment fingerprint NF based on the basic network parameter set and the neighboring network parameter set; NF = {primaryID, RSSI, SINR, N_list, S_list, timestamp}, where primaryID is the primary base station ID, timestamp is the timestamp, N_list is the neighboring cell list, and S_list is the signal strength list. This embodiment differs from traditional single signal strength records and can accurately identify different network environments, even if the signal strength values are similar. S14. Based on the primary base station ID matching, the cosine similarity of the neighboring cell lists, and the signal strength difference, a similarity function is established to evaluate the similarity between network environment fingerprints. The similarity function is: SimFunc(NF1, NF2) = w1 I(NF1.primaryID == NF2.primaryID) + w2 CosSim(NF1.N_list, NF2.N_list) + w3 (1 - |NF1.RSSI - NF2.RSSI| / MaxRSSDiff); where I() is the indicator function, CosSim() is the cosine similarity function, and MaxRSSDiff is the maximum signal strength difference. w1, w2, and w3 are weight coefficients. This allows for accurate identification of network environment changes.

[0021] According to one aspect of the present application, constructing a network environment fingerprint includes: The basic network parameter set and the adjacent network parameter set are read to construct an original network dataset. Key features are extracted from the dataset and stability is analyzed to obtain feature reliability weights. Based on the key features and feature reliability weights, a structured network environment fingerprint is constructed, compressed, and optimized to generate an optimized fingerprint. Calculate the similarity between the optimized fingerprint and the historical network environment fingerprint within a predetermined period to verify the recognition effectiveness of the fingerprint; The optimized fingerprint that passes the verification is stored in the historical fingerprint set as the final network environment fingerprint.

[0022] Specifically, based on the basic network parameter set P_base, the neighboring cell list N_list, and the signal strength list S_list, raw network environment data is collected: the primary base station ID (primaryID), signal strength RSSI, and signal quality SINR are extracted from the basic network parameter set; each cell's cell ID (cellID) and corresponding signal strength (cellRSSI) are extracted from the neighboring cell list and signal strength list; the current timestamp (timestamp) is recorded; and a raw network dataset containing all of the above information is formed. This ensures that complete network environment data is captured, providing a foundation for subsequent fingerprint construction.

[0023] Perform feature selection and extraction. Select key features from the original network data set and extract them: retain the primary base station ID primaryID as the main identification feature Primary_feature; calculate the normalized value RSSI_norm of the primary base station signal strength: RSSI_norm = (RSSI - Min_RSSI) / (Max_RSSI - Min_RSSI), where Min_RSSI and Max_RSSI are the preset signal strength range boundaries; extract the adjacent cell topology feature Topo_feature: sort the adjacent cell list N_list in descending order of signal strength to form a sorted cell list Sorted_cells, retain the N cells with the strongest signals (default N=5) as the core topology feature Core_topo, calculate the relative signal strength Relative_RSSI of each retained cell: Relative_RSSI = cellRSSI / RSSI; construct the spectrum occupancy feature Spectrum_feature: count the number of cells detected on each frequency band, form the frequency band distribution vector Freq_distribution, and calculate the occupancy rate of the main frequency band. This embodiment is different from the traditional single signal strength record and provides a more comprehensive representation of network environment features.

[0024] Perform feature stability analysis. Read historical network data and analyze the stability of the above features: calculate the coefficient of variation of the RSSI of the main base station within a certain time window: RSSI_CV = StandardDeviation(History_RSSI) / Mean(History_RSSI); calculate the change rate Cell_change_rate of the adjacent cell list N_list within the time window: Cell_change_rate = Count(Cell_changes) / Window_size; based on the above stability indicators, calculate the reliability weight of each feature: RSSI_weight = 1 / (1 + RSSI_CV); Topo_weight = 1 / (1 + Cell_change_rate). Where StandardDeviation is the standard deviation function, History_RSSI is the historical signal strength, Mean is the mean function, Count is the count function, Cell_changes is the number of cell switches, Window_size is the window size, RSSI_weight is the signal strength weight, and Topo_weight is the topology weight. This embodiment enables the system to downgrade unstable features and improve the reliability of the fingerprint.

[0025] Fingerprint vectorization is performed. The extracted multidimensional features are converted into a structured fingerprint vector: the basic fingerprint structure is created, including: primaryID (primary base station ID); RSSI (signal strength); SINR (signal quality); N_list (neighboring cell list [{cellID, cellRSSI}]); and timestamp (timestamp). Derived features are added: RSSI_stability (signal stability indicator); Topo_complexity (topological complexity indicator); and Features_weight (reliability weight of each feature). This complete network environment fingerprint (NF) is assembled. This retains the key information of the original network data while adding derived features and reliability information, providing a rich basis for subsequent similarity calculations.

[0026] Fingerprint compression and optimization are performed. The network environment fingerprint is compressed and optimized to improve storage and computational efficiency. The neighboring cell list is compressed, removing cells with signal strength below the threshold Threshold_RSSI. The remaining cells are grouped by signal strength, with cells in the same group sharing a common representation. Numerical features are quantized, discretizing the signal strength RSSI and signal quality SINR into pre-set intervals. The relative signal strength Relative_RSSI is quantized to a finite-precision value. This generates an optimized fingerprint containing the same information as the network environment fingerprint, but with a more compact data structure. This reduces storage overhead and similarity calculation complexity while maintaining recognition accuracy.

[0027] Perform fingerprint validity verification. Verify the recognition effectiveness of the constructed optimized fingerprint, ensure the robustness of the fingerprint modeling process, and prevent misjudgment caused by abnormal data. Perform fingerprint storage and indexing. Store and index the optimized fingerprint that has passed the verification: add the optimized fingerprint to the short-term historical fingerprint set; establish a quick search index Quick_lookup_index based on the primary base station ID primaryID; regularly clean up expired fingerprints in the short-term historical fingerprint set; output the optimized fingerprint to the environment change detection module and the current environment discrimination module for use. The persistence and index optimization of the fingerprint are completed, providing efficient query support for subsequent environmental discrimination. This embodiment fully realizes the construction process of the wireless network fingerprint, and provides an accurate and reliable basis for environmental identification and segmentation through the extraction and fusion of multi-dimensional features, feature stability analysis and weight adjustment, structured fingerprint representation and optimization, etc.

[0028] According to one aspect of the present application, establishing a similarity function includes: Set the initial weight coefficient for fingerprint similarity calculation and verify its effectiveness; establish a dynamic weight adjustment mechanism to update the weight after each environmental identification result verification; based on the verified weight coefficient, construct a fingerprint similarity calculation formula, comprehensively considering the main base station ID matching indicator function, the cosine similarity of the adjacent cell list, and the normalized difference in signal strength to calculate the similarity score of the two network environment fingerprints.

[0029] like Figure 3 As shown, according to one aspect of the present application, generating environmental parameters includes: S21. Compare the historical network environment fingerprints in the current period with those in the previous period, and calculate the degree of environment change that comprehensively considers signal strength changes, primary base station switching, and changes in neighboring cell topology; S22. When the degree of environmental change is greater than a preset threshold, the environment changes significantly, and a new environmental segment is created and assigned an ID; otherwise, the environment changes continuously, and the current environmental segment information is updated to form an updated network environment database; S23. Every predetermined period, use a similarity function to calculate the similarity between the current network environment fingerprint and each network environment fingerprint in the network environment library, and obtain the most similar environment ID and similarity score; S24: Based on the similarity score, determine whether the current environment is a known environment or a new environment, generate the current environment ID and environment stability score as environment parameters, and update the network environment database based on the result of the environment change type determination.

[0030] Specifically, the current network fingerprint is compared with the historical network fingerprint, and the degree of environmental change ΔE is calculated as α|RSSI_current - RSSI_history| + β(1-I(primaryID_current == primaryID_history)) + γ(1-|N_list_common| / |N_list_union|). Here, |N_list_common| is the intersection of the current and historical neighboring cell lists, |N_list_union| is the union, α, β, and γ are weighting coefficients, RSSI_current is the current signal strength, and RSSI_history is the historical signal strength. This method breaks through the limitation of traditional network change judgment based solely on signal strength and introduces changes in network topology. The degree of environmental change ΔE is compared with a preset threshold Eth. When ΔE > Eth, it is determined to be a significant environmental change, triggering the creation of a new environmental segment. Otherwise, it is determined to be a continuous environmental change and belongs to the current environmental segment. This solves the problem that traditional continuous time series models fail when the network environment suddenly changes. Maintain a network environment database, assign a unique environment ID (EnvID) to each identified network environment, and record its network fingerprint NF, duration, and corresponding communication parameter set. Unlike traditional methods that rely on a single model, this approach enables the coexistence of multiple models and intelligent switching.

[0031] The current network fingerprint NF_current is compared with the fingerprints of each environment stored in the environment library EnvLib to obtain the most similar environment ID and similarity score Sim_score. If the Sim_score exceeds the identification threshold Sth, the environment is identified as a known one; otherwise, it is identified as a new environment and assigned a new environment ID. This allows the system to quickly restore historical successful parameters and avoid repeated learning. Environmental change measurement and segmentation determination primarily focus on changes between consecutive measurement points in the temporal dimension. Specifically, they determine whether the current network environment has significantly changed from the previous one, necessitating the creation of a new environmental segment. This is a temporal comparison process. Current environment determination focuses on spatial similarity with known environments in the historical environment library. Specifically, they determine whether the current network environment is similar to a previously encountered environment, allowing its parameter settings to be reused. This is a pattern matching process.

[0032] According to one aspect of the present application, determining the type of environmental change further includes: Decompose the environmental change degree into signal strength change component, main base station change flag and cell topology change rate, and construct the environmental change feature vector; The environmental change characteristic vector is combined with the historical environmental stability index to evaluate the actual impact of environmental changes on communications, and the physical changes are converted into communication impact assessment to obtain the significance of environmental changes; Based on historical records of environmental changes and communication success rate changes, the environmental segmentation threshold is calculated. When the environment changes frequently, the threshold is increased, and when the communication success rate is low, the threshold is lowered. The significance of environmental changes is compared with that of environmental changes, and environmental changes are divided into three categories: major changes, minor changes, and noise. Whether to trigger environmental segmentation is determined according to the type of environmental change.

[0033] According to one aspect of the present application, calculating an environment segmentation threshold includes: Read the basic threshold of the system configuration as the initial calculation basis for the environmental segmentation threshold; Calculate the average communication success rate within a predetermined time window based on historical communication records and generate a success rate adjustment factor; Analyze historical environmental change records, calculate the frequency of environmental changes per unit time, and generate a change frequency adjustment factor; Taking into account the stability, importance and typical scene characteristics of the current environment, an environment-specific adjustment factor is generated; The final adaptive environment segmentation threshold is calculated by multiplying the basic threshold by (1 + change frequency adjustment factor), (1 - success rate adjustment factor) and the environment-specific adjustment factor.

[0034] Specifically, a multi-dimensional decomposition of environmental change characteristics is performed. The environmental change degree ΔE and its constituent raw features (current RSSI_current, historical RSSI_history, current primary base station ID primaryID_current, historical primary base station ID primaryID_history, current neighboring cell list N_list_current, and historical neighboring cell list N_list_history) are read and decomposed into three dimensional features: The signal strength change component ΔS is calculated as |RSSI_current - RSSI_history|; the primary base station change flag ΔP is determined as (primaryID_current != primaryID_history) ? 1 : 0; and the cell topology change rate ΔN is calculated as: N_list_common = Intersection(N_list_current, N_list_history); N_list_union = Union(N_list_current, N_list_history); ΔN = 1 - |N_list_common| / |N_list_union|. Unlike traditional methods that only consider changes in signal strength, this method explicitly quantifies changes in network topology. primaryID_current is the ID of the primary base station at the current moment, and primaryID_history is the ID of the historical primary base station.

[0035] Calculate historical environment stability. Read the historical environment record set History_records corresponding to the current environment ID and calculate the environment stability duration Stable_duration and environment change frequency Change_freq: Stable_duration = Current_timestamp - Last_segment_timestamp; Change_freq = Segment_count / Total_monitoring_time; where Last_segment_timestamp is the timestamp of the most recent environment segment, Segment_count is the total number of environment segments during the monitoring period, Current_timestamp is the current timestamp, and Total_monitoring_time is the total monitoring time. This embodiment is used for adaptive threshold adjustment of subsequent environment segments.

[0036] Assess the significance of environmental changes. Combining the three-dimensional characteristics of environmental change (ΔS, ΔP, ΔN) with historical stability indicators (Stable_duration, Change_freq), calculate the actual communication impact significance of environmental changes: Signal_impact = min(1, ΔS / Max_meaningful_RSSI_change); Topology impact = Topo_impact = w_p * ΔP + w_n * ΔN; Stability_factor = 1 - exp(-λ * Stable_duration); Change_significance = (Signal_impact + Topo_impact) * Stability_factor. w_p and w_n are the weighting coefficients for changes in the primary base station and cell topology, λ is the stability duration impact factor, and Max_meaningful_RSSI_change is the effective signal change threshold. Translate physical changes into communication impact assessments.

[0037] Calculate the adaptive threshold. Based on the historical environmental change records (History_changes) and the communication success rate change records (Success_rate_changes), the environmental segmentation threshold (Eth) is dynamically calculated: Base_threshold = Config_base_threshold; Adjustment = k1 * Change_freq + k2 * (1 - Avg_success_rate); Eth = Base_threshold * (1 - Adjustment) * Env_specific_factor. Config_base_threshold is the preset base threshold, k1 and k2 are adjustment factors, Avg_success_rate is the average communication success rate, Env_specific_factor is the environment-specific adjustment factor, Base_threshold is the base threshold, and Adjustment is the adjustment factor. Unlike traditional fixed threshold methods, this method can dynamically adjust segmentation sensitivity based on environmental characteristics and communication requirements.

[0038] Perform multi-level change classification. The environmental change significance (Change_significance) is compared with the adaptive threshold Eth and the secondary threshold Eth_minor to classify environmental changes into three categories: if Change_significance > Eth: Change_type = "Major_change" / / Major change, requiring the creation of a new environmental segment; if Change_significance > Eth_minor: Change_type = "Minor_change" / / Minor change, recorded but not segmented; else: Change_type = "Noise" / / Noise, ignored. This avoids over-response to changes in noise levels while maintaining sensitivity to significant changes.

[0039] Perform environmental segmentation decisions and state transitions. Based on the change type Change_type and the current environment state Env_state, perform environmental segmentation decisions: If Change_type is "Major_change": Generate a new environment segment identifier New_segment_ID; Use the current network fingerprint as the initial fingerprint of the new environment segment NF_init; Update the environment segment timestamp Segment_timestamp to the current time; Save the previous environment segment information Old_segment to the environment history record Environment_history; Convert the environment state Env_state to "New_segment". If Change_type is "Minor_change": Keep the current environment segment unchanged; Record minor changes to the environment change log Change_log; Update the environment change frequency Change_freq; Convert the environment state Env_state to "Fluctuating"; If Change_type is "Noise": Do not perform any segmentation operations; Keep the environment state Env_state to "Stable". This embodiment implements fine control of environmental segmentation, which is different from traditional fixed rule segmentation.

[0040] Perform segmentation event notifications and data synchronization. When an environment segmentation occurs (Change_type is "Major_change"), an environment segmentation event notification is generated and the following data is synchronized: the new environment segment identifier (New_segment_ID) is notified to the environment repository management module; the previous environment segment information (Old_segment) is archived to the environment history repository (History_repository); and a new environment parameter initialization request (Init_request) is triggered, passing the new environment segment identifier (New_segment_ID) and the environment transition context (Transition_context) to the parameter initialization module.

[0041] like Figure 4 As shown, according to one aspect of the present application, calculating the comprehensive signal quality index includes: S31. Determine the initial monitoring level based on the current environment ID and environment stability score in the environmental parameters; the monitoring levels include basic monitoring that only monitors signal strength, intermediate monitoring that monitors signal strength and signal quality, and comprehensive monitoring that monitors signal strength, signal quality, network latency, and packet loss rate.

[0042] S32. Calling a monitoring trigger function based on the initial monitoring level determines whether to trigger a higher level of monitoring based on the current signal strength, environmental stability score, and priority of the data to be transmitted, thereby obtaining the actual monitoring level to be implemented. Differentiated validity periods are set for monitoring indicators based on the signal variation coefficient and environmental change frequency. S33. Obtain a multi-dimensional monitoring indicator set according to the monitoring level actually performed, normalize and weight it, and obtain a comprehensive signal quality index.

[0043] Specifically, the applicable monitoring level is determined based on the current environment ID and the environment stability score Stab_score: Level 1 (basic monitoring): monitors only RSSI; Level 2 (intermediate monitoring): monitors RSSI and SINR; Level 3 (comprehensive monitoring): monitors RSSI, SINR, network latency, and packet loss rate. Traditional methods rigidly monitor all indicators, while this embodiment reduces monitoring overhead through a hierarchical monitoring strategy. A monitoring trigger function, TrigFunc, is constructed to dynamically determine whether to trigger a higher level of monitoring based on the current RSSI, the environment stability score Stab_score, and the importance of the data to be transmitted, Priority: TrigFunc(RSSI, Stab_score, Priority) = Level 1, if RSSI > Th_high; Level 2, if Th_low ≤ RSSI ≤ Th_high; Level 3, if RSSI < Th_low && (Priority > P_th || Stab_score < S_th). Th_high and Th_low are signal thresholds, P_th is the priority threshold, and S_th is the stability threshold. This embodiment is different from traditional fixed-period monitoring and achieves efficient resource utilization.

[0044] Different TTL validity periods are set for different monitoring indicators: TTL_RSSI = Base_TTL; TTL_SINR = Base_TTL × (1-k1 × Var_RSSI); TTL_NetParams = Base_TTL × (1-k2 × Var_RSSI) × (1-k3 × Env_Change_Freq). Base_TTL is the base validity period, Var_RSSI is the signal variation coefficient, Env_Change_Freq is the environmental change frequency, and k1, k2, and k3 are adjustment coefficients. This solves the problem of indicator expiration or excessive updates caused by traditional fixed validity periods. Based on the acquired multi-dimensional monitoring indicator set, the comprehensive signal quality index (SQI) is calculated as: w1 × RSSI_norm + w2 × SINR_norm - w3 × Delay_norm - w4 × PacketLoss_norm. Each indicator is normalized (labeled with _norm), and w1 to w4 are weighting coefficients. Different from traditional single-indicator evaluation, this embodiment implements a multi-dimensional comprehensive evaluation to more accurately reflect the actual communication quality.

[0045] According to one aspect of this application, differentiated validity periods are set for monitoring indicators, including: Based on the historical network environment fingerprint set corresponding to the current environment ID, the short-term and long-term coefficients of variation of signal strength, the frequency of network topology changes, and the frequency of master base station switching are analyzed to form an environment dynamic characteristic vector. Based on the environment dynamic characteristic vector, different basic validity periods are calculated for signal strength, signal quality, and network parameters respectively. The highest priority of the data to be transmitted and the current energy status of the device are read, and the basic validity period is adjusted to obtain the optimized validity period. Based on the transmission failure rate caused by the expiration of indicators in historical transmission records and the change of indicators within the validity period, the comprehensive failure impact coefficient is calculated, and the optimized validity period is further adjusted to obtain the final validity period.

[0046] According to one aspect of the present application, the steps of calculating the comprehensive failure impact coefficient and further adjusting the optimized validity period to obtain the final validity period include: From historical transmission records, statistics are collected on the transmission failure rates caused by indicator expiration and indicator changes within the validity period. When the transmission failure rate caused by the expiration of the indicator exceeds the preset threshold, the expiration adjustment factor is calculated, and its value is proportional to the degree of exceeding the threshold; when the transmission failure rate caused by the change of the indicator during the validity period exceeds the preset threshold, the change adjustment factor is calculated, and its value is proportional to the degree of exceeding the threshold; Use the balance coefficient to perform a weighted combination of the expiration and change adjustment factors. Take the maximum adjustment factor plus the balance coefficient multiplied by the minimum adjustment factor to obtain the comprehensive failure impact coefficient. The validity period adjusted based on the energy status is multiplied by (1-comprehensive failure impact coefficient) to calculate the final validity period, and a lower limit protection for the validity period adjustment is set to ensure that the validity period is not too short and causes excessive monitoring.

[0047] Specifically, based on the historical network fingerprint set (Historical_NFs) corresponding to the current environment ID, the dynamic characteristics of the environment are analyzed: the short-term coefficient of variation (Short_term_RSSI_CV) and the long-term coefficient of variation (Long_term_RSSI_CV) of the RSSI signal strength are calculated: Short_term_RSSI_CV = StandardDeviation(Recent_RSSI) / Mean(Recent_RSSI); Long_term_RSSI_CV = StandardDeviation(All_RSSI) / Mean(All_RSSI). The network topology change frequency (Topo_change_freq) is calculated: Topo_change_freq = Count(Cell_list_changes) / Monitoring_period; the primary base station switching frequency (Cell_switch_freq) is calculated: Cell_switch_freq = Count(Primary_cell_changes) / Monitoring_period. Based on the above indicators, an environmental dynamics vector, Env_dynamics, is constructed: Env_dynamics = {Short_term_RSSI_CV, Long_term_RSSI_CV, Topo_change_freq, Cell_switch_freq}. This embodiment provides the basis for precise management of indicator validity periods. Recent_RSSI is the RSSI data set for a short period of time, All_RSSI is the RSSI data set for the entire monitoring period, Cell_list_changes is the number of cell list changes, Monitoring_period is the monitoring period, and Primary_cell_changes is the number of primary base station switches.

[0048] A basic validity period calculation formula is established to calculate the basic validity period of different indicators: for RSSI (signal strength), Base_TTL_RSSI = Max_TTL_RSSI * (1 - k1 * Short_term_RSSI_CV); for SINR (signal quality), Base_TTL_SINR = Max_TTL_SINR * (1 - k2 * Short_term_RSSI_CV); and for Delay and Packet Loss, Base_TTL_NetParams = Max_TTL_NetParams * (1 - k3 * Short_term_RSSI_CV) * (1 - k4 * Topo_change_freq). Max_TTL_X represents the maximum validity period for each indicator type, and k1, k2, k3, and k4 represent influence coefficients. Unlike traditional fixed validity period methods, this method can adaptively adjust to changing environmental characteristics.

[0049] The validity period of the indicator is adjusted based on the priority of the data to be transmitted: the highest priority (Max_priority) of the currently transmitted data set (Pending_data) is read; the priority adjustment factor (Priority_factor) is calculated: Priority_factor = 1 - (Max_priority / Max_possible_priority) * Priority_impact_factor, where Priority_impact_factor represents the priority impact factor. The validity period is then adjusted using the Priority_factor: TTL_adjusted = Base_TTL * Priority_factor. This ensures that high-priority data receives more timely network status information, improving transmission success rates.

[0050] Read the device's current energy state (Energy_state) and adjust the validity period based on energy. Calculate the energy adjustment factor (Energy_factor) based on the remaining power (Remaining_energy): if Remaining_energy > High_energy_threshold: Energy_factor = 1.0 / / Energy is sufficient, no adjustment. elif Remaining_energy > Low_energy_threshold: Energy_factor = 1.0 + (Energy_extension_ratio - 1.0) *(1 - (Remaining_energy - Low_energy_threshold) / (High_energy_threshold - Low_energy_threshold)). Else: Energy_factor = Energy_extension_ratio / / Energy is insufficient, maximize the validity period. High_energy_threshold is the highest energy threshold, Low_energy_threshold is the lowest energy threshold, and Energy_extension_ratio is the energy extension ratio. Applying the energy adjustment factor, Energy_factor, further modifies the validity period: TTL_energy_adjusted = TTL_adjusted * Energy_factor. This can extend the indicator's validity period under energy-constrained conditions, reduce monitoring frequency, and achieve energy conservation.

[0051] Feedback adjustments are made based on historical indicator effectiveness evaluations: The system reads historical transmission records for the transmission failure rate (Expiry_failure_rate) caused by indicator expiration and the failure rate (Change_failure_rate) caused by indicator changes during the validity period. If the transmission failure rate (Expiry_failure_rate) exceeds the threshold (Threshold_expiry), the validity period is shortened: TTL_feedback_adjusted = TTL_energy_adjusted * (1 - Expiry_correction_factor). If the failure rate (Change_failure_rate) caused by indicator changes during the validity period exceeds the threshold (Threshold_change), the validity period is shortened: TTL_feedback_adjusted = TTL_energy_adjusted * (1 - Change_correction_factor). This achieves closed-loop optimization of validity period management, dynamically adjusting the validity period policy based on actual transmission results.

[0052] Build an abnormal event detection and validity period reset mechanism: Define a list of abnormal events that may cause sudden network status changes: primary base station ID change; sudden signal strength change (exceeding the preset change threshold Sudden_change_threshold); significant change in the neighboring cell list (exceeding the list update threshold Cell_list_update_threshold); device restart or movement detection. Monitor continuous changes in the network environment fingerprint NF and detect anomaly events: if |RSSI_current - RSSI_previous| > Sudden_change_threshold: Trigger_anomaly("RSSI_sudden_change"); if primaryID_current != primaryID_previous: Trigger_anomaly("Cell_switch") / / ... other abnormal event detection. When an anomaly is detected, the metric expiration date is reset: foreach Metric in Active_metrics: if Is_affected_by(Metric, Anomaly_event): Reset_TTL(Metric) / / Immediately expire the metric; Schedule_immediate_monitoring(Metric) / / Schedule immediate re-monitoring. Unlike traditional time-based mechanisms, this approach allows for timely updates of metrics when sudden changes occur in the network environment, preventing incorrect decisions caused by expired metrics.

[0053] Apply the calculated final validity period (TTL_final) to different monitoring indicators: Assign a specific final validity period (TTL_final) to each monitoring indicator: TTL_final_RSSI = TTL_feedback_adjusted_RSSI; TTL_final_SINR = TTL_feedback_adjusted_SINR; TTL_final_Delay = TTL_feedback_adjusted_NetParams; TTL_final_PacketLoss = TTL_feedback_adjusted_NetParams. TTL_feedback_adjusted_RSSI is the RSSI validity period adjusted based on historical data and environmental changes, and TTL_feedback_adjusted_NetParams is the validity period adjusted based on network delay characteristics or packet loss. Update the expiry date Expiry_timestamp for each metric: for each metric in Active_metrics: Metric.Expiry_timestamp = Current_timestamp + Metric.TTL_final; maintain the metric expiry date queue TTL_queue, sorting by Expiry_timestamp; schedule the next monitoring task based on the metric expiry date queue TTL_queue: Next_monitoring_time = Min(Expiry_timestamps) Schedule_monitoring_task(Next_monitoring_time). Output each metric's expiry date information TTL_info to the monitoring scheduling module and signal quality assessment module for use. Metric.TTL_final is the final expiry date for the metric, and Next_monitoring_time is the next monitoring time.

[0054] According to one aspect of the present application, obtaining a transmission execution result includes: S41. When a new environment is identified, similar environments are searched in the network environment library and parameter sets are extracted. An initial parameter set is calculated based on similarity weighting. The signal quality assessment parameters and decision thresholds are optimized based on historical transmission result records. S42. Calculate a transmission decision score based on the signal quality evaluation parameter, the comprehensive signal quality index, and the priority of the data to be transmitted, and compare it with the decision threshold to make a decision on whether to transmit or wait; S43. If transmitting, based on the current environment characteristics and the data to be transmitted, select the optimal transmission parameters, prioritize and appropriately package the data to be transmitted, execute the transmission and obtain the transmission execution result.

[0055] Specifically, when a new environment is identified, a small-sample rapid adaptation algorithm is implemented: the top-K similar environments are searched in the environment database, and their parameter sets (Params_similar) are extracted. The initial parameter set (Params_init) is calculated based on similarity weighting: Params_init = ∑(Sim_scorei × Params_similari) / ∑Sim_scorei, where Sim_scorei is the similarity score of the i-th similar environment. Unlike traditional methods that require a large number of samples to build an accurate model, this algorithm achieves small-sample rapid adaptation by transferring environment similarity. For the current environment ID, the signal quality assessment parameters (SQI_params) and decision threshold (Decision_th) are optimized based on historical transmission results (Records): SQI_params_new = SQI_params_old + η × (ActualSuccess - PredictedSuccess) × FeatureValues; Decision_th_new = Decision_th_old + Δ × (TargetRate - ActualRate). η is the learning rate, and Δ is the adjustment step size. This embodiment breaks through the limitations of traditional global models and maintains independent optimal parameters for each environment.

[0056] Based on the current environment ID, the corresponding signal quality index (SQI), the priority of the data to be transmitted (Priority), and the decision threshold (Decision_th), the transmission decision score (Tscore) is calculated: Tscore = SQI + Priority_weight × Priority. When Tscore > Decision_th, the system is considered to be in a transmission state; otherwise, it enters standby mode. This multi-factor decision-making mechanism, unlike traditional decisions based solely on signal strength, enables more intelligent transmission control.

[0057] The results of transmission attempts (success / failure) are recorded, and the following parameters in the current environment model are updated: Stab_score (environment stability score); Success_rate (transmission success rate); and Opt_threshold (optimal transmission threshold). Through a closed-loop feedback mechanism, the environment-specific model is continuously optimized, enabling the system to adapt to changes in the network environment. Based on the current environment characteristics and the data set to be transmitted, the following optimizations are performed: optimal transmission parameters (such as data fragment size and retransmission timeout) are selected based on the environment ID; data to be transmitted is prioritized and appropriately packaged; and the transmission is executed and the results are recorded. Unlike traditional fixed parameter strategies, this dynamically adjusts transmission parameters based on network environment characteristics to maximize transmission efficiency.

[0058] In another embodiment of the present application, the signal quality weight SQI_weights and decision threshold Decision_threshold in the optimization parameter set Optimized_params and the current monitoring indicator set I_set are read to make a transmission timing decision: the signal quality weight SQI_weights is used to perform weighted calculation on each indicator in the monitoring indicator set I_set to obtain the signal quality index: SQI = SQI_weights.w1×RSSI_norm + SQI_weights.w2×SINR_norm- SQI_weights.w3×Delay_norm - SQI_weights.w4×PacketLoss_norm; wherein each indicator is normalized (marked as _norm). The priority of the data to be transmitted Priority is read and the transmission decision score is calculated: Tscore= SQI + Priority_weight × Priority; the transmission decision score Tscore is compared with the decision threshold Decision_threshold: when Tscore > Decision_threshold, it is determined to be in a transmittable state; otherwise, it is determined to be in a waiting state. The transmission status determination result, Transmission_status (transmittable / waiting), is output to the resource optimization and transmission execution module. The current signal quality index (SQI), transmission decision score (Tscore), and determination result, Transmission_status, are recorded to form a decision record (Decision_record) for subsequent parameter optimization and feedback analysis. Transmission timing decisions, by comprehensively considering signal quality and data priority, achieve precise transmission timing control and improve transmission success rates.

[0059] After a transmission attempt completes, the system reads the transmission results (success / failure) and transmission statistics (delay, number of retries, etc.) and updates the environment model. Based on the transmission results and statistics, the system calculates the actual performance metrics for this transmission. The system compares the performance metrics predicted before the transmission with the actual performance metrics to calculate the prediction error. Based on the prediction error, the system updates the following parameters in the environment record (Env_record) corresponding to the current environment ID: Stab_score, Success_rate, SQI_weights, Decision_threshold, and other communication control parameters. The updated parameters are saved back to the environment database, forming an updated optimized parameter set. This optimized parameter set is then fed to the Transmission Timing Decision Module and the Resource Optimization and Transmission Execution Module for use in the next transmission decision. Through a closed-loop feedback mechanism, the environment-specific model is continuously optimized, enabling the system to continuously adapt to changes in the network environment.

[0060] The system reads the transmission status determination result, the optimization parameter set, and the data set to be transmitted, and performs transmission resource optimization and execution. When the transmission status determination result is "transferable," the system performs the following resource optimization steps: extracts transmission control parameters (such as data segment size and retransmission timeout) from the optimization parameter set; fine-tunes the transmission control parameters based on the characteristics of the current environment ID to generate the final transmission parameters; prioritizes and appropriately packages the data sets to be transmitted to generate transmission data packets; uses the final transmission parameters to control the sending process of the transmission data packets; monitors the transmission process, records the transmission results and transmission statistics, and outputs the transmission results and transmission statistics to the environment feedback and model update module; updates the status of the data set to be transmitted based on the transmission results: successfully transmitted data is removed from the data set to be transmitted; and data that failed to transmit is requeued or downgraded based on the policy.

[0061] According to one aspect of the present application, the initial parameter set is calculated based on similarity weighting, including: Using a similarity function, the current network environment fingerprint is compared with the historical network environment fingerprints in the network environment library. Environments with similarities exceeding a preset threshold are selected to form a similar environment set and parameter stability is verified. The reliability score and similarity score of each environment are calculated and multiplied to obtain a comprehensive weight. Extract the communication parameter sets of each environment from the similar environment set and perform weighted calculations on them using the corresponding comprehensive weights to obtain the weighted parameter sets. The weighted calculations are as follows: standardize numerical parameters and then perform weighted averaging, use weighted voting for Boolean parameters, and use the parameter value of the environment with the highest comprehensive weight for enumerated parameters. The weighted parameter set is adaptively adjusted according to the characteristics of the current network environment fingerprint to form an initial parameter set.

[0062] Specifically, the system reads the current network fingerprint and the environment library EnvLib and performs a search for similar environments: It iterates through each environment record in the environment library EnvLib and uses the fingerprint similarity calculation function SimFunc to calculate the similarity between the current fingerprint and the previous fingerprint: for each Env_record in EnvLib: Sim_score = SimFunc(NF_current, Env_record.NF)Similarity_list.append({EnvID: Env_record.EnvID, Score: Sim_score}); The similarity list Similarity_list is sorted in descending order by similarity score to obtain the sorted similarity list Sorted_similarity_list; The K environments with the highest similarity scores are selected as the Top-K similar environment set, with K = 5 by default. Unlike traditional single-metric matching, this method can evaluate environment similarity from multiple dimensions.

[0063] Evaluate the effectiveness of similar environments. For each environment in the Top-K similar environment set, evaluate its parameter validity: read the communication success rate (Success_rate), environment duration (Duration), and last update time (Last_update) for each similar environment. Calculate the parameter reliability score (Reliability_score) for each environment: Reliability_score = Success_rate * min(1, Duration / Min_reliable_duration) * exp(-λ * (Current_time - Last_update)), where Min_reliable_duration is the minimum duration threshold for reliability evaluation and λ is the time decay factor. Select environments with reliability scores greater than the reliability threshold (R_threshold) to form a reliable similar environment set. Ensure that only fully verified environment parameters are used for new environment initialization to avoid interference from unreliable parameters.

[0064] Based on the set of reliable similar environments, weighted parameter fusion is performed. If the set of reliable similar environments is empty, the default parameter set is used as the initial parameter set Params_init. Otherwise, the communication parameter set Params is extracted from each reliable similar environment, including: SQI_weights (signal quality index weights); decision_threshold (transmission decision threshold); monitoring_intervals (monitoring intervals); and time-to-live (TTL_factors). The combined weight Total_weight of each environment is calculated based on the similarity score Sim_score and the reliability score Reliability_score: Total_weight = Sim_score * Reliability_score. The combined weight Total_weight is weighted averaged to calculate the initial parameter set Params_init: Params_init = ∑(Total_weighti * Paramsi) / ∑Total_weighti.

[0065] The system reads the characteristics of the current network fingerprint and adaptively adjusts the initial parameter set: It analyzes features such as signal strength, signal quality, and the number of neighboring cells |N_list| in the current network fingerprint; calculates the environmental feature vector Env_features based on these features; and applies the environmental adaptive adjustment function Adapt_func to fine-tune the initial parameters: Params_adjusted = Adapt_func(Params_init, Env_features). Specific adjustments include: if the RSSI (signal strength RSSI) is poor, increasing the conservativeness of the decision threshold Decision_threshold; if the number of neighboring cells is large, increasing the weight of topology changes in environmental change detection; and if the SINR (signal quality unreliable ratio) fluctuates significantly, reducing the signal indicator validity period (TTL). This ensures that the initial parameters are more consistent with the actual conditions of the new environment and reduces the number of samples required for parameter convergence.

[0066] Evaluate the reliability and uncertainty of the adjusted parameter set Params_adjusted: Based on the size and distribution of the reliable set of similar environments, calculate the parameter uncertainty index Uncertainty_index: Param_variance = Variance(Params_reliable_envs); Similarity_factor = Average(Sim_scores); Uncertainty_index = Param_variance * (1 - Similarity_factor). Set the learning rate Learning_rate and initial exploration rate Exploration_rate based on the uncertainty index Uncertainty_index: Learning_rate = Base_learning_rate * (1 + Uncertainty_index); Exploration_rate = Base_exploration_rate * Uncertainty_index. This allows the system to perform more aggressive parameter exploration and adjustment when parameter uncertainty is high, accelerating parameter convergence.

[0067] Apply the finalized parameters to the new environment: combine the adjusted parameter set, learning rate, and exploration degree into the initial configuration of the new environment; create a new environment record containing: environment ID, network fingerprint, initial configuration, and creation timestamp. Write the new environment record into the environment library; output the new environment parameter set to the communication control decision module for use. This embodiment fully implements the process of quickly determining the initial parameters of the new environment. Its core lies in utilizing empirical knowledge from historical similar environments and combining it with the characteristics of the current environment to perform parameter migration and adaptive adjustment, achieving rapid parameter initialization in small or even zero sample situations, and reducing the convergence time of parameter optimization in the new environment.

[0068] According to one aspect of the present application, optimizing signal quality assessment parameters involves defining an environment-specific optimization objective based on the current environment ID and service requirement configuration: reading the success rate target Target_success_rate, delay requirement Max_delay, and energy consumption limit Energy_constraint from the service requirement configuration; and determining optimization weights based on the environment stability score Stab_score corresponding to the current environment ID: if Stab_score > High_stability_threshold: Success_weight = 0.7; Energy_weight = 0.3; if Stab_score > Medium_stability_threshold: Success_weight = 0.5; Energy_weight = 0.5; else: Success_weight = 0.3; Energy_weight = 0.7. Constructing an environment-specific optimization objective Env_specific_objective, expressed as a multi-objective weighted function: Objective = Success_weight * Success_rate - Energy_weight * Energy_consumption. Unlike traditional fixed-target approaches, this approach allows for flexible adjustment of optimization priorities based on environmental characteristics. Where High_stability_threshold is the high stability threshold, Medium_stability_threshold is the medium stability threshold, Success_weight is the success rate weight, and Energy_weight is the energy consumption weight. Read historical transmission records corresponding to the current environment ID and perform data analysis: divide the historical transmission records into a training set (Training_set) and a validation set (Validation_set) by time. From the training set, extract the following for each transmission: signal quality indicators (Signals, RSSI, SINR, etc.); transmission parameters (TX_params, decision threshold, data size, etc.); and transmission results (Results, success / failure, latency, and number of retries). Calculate the success rate (Success_rate) and energy consumption (Energy_consumption) for different parameter settings to generate a parameter-performance mapping. This provides a solid basis for subsequent parameter optimization.Based on the parameter-performance mapping, analyze the sensitivity of each parameter to performance: For each parameter to be optimized, calculate its sensitivity coefficient: for each Parameter in [SQI_weights, Decision_threshold, ...]: Delta_performance = Performance(Parameter + Delta) - Performance(Parameter); Sensitivity = Delta_performance / Delta. Parameters are ranked by importance based on the sensitivity coefficient, resulting in a parameter importance ranking (Param_importance) to determine the highly sensitive parameter set (High_sensitivity_params) and the less sensitive parameter set (Low_sensitivity_params). Unlike traditional equal-weight optimization, this method focuses optimization resources on the parameters with the greatest impact. Delta_performance represents the change in performance after the parameter change, and Delta represents the small adjustment to the parameter.

[0069] For highly sensitive parameter sets, gradient estimation and parameter updates are performed: For each highly sensitive parameter, its performance gradient is estimated based on historical transmission records: for each Parameter in High_sensitivity_params: Success_gradient = ΨSuccess_rate / ΨParameter; Energy_gradient = ΨEnergy_consumption / ΨParameter; Objective_gradient = Success_weight * Success_gradient - Energy_weight * Energy_gradient. The adaptive learning rate Adaptive_learning_rate is applied to update the parameters: Parameter_new = Parameter_old + Adaptive_learning_rate * Objective_gradient. The adaptive learning rate is dynamically adjusted based on the performance fluctuations of the current parameters and environmental stability. For less sensitive parameter sets, a smaller fixed learning rate is used for updates to reduce computational overhead. Unlike traditional fixed-step updates, this method can ensure convergence speed while avoiding parameter oscillation. Where Ψ is the partial derivative, Success_gradient is the success rate gradient, Energy_gradient is the energy consumption gradient, and Objective_gradient is the optimization target gradient.

[0070] Dynamically adjust the exploration-exploitation balance based on the data sufficiency and performance stability of the current environment ID: Data sufficiency is calculated as: Data_sufficiency = min(1, |Training_set| / Sufficient_sample_size), where Sufficient_sample_size is the sufficient sample size. Performance stability is calculated as: Performance_stability = 1 - StdDev(Recent_success_rates) / Mean(Recent_success_rates), where Recent_success_rates is the recent success rate. The exploration rate is determined as: Exploration_rate = Base_exploration * (1 - Data_sufficiency) * (1 - Performance_stability), where Base_exploration is the base exploration rate. An exploration mechanism is applied during parameter updates: parameters are randomly perturbed at a probabilistic exploration rate; the perturbation amplitude decreases with increasing data sufficiency. Unlike traditional fixed exploration rate methods, this method increases exploration when data is insufficient and focuses on exploitation when performance is stable.

[0071] Use the validation set (Validation_set) to verify optimized parameter performance and implement rollback protection: Apply the updated parameters to the validation set and calculate the new parameter performance (New_performance); Compare the new parameter performance (New_performance) with the old parameter performance (Old_performance): if New_performance < Old_performance * (1 - Degradation_tolerance): / / Performance degradation exceeds the tolerance, triggering the rollback mechanism; Rollback_to_previous_parameters(); Reduce_learning_rate(); Mark_update_as_failed(); else: / / Performance improves or is within an acceptable range, confirm the update; Confirm_parameter_update(); if New_performance > Old_performance * (1 + Improvement_threshold): Increase_learning_rate() / / Performance has significantly improved, and the learning rate can be appropriately increased. This ensures the robustness of the parameter optimization process and prevents performance degradation caused by erroneous updates. Among them, Rollback_to_previous_parameters() means rolling back to the previous parameters, Reduce_learning_rate() means reducing the learning rate, Mark_update_as_failed() means marking the update failure, and Improvement_threshold is the performance improvement threshold. The finally confirmed optimization parameters are persisted and distributed to relevant modules: the latest optimization parameter set Optimized_params is updated to the environment record Env_record corresponding to the current environment ID, the current parameter version number Param_version and update timestamp are recorded, and the corresponding records in the environment library are updated; the optimization parameter set is output to the communication control decision module for use. Ensure that all relevant modules use the latest optimization parameters. This embodiment fully implements the environment-specific parameter optimization process. Its core lies in the optimization goal setting of environment adaptation, parameter sensitivity analysis and importance distinction, adaptive learning rate adjustment, and exploration-utilization balance mechanism, etc., realizing the refined optimization of parameters for different network environments and improving communication efficiency and reliability.

[0072] In another embodiment of the present application, weighted fusion of similar environments can also be performed as follows: based on a reliable similar environment set, parameter weighted fusion is performed: if the reliable similar environment set is empty, the default parameter set Default_params is used as the initial parameter set Params_init; otherwise, the communication parameter set Params is extracted from each reliable similar environment and grouped by parameter type: numerical parameters (such as signal quality index weight, decision threshold, monitoring interval): weighted average is used; Boolean parameters: majority voting is used, and the voting weight is weighted according to the similarity; enumerated parameters (such as transmission mode): the parameter value of the environment with the highest similarity is used; interval parameters: weighted average is applied to the upper and lower bounds respectively. For numerical parameters, normalization is first performed to eliminate the dimension effect: Param_normalized = (Param - Min_param) / (Max_param - Min_param), where Min_param and Max_param are the preset minimum and maximum values of the parameter. Calculate the combined weight (Total_weight) for each environment based on the similarity score (Sim_score) and the reliability score (Reliability_score): Total_weight = Sim_score * Reliability_score. Use the combined weight (Total_weight) to perform a weighted average and calculate the initial numerical parameter set: Param_init_normalized = ∑(Total_weighti *Param_normalizedi) / ∑Total_weighti. Denormalize the normalized parameters to restore their original dimensions: Param_init = Min_param + Param_init_normalized * (Max_param - Min_param). Combine the parameter results of different types to form the complete initial parameter set (Params_init).

[0073] In another embodiment of the present application, the adaptive threshold calculation can also be as follows: based on the historical environment change record History_changes and the communication success rate change record Success_rate_changes, the environment segment threshold Eth is dynamically calculated: the base threshold adjustment factor is calculated: Base_threshold = Config_base_threshold; the success rate adjustment factor is calculated: Success_rate_adjustment = k2 * (1-Avg_success_rate); when the success rate is low, Success_rate_adjustment increases, the final threshold is lowered, and sensitivity is increased. The change frequency adjustment factor is calculated: Frequency_adjustment = k1 * Change_freq; when changes are frequent, Frequency_adjustment increases, the final threshold is raised, and sensitivity is reduced. The overall adjustment factor is calculated to distinguish and process different impacts: Eth = Base_threshold * (1 + Frequency_adjustment) * (1 - Success_rate_adjustment) * Env_specific_factor. This adjustment mechanism ensures that: when the communication success rate is low, the threshold is lowered, the sensitivity of environmental segmentation is increased, and it is easier to trigger environmental reconstruction; when the environment changes frequently, the threshold is increased, the sensitivity is reduced, and excessive segmentation is avoided; Env_specific_factor, as an environment-specific adjustment factor, can be fine-tuned according to the specific environment type.

[0074] In another embodiment of the present application, the historical accuracy feedback adjustment can also be: feedback adjustment is performed based on the historical indicator validity evaluation: the transmission failure rate Expiry_failure_rate caused by the expiration of the indicator and the failure rate Change_failure_rate caused by the change of the indicator during the validity period in the historical transmission record are read; the comprehensive failure impact coefficient Compound_failure_factor is calculated: Expiry_adjustment = (Expiry_failure_rate > Threshold_expiry) ? Expiry_correction_factor * (Expiry_failure_rate - Threshold_expiry) / (1 - Threshold_expiry): 0; Change_adjustment = (Change_failure_rate > Threshold_change) ? Change_correction_factor*(Change_failure_rate-Threshold_change) / (1-Threshold_change): 0; / / Use the Max function to avoid over-adjustment caused by simply adding two adjustment factors; / / α is the balance factor between indicator expiration failure and indicator change failure, with a default value of 0.7. Compound_failure_factor = Max(Expiry_adjustment, Change_adjustment) + α * Min(Expiry_adjustment, Change_adjustment). Apply the compound failure impact factor to adjust the validity period: TTL_feedback_adjusted = TTL_energy_adjusted * (1 - Compound_failure_factor). Set the lower limit for validity period adjustment: TTL_feedback_adjusted = Max(TTL_feedback_adjusted, Min_acceptable_TTL). This considers the impact of both failure scenarios while avoiding over-adjustment caused by simply over-adding them.

[0075] In another embodiment of the present application, fingerprint validity verification can also include: verifying the recognition effectiveness of the constructed optimized fingerprint: defining a short-term historical fingerprint set: data structure: a circular buffer with a maximum capacity of Max_recent_fingerprints (default value: 20); initial state: an empty set; update mechanism: each successfully constructed new fingerprint is added to the set; when the set reaches the maximum capacity, the oldest fingerprint is replaced. Fingerprint validity management is defined: each fingerprint record includes a timestamp; a fingerprint validity period, Fingerprint_validity_period, is set (default value: 4 hours); fingerprints whose current time exceeds timestamp + Fingerprint_validity_period are marked as expired. Fingerprint validity verification process: selecting the most recent N non-expired fingerprints (default value: N=3) from the short-term historical fingerprint set, calculating the similarity between the current fingerprint and these historical fingerprints; and confirming that the temporally continuous fingerprint similarity satisfies the continuity condition, Continuity_condition: SimFunc(NF_current, NF_recent) > Continuity_threshold. Where SimFunc is the similarity function, NF_recent is the historical fingerprint, and Continuity_threshold is the continuity threshold. Exception handling mechanism: If the continuity condition is not met, fingerprint exception handling is triggered: the current fingerprint is marked as a suspicious fingerprint (Suspicious_fingerprint); the priority of the next network data collection is increased; and the previous valid fingerprint is temporarily retained for environmental identification. Regular maintenance mechanism: During system idle time (such as every morning), a complete cleanup of the short-term historical fingerprint set is performed: all expired fingerprints are removed; if the set size exceeds 80% of Max_recent_fingerprints, the oldest 20% of fingerprints are removed.

[0076] In another embodiment of the present application, the fingerprint similarity calculation method may further include: establishing a fingerprint similarity calculation function SimFunc for evaluating the similarity between two network fingerprints: Initial weight setting: During system initialization, default weights are set: w1_init = 0.5, w2_init = 0.3, and w3_init = 0.2. These initial values are determined based on preliminary experimental results and reflect the general importance of different features for environment recognition. A specific set of weight coefficients {w1_env, w2_env, w3_env} is maintained for each environment ID. New environments initially use the global default weights, which are then gradually adjusted based on the recognition performance in that environment. A dynamic weight adjustment mechanism is implemented: a weight adjustment learning rate η is defined (default value 0.05). After each environment recognition result is verified, the weights are updated: if the recognition is correct, the weight of the feature with the greatest contribution is increased; if the recognition is incorrect, the weight of the feature with the greatest difference between the incorrect environment and the correct environment is increased; and weight normalization is performed: the adjusted weights are renormalized to ensure that w1 + w2 + w3 = 1. Weight Validation: Regularly evaluate the validity of the current weights (e.g., after every 50 environment recognitions). If recognition accuracy decreases, roll back to the previous version of the weights. If accuracy continues to improve, reduce the learning rate η appropriately to avoid overfitting. Calculate similarity: SimFunc(NF1, NF2) = w1·I(NF1.primaryID == NF2.primaryID) + w2·CosSim(NF1.N_list, NF2.N_list) + w3·(1-|NF1.RSSI-NF2.RSSI| / MaxRSSDiff). Use environment-specific weight coefficients {w1_env, w2_env, w3_env}. Weight Persistence: Persistently store all environment weight coefficients before system shutdown. Load historical weight coefficients after system restart. Periodically restore weight coefficients (e.g., every 24 hours) to prevent loss of learning results due to unexpected power outages.

[0077] In another embodiment of the present application, environment-specific parameters are optimized. For the current environment ID, based on historical transmission results, the parameters are optimized: the parameter learning rate and exploration degree in the environment record are obtained: the initial values in the environment record corresponding to the current environment ID are read. If these values are not in the environment record, the default values are used. Dynamic adjustment of parameter learning rate: based on the effects of the last N (default N=10) parameter updates, the parameter adjustment effectiveness index Adjustment_effectiveness is calculated; if Adjustment_effectiveness > Effectiveness_threshold, Learning_rate is reduced: Learning_rate=Learning_rate*Learning_rate_decay; if Adjustment_effectiveness<Low_effectiveness_threshold, Learning_rate is increased: Learning_rate = Min(Learning_rate * Learning_rate_boost, Max_learning_rate). Dynamically adjust the exploration rate: Decrease with increasing environmental data volume: Exploration_rate = Exploration_rate * (1 - Data_increment_factor); but never fall below the minimum exploration rate: Exploration_rate = Max(Exploration_rate, Min_exploration_rate). Optimize the adjusted parameters: SQI_params_new = SQI_params_old + Learning_rate * (ActualSuccess - PredictedSuccess) * FeatureValues; Decision_th_new = Decision_th_old + Learning_rate * (TargetRate - ActualRate). Implementation of the exploration mechanism: Randomly perturb the optimized parameters with a probability of Exploration_rate: Parameter_final = Parameter_optimized + Random_disturbance * Exploration_rate. This directly influences the subsequent parameter optimization process, completing a complete data chain from parameter initialization to parameter optimization.

[0078] According to another aspect of the present application, a communication control method for a smart energy unit security access module includes: the smart energy unit security access module has an independent dialing function, and when the dialing fails, it retries, and after a certain number of retries, it sleeps for a period of time and then dials again; when the 4G signal is normal but the login platform fails, it retries, and after a certain number of retries, it sleeps for a period of time in an incremental manner and then dials again and logs in again; real-time monitoring and recording of the 4G signal in the current environment, forming a curve and predicting the signal curve in the future based on it; after data transmission fails, it retries, and after a certain number of retries, it writes to a historical data record file; according to the signal prediction curve, when the signal is good, it automatically generates a sending task and sends historical data.

[0079] Specifically, the smart energy unit's secure access module is a relatively independent module responsible for data exchange between the smart energy unit and the platform. It provides communication detection and exception handling functions: It re-dials upon system startup or after an abnormal 4G disconnection. If the dialup fails, the software automatically locates the cause of the anomaly, uses a built-in anomaly signature library to identify and resolve the issue, and then re-dials. After three consecutive unsuccessful retries, the dialing function enters a dormant state for 10 minutes. The smart energy unit needs to log in to the platform through the secure access module. Login failures occur when the network is congested or the platform is abnormal. After three consecutive unsuccessful retries, the login function is delayed in one-minute increments, reaching a maximum of 30 minutes and then no longer increasing. Upon successful login, the initial one-minute delay is restored. Failure in data reporting, such as three consecutive unsuccessful retries, is recorded as a transmission failure. After five consecutive transmission failures, the pending data is recorded in a history file. The secure access module records 4G signal values in real time, creating a historical signal curve. When there is historical data to be sent, the future signal curve is predicted based on the historical signal curve, and the optimal time period is selected to generate the data sending task and wait for execution.

[0080] In a specific embodiment of this application, a communication control method for a secure access module of a smart energy unit is demonstrated. This method addresses communication control issues for smart energy units in diverse network environments through segmented environmental management based on network topology changes and a small sample rapid adaptation algorithm. The following details the implementation of this method, using specific data and calculations.

[0081] Step 1: Network environment perception and fingerprint construction.

[0082] 1.1. Base station information collection and neighboring cell information acquisition. In this embodiment, the device is located in a commercial area and connected to a 4G base station. The following basic network parameter set is obtained through the API: main base station ID: "4G-BS-2305"; signal strength (RSSI): -75dBm; signal quality (SINR): 12dB. At the same time, the device scans the surrounding network environment and obtains a list of neighboring cells and their signal strengths: Neighboring cell list (N_list): ["4G-BS-2306", "4G-BS-2304", "4G-BS-2401"]; Signal strength list (S_list): [-82dBm, -88dBm, -95dBm].

[0083] 1.2. Wireless network fingerprint construction.

[0084] 1.2.1. Raw network data collection. Combine the basic network parameter set, neighboring cell list, and signal strength list to form the raw network data set: Raw_network_data = {"primaryID": "4G-BS-2305", "RSSI": -75, "SINR": 12, "N_list": ["4G-BS-2306", "4G-BS-2304", "4G-BS-2401"], "S_list": [-82, -88, -95], "timestamp": "2025-05-10T14:30:25"}.

[0085] 1.2.2 Feature Selection and Extraction. Key features were selected from the original network dataset: the primary base station ID was used as the primary identification feature, and the signal strength normalization value was RSSI_norm = (RSSI - (-100)) / ((-60) - (-100)) = (-75 - (-100)) / 40 = 0.625. Neighboring cell topology features were extracted: cell IDs and corresponding signal strengths were extracted to construct a topology representation.

[0086] 1.2.3. Feature Stability Analysis. The stability of each feature was calculated using historical network data: The primary base station's RSSI measurement samples over the past 30 minutes were: [-76, -78, -74, -75, -77, -75, -73, -75]; RSSI average: -75.375 dBm; RSSI standard deviation: 1.5 dBm; RSSI coefficient of variation (RSSI_CV): standard deviation / |average| = 1.5 / 75.375 = 0.0199. Number of neighbor cell list changes over the past 30 minutes: 2; cell change rate (cell_change_rate): 2 / 30 = 0.0667 times / minute. Based on the above stability indicators, the feature reliability weights are calculated: RSSI_weight = 1 / (1 + RSSI_CV) = 1 / (1 + 0.0199) = 0.9805; Topo_weight = 1 / (1 + Cell_change_rate) = 1 / (1 + 0.0667) = 0.9375.

[0087] 1.2.4. Fingerprint vectorization representation. Construct a structured network environment fingerprint: NF_structure = {"basic_features":{"primaryID":"4G-BS-2305","RSSI":-75,"SINR":12,"N_list":["4G-BS-2306","4G-BS-2304","4G-BS-2401"],"S_list":[-82, -88, -95]},"derived_features":{"RSSI_norm":0.625,"Features_weight":{"RSSI_weight":0.9805,"Topo_weight":0.9375}}},"metadata":{"timestamp":"2025-05-10T14:30:25"}}.

[0088] 1.2.5 Fingerprint Compression and Optimization. Compress and optimize the structured network environment fingerprint: remove cells with signal strength below -90dBm: "4G-BS-2401" (-95dBm); update the neighboring cell list and signal strength list: N_list_optimized = ["4G-BS-2306", "4G-BS-2304"]; S_list_optimized = [-82, -88]; quantize RSSI values to 2dBm accuracy: -75dBm is quantized to -76dBm. Optimized fingerprint: NF_optimized = {"primaryID":"4G-BS-2305","RSSI":-76, "SINR":12,"N_list":["4G-BS-2306","4G-BS-2304"],"S_list":[-82, -88],"RSSI_norm":0.6,"Features_weight":{"RSSI_weight":0.98,"Topo_weight":0.94},"timestamp":"2025-05-10T14:30:25"}.

[0089] 1.3. Fingerprint similarity calculation function definition. Set the initial weight coefficients for fingerprint similarity calculation: w1_init = 0.5 (master base station ID matching weight); w2_init = 0.3 (neighboring cell topology similarity weight); w3_init = 0.2 (signal strength difference weight). Similarity calculation formula: SimFunc(NF1, NF2) = w1 I(NF1.primaryID == NF2.primaryID) + w2 CosSim(NF1.N_list, NF2.N_list) + w3 (1-|NF1.RSSI -NF2.RSSI| / MaxRSSDiff); where I(NF1.primaryID == NF2.primaryID) is the primary base station ID match indicator function, which is 1 when the primary base station IDs of the two fingerprints are the same, and 0 otherwise; CosSim(NF1.N_list, NF2.N_list) is the cosine similarity of the neighboring cell lists; |NF1.RSSI - NF2.RSSI| is the absolute value of the signal strength difference between the two fingerprints; MaxRSSDiff is the set maximum signal strength difference, which is 40dBm; w1, w2, and w3 are weight coefficients, and w1+w2+w3=1. Taking the current network environment fingerprint and the historical network environment fingerprint NF_history as an example, calculate the similarity: NF_current.primaryID = "4G-BS-2305"; NF_history.primaryID = "4G-BS-2305"; NF_current.N_list = ["4G-BS-2306", "4G-BS-2304"]; NF_history.N_list = ["4G-BS-2306", "4G-BS-2304", "4G-BS-2308"]; NF_current.RSSI = -76dBm; NF_history.RSSI = -72dBm. Calculation process: I(NF_current.primaryID == NF_history.primaryID) = 1 (the primary base station ID is the same); calculate the cosine similarity CosSim: a vector indicating the presence or absence of each cell: the eigenvector of NF_current is [1, 1, 0] ("4G-BS-2306" exists, "4G-BS-2304" exists, "4G-BS-2308" does not exist); the eigenvector of NF_history is [1, 1, 1]; CosSim = (1×1 + 1×1 + 0×1) / (sqrt(1 2 +1 2 +0 2) × sqrt (1 2 +1 2 +1 2 )) = 2 / (sqrt (2) × sqrt (3)) = 2 / sqrt (6) ≈ 0.816. Signal strength difference normalization: 1 - |NF_current.RSSI - NF_history.RSSI| / MaxRSSDiff = 1 - |-76 - (-72)| / 40 = 1 - 4 / 40 = 0.9; Final similarity score: SimFunc(NF_current, NF_history) = 0.5×1 +0.3×0.816 + 0.2×0.9 = 0.5 + 0.2448 + 0.18 = 0.9248.

[0090] Step 2: Environmental change detection and model segmentation management.

[0091] 2.1. Calculation of Environmental Change Metrics. Compare the current network environment fingerprint with the historical fingerprint and calculate the degree of environmental change: Environmental Change (ΔE) = w1 (1-I(NF_current.primaryID == NF_history.primaryID)) + w2 (1-CosSim(NF_current.N_list, NF_history.N_list)) + w3 |NF_current.RSSI - NF_history.RSSI| / MaxRSSDiff; where I(NF_current.primaryID == NF_history.primaryID) is the primary base station ID match indicator function; CosSim(NF_current.N_list, NF_history.N_list) is the cosine similarity of the neighboring cell list; |NF_current.RSSI - NF_history.RSSI| is the absolute value of the signal strength difference; MaxRSSDiff is the maximum signal strength difference, which is 40dBm; w1, w2, and w3 are weight coefficients, and w1+w2+w3=1. Calculate the environmental change based on the data in the previous similarity calculation: ΔE = 0.5×(1-1) +0.3×(1-0.816) + 0.2×4 / 40 = 0 + 0.0552 + 0.02 = 0.0752.

[0092] 2.2. Environment segmentation determination based on topology changes.

[0093] 2.2.1. Multi-dimensional decomposition of environmental change characteristics. The environmental variability is decomposed into three dimensional features: signal strength change component (ΔS) = |RSSI_current - RSSI_history| = |-76 - (-72)| = 4dBm; primary base station change flag (ΔP) = (primaryID_current != primaryID_history) ? 1:0=0; cell topology change rate (ΔN): N_list_common = Intersection(N_list_current, N_list_history) = ["4G-BS-2306", "4G-BS-2304"]; N_list_union = Union(N_list_current, N_list_history) = ["4G-BS-2306", "4G-BS-2304", "4G-BS-2308"]; ΔN = 1 - |N_list_common| / |N_list_union| = 1 - 2 / 3 = 0.333. Environmental change feature vector ΔVector = [4, 0, 0.333].

[0094] 2.2.2. Environmental Change Significance Assessment. Evaluate the actual impact of environmental changes on communications: a 4dBm change in signal strength corresponds to a communication impact coefficient of 4 / 40 = 0.1; a primary base station change impact coefficient of 0 (no primary base station change); a cell topology change rate of 0.333 corresponds to a communication impact coefficient of 0.333 × 0.5 = 0.1665; and the environmental change significance (Change_significance) = 0.1 + 0 + 0.1665 = 0.2665.

[0095] 2.2.3. Adaptive Threshold Calculation. Basic Threshold Setting: Read the system configuration's basic threshold: Config_base_threshold = 0.3. Success Rate Adjustment Factor Calculation: Average success rate of the last 30 communications: Avg_success_rate = 0.85. Success Rate Adjustment Factor: k2 = 0.5. Success Rate Adjustment Factor = k2 × (1 - Avg_success_rate) = 0.5 × (1 - 0.85) = 0.075. Change Frequency Adjustment Factor Calculation: Environment Change Frequency: Change_freq = 0.08 times / minute. Change Frequency Adjustment Factor: k1 = 5. Change Frequency Adjustment Factor = k1 × Change_freq = 5 × 0.08 = 0.4. Environment-specific adjustment factor: Based on the current environment characteristics: Env_specific_factor = 1.1; calculate the final adaptive environment segmentation threshold: Eth = Base_threshold × (1 + Frequency_adjustment) × (1 - Success_rate_adjustment) × Env_specific_factor = 0.3 ×(1 + 0.4) × (1 - 0.075) × 1.1 = 0.3 × 1.4 × 0.925 × 1.1 = 0.428.

[0096] 2.2.4. Multi-level change classification. Compare the environmental change significance with the adaptive threshold: Environmental change significance: Change_significance = 0.2665; Adaptive threshold: Eth = 0.428. Since Change_significance (0.2665) < Eth (0.428), the current environmental change is considered "minor" and does not trigger environmental segmentation, so it is classified as belonging to the current environmental segment.

[0097] 2.3 Environment Library Management. Assume that the environment IDs in the current environment library are: ["Env001", "Env002", "Env003"]. Since it is determined to be a "minor change", no new environment segment is created. Instead, the information of the current environment "Env003" is updated: the environment fingerprint is updated to the current latest fingerprint; the environment duration segment is extended; and the communication parameter set is updated.

[0098] 2.4. Current Environment Identification. Calculate the similarity between the current network environment fingerprint and each environment fingerprint in the environment database: SimFunc(NF_current, NF_Env001) = 0.5502; SimFunc(NF_current, NF_Env002) = 0.7123; SimFunc(NF_current, NF_Env003) = 0.9248; the most similar environment ID is "Env003", and the similarity score is 0.9248. Calculate the environment stability score based on the similarity scores: Stab_score = min(1.0, SimFunc(NF_current, NF_EnvBest) × 1.2) = min(1.0, 0.9248 × 1.2) = min(1.0, 1.10976) = 1.0. Generated environment parameters: Current environment ID (EnvID) = "Env003"; Environment stability score (Stab_score) = 1.0.

[0099] Step 3: Multi-index signal quality assessment and hierarchical monitoring.

[0100] 3.1. Monitoring Level Setting. The initial monitoring level is determined based on environmental parameters: Environmental Stability Score (Stab_score) = 1.0 (Extremely Stable); when Stab_score > 0.8, the basic monitoring level is set; Initial Monitoring Level (Level) = 1 (Basic Monitoring, monitoring only signal strength).

[0101] 3.2 Adaptive monitoring trigger. Design the monitoring trigger function: Trigger(RSSI, Stab_score, Priority) = α × (RSSI_threshold - RSSI) / RSSI_range + β × (1 - Stab_score) + γ × Priority / MaxPriority; where RSSI is the current signal strength; RSSI_threshold is the signal strength threshold, set to -80dBm; RSSI_range is the signal strength range, set to 40dBm; Stab_score is the environmental stability score; Priority is the priority of the data to be transmitted; MaxPriority is the highest priority, set to 10; α, β, and γ are weight coefficients, set to 0.3, 0.3, and 0.4, respectively, and α + β + γ = 1. Using the current environment as an example to calculate the trigger value: Current signal strength (RSSI) = -76 dBm; Environment stability score (Stab_score) = 1.0; Priority of data to be transmitted (Priority) = 8 (High); RSSI_threshold = -80 dBm; RSSI_range = 40 dBm; MaxPriority = 10; Trigger_value = 0.3 × (-80 - (-76)) / 40 + 0.3 × (1 - 1.0) + 0.4 × 8 / 10 = 0.3 × (-4) / 40 + 0.3 × 0 + 0.4 × 0.8 = -0.03 + 0 + 0.32 = 0.29. Set the trigger threshold: Trigger_threshold = 0.4; because Trigger_value(0.29) < Trigger_threshold(0.4), no higher-level monitoring is triggered, and the actual monitoring level is still 1 (basic monitoring).

[0102] 3.3. Indicator validity period management.

[0103] 3.3.1. Analysis of Dynamic Environment Characteristics. Based on the historical network environment fingerprint set, the following dynamic environment characteristics were analyzed: Signal Strength Short-Term Coefficient of Variation (RSSI_short_CV) = 0.0199; Signal Strength Long-Term Coefficient of Variation (RSSI_long_CV) = 0.0335; Network Topology Change Frequency (Topo_change_freq) = 0.0667 times / minute; Master Base Station Switching Frequency (BS_switch_freq) = 0.0083 times / minute. The dynamic environment feature vector was constructed: DynamicFeature = [0.0199, 0.0335, 0.0667, 0.0083].

[0104] 3.3.2. Calculation of Base Validity Period. Formulas for calculating base validity periods are designed for different indicator types: TTL_base_RSSI = Base_TTL × (1 / (1 + RSSI_short_CV × β1)); Base_TTL is the base validity period, set to 300 seconds; RSSI_short_CV is the short-term coefficient of variation of signal strength; β1 is the adjustment factor, set to 10. TTL_base_SINR = Base_TTL × (1 / (1 + RSSI_short_CV × β2)); Base_TTL is the base validity period, set to 300 seconds; RSSI_short_CV is the short-term coefficient of variation of signal strength; β2 is the adjustment factor, set to 15. TTL_base_NetParams = Base_TTL × (1 / (1 + Topo_change_freq × β3)); Base_TTL is the base validity period, set to 300 seconds; Topo_change_freq is the frequency of network topology changes; β3 is the adjustment factor, set to 20. Calculate the basic validity period of each indicator: TTL_base_RSSI = 300 × (1 / (1 + 0.0199 × 10)) = 300 × (1 / 1.199) = 300 × 0.834 = 250.2 seconds; TTL_base_SINR = 300 × (1 / (1 + 0.0199 × 15)) = 300 × (1 / 1.2985) = 300 × 0.77 = 231.0 seconds; TTL_base_NetParams = 300 × (1 / (1 + 0.0667 ×20)) = 300 × (1 / 2.334) = 300 × 0.4285 = 128.55 seconds.

[0105] 3.3.3 Data Priority and Energy Status Adjustment. Read the highest priority of the data to be transmitted and the current energy status of the device: Maximum priority (MaxPriority) = 8 (high); Current device battery percentage (Battery_percent) = 65%. Adjust the validity period based on data priority: Priority_factor = 1 - Priority_weight × Priority / MaxPriority; where Priority_weight is the priority weight, set to 0.4; Priority is the current priority, value 8; MaxPriority is the highest priority, value 10. Priority_factor = 1 - 0.4 × 8 / 10 = 1 - 0.32 = 0.68. Adjust the validity period based on the energy status: Energy_factor = 1 + Energy_weight × (1 - Battery_percent / 100); where Energy_weight is the energy weight, set to 0.6; Battery_percent is the battery percentage, value 65%. Energy_factor = 1 + 0.6 × (1 - 65 / 100) = 1 + 0.6 × 0.35 = 1 + 0.21 = 1.21. Apply these two factors to adjust the validity period: TTL_adjusted_RSSI = TTL_base_RSSI × Priority_factor × Energy_factor = 250.2 × 0.68 × 1.21 = 205.36 seconds; TTL_adjusted_SINR = TTL_base_SINR × Priority_factor × Energy_factor = 231.0 × 0.68 × 1.21 = 189.61 seconds; TTL_adjusted_NetParams = TTL_base_NetParams × Priority_factor × Energy_factor = 128.55 × 0.68 × 1.21 = 105.49 seconds.

[0106] 3.3.4. Historical Accuracy Feedback Adjustment. Based on historical transmission records, the following statistics were used: the transmission failure rate due to expiration of indicators (Expiry_failure_rate) = 0.15; the failure rate due to changes in indicators during the validity period (Change_failure_rate) = 0.08; the preset thresholds: Threshold_expiry = 0.1, Threshold_change = 0.05; the adjustment factors: Expiry_correction_factor = 0.5, Change_correction_factor = 0.3; and the balance coefficient: α = 0.7. Calculate the expiration adjustment factor: Expiry_adjustment = (Expiry_failure_rate > Threshold_expiry) ? Expiry_correction_factor×(Expiry_failure_rate-Threshold_expiry) / (1- Threshold_expiry): 0 Expiry_adjustment = 0.5×(0.15 - 0.1) / (1 - 0.1) = 0.5 × 0.05 / 0.9 = 0.0278. Calculate the change adjustment factor: Change_adjustment = (Change_failure_rate > Threshold_change) × Change_correction_factor × (Change_failure_rate - Threshold_change) / (1 - Threshold_change): 0 Change_adjustment = 0.3 × (0.08 - 0.05) / (1 - 0.05) = 0.3 × 0.03 / 0.95 = 0.0095. Calculate the composite failure impact factor: Compound_failure_factor = max(Expiry_adjustment, Change_adjustment) + α × min(Expiry_adjustment, Change_adjustment) = 0.0278 + 0.7 × 0.0095 = 0.0278 + 0.00665 = 0.0345.Apply the composite failure impact factor to adjust the validity period: TTL_final_RSSI = TTL_adjusted_RSSI × (1 - Compound_failure_factor) = 205.36 × (1 - 0.0345) = 205.36 × 0.9655 = 198.28 seconds; TTL_final_SINR = TTL_adjusted_SINR × (1 - Compound_failure_factor) = 189.61 × (1 - 0.0345) = 189.61 × 0.9655 = 183.07 seconds; TTL_final_NetParams = TTL_adjusted_NetParams × (1 - Compound_failure_factor) = 105.49 × (1 - 0.0345) = 105.49 × 0.9655 = 101.85 seconds.

[0107] 3.4 Comprehensive Signal Quality Assessment. Obtain a set of multi-dimensional monitoring indicators based on the monitoring level: Current signal strength (RSSI) = -76dBm. Since the current monitoring level is 1 (basic monitoring), only signal strength is obtained, and other indicators use historical values or default values: Signal quality (SINR) = 12dB (historical value); Network latency = 180ms (default value); Packet loss rate = 2% (default value). Normalize each indicator: RSSI_norm = (RSSI - RSSI_min) / (RSSI_max -RSSI_min) = (-76 - (-100)) / (-60 - (-100)) = 24 / 40 = 0.6; SINR_norm = (SINR -SINR_min) / (SINR_max - SINR_min) = (12 - 0) / (30 - 0) = 12 / 30 = 0.4; Delay_norm = 1 - (Delay - Delay_min) / (Delay_max - Delay_min) = 1 - (180 - 50) / (500 - 50) = 1 - 130 / 450 = 1 - 0.289 = 0.711; Loss_norm = 1 - (Loss - Loss_min) / (Loss_max- The signal quality index (SQI) is calculated by weighted calculation: SQI = w_RSSI × RSSI_norm + w_SINR × SINR_norm + w_Delay × Delay_norm + w_Loss × Loss_norm, where w_RSSI, w_SINR, w_Delay, and w_Loss are the weights of the indicators, set to 0.4, 0.3, 0.2, and 0.1, respectively, and their sum is 1. SQI = 0.4 × 0.6 + 0.3 × 0.4 + 0.2 × 0.711 + 0.1 × 0.9 = 0.24 + 0.12 + 0.1422 + 0.09 = 0.5922.

[0108] Step 4: Fast adaptation and communication control decision-making with small samples.

[0109] 4.1. Determine the initial parameters of the new environment. Assume that a new environment "Env004" is identified and its initial parameters need to be determined: 4.1.1. Similar Environment Retrieval. Use the fingerprint similarity calculation function to compare the network environment fingerprint of "Env004" with the historical environment fingerprints in the environment database: SimFunc(NF_Env004, NF_Env001) = 0.6821; SimFunc(NF_Env004, NF_Env002) = 0.7932; SimFunc(NF_Env004, NF_Env003) = 0.5547. Set the similarity threshold to 0.65 and select environments that exceed this threshold to form a similar environment set: Similar_envs = ["Env001", "Env002"].

[0110] 4.1.2. Validation Evaluation of Similar Environments. Parameter stability in similar environments was evaluated through cross-validation: the reliability score of Env001 (Reliability_score_Env001) = 0.85; the reliability score of Env001 (Reliability_score_Env001) = 0.85; and the reliability score of Env002 (Reliability_score_Env002) = 0.92. Calculate the comprehensive weight of each environment: Total_weight_Env001 = SimFunc(NF_Env004, NF_Env001) × Reliability_score_Env001 = 0.6821 × 0.85 = 0.5798; Total_weight_Env002 = SimFunc(NF_Env004, NF_Env002) × Reliability_score_Env002 = 0.7932 × 0.92 = 0.7297.

[0111] 4.1.3. Weighted fusion of similar environments. Extract the communication parameters of each environment from the similar environment set: Communication parameter set of Env001: Params_Env001 = {"SQI_weights": {"w_RSSI": 0.42, "w_SINR": 0.28, "w_Delay": 0.22, "w_Loss": 0.08}, "Decision_th": 0.62, "Monitor_level": 2, "TTL_factors": {"RSSI_factor": 0.95, "SINR_factor": 0.90, "NetParams_factor": 0.85}}. Communication parameter set for Env002: Params_Env002 = {"SQI_weights": {"w_RSSI": 0.38, "w_SINR": 0.32, "w_Delay": 0.18, "w_Loss": 0.12}, "Decision_th": 0.58, "Monitor_level": 1, "TTL_factors": {"RSSI_factor": 1.05, "SINR_factor": 0.92, "NetParams_factor": 0.88}}. Parameters are processed separately based on their type: Numeric parameters are first normalized, then weighted averaged, and finally denormalized. Taking the RSSI weight of SQI as an example: Normalization: Param_Env001_normalized = (w_RSSI_Env001 - Min_w_RSSI) / (Max_w_RSSI - Min_w_RSSI) = (0.42 - 0.2) / (0.6 -0.2) = 0.55; Param_Env002_normalized = (w_RSSI_Env002 - Min_w_RSSI) / (Max_w_RSSI - Min_w_RSSI) = (0.38 - 0.2) / (0.6 - 0.2) = 0.45.Weighted average: Param_init_normalized = (Total_weight_Env001 × Param_Env001_normalized + Total_weight_Env002 × Param_Env002_normalized) / (Total_weight_Env001 + Total_weight_Env002) = (0.5798 × 0.55 + 0.7297 × 0.45) / (0.5798 + 0.7297) = (0.3189 +0.3284) / 1.3095 = 0.6473 / 1.3095 = 0.4943. Denormalization: w_RSSI_init = Min_w_RSSI + Param_init_normalized × (Max_w_RSSI - Min_w_RSSI) = 0.2 + 0.4943 ×(0.6 - 0.2) = 0.2 + 0.4943 × 0.4 = 0.2 + 0.1977 = 0.3977 ≈ 0.40.

[0112] Similarly, calculate other numerical parameters: w_SINR_init = 0.30; w_Delay_init = 0.20; w_Loss_init = 0.10; and decision_th_init = 0.60. Boolean parameters (weighted voting): Monitor_level (discrete value, but treated like a Boolean): Env001: Monitor_level = 2, weight = 0.5798; Env002: Monitor_level = 1, weight = 0.7297; Weighted score: Level 1 score = 0.7297, Level 2 score = 0.5798; Since Level 1 has a higher score, Monitor_level_init = 1; Interval parameters: TTL_factors, apply weighted average to the upper and lower bounds: RSSI_factor_init = (0.5798 × 0.95 + 0.7297 × 1.05) / (0.5798 + 0.7297) = (0.5508 + 0.7662) / 1.3095 = 1.317 / 1.3095 = 1.006; SINR_factor_init = (0.5798 × 0.90 + 0.7297 × 0.92) / (0.5798+ 0.7297) = (0.5218 + 0.6713) / 1.3095 = 1.1931 / 1.3095 = 0.911; NetParams_factor_init = (0.5798 × 0.85 + 0.7297 × 0.88) / (0.5798 + 0.7297) = (0.4928+ 0.6421) / 1.3095 = 1.1349 / 1.3095 = 0.867.

[0113] 4.1.4. Adaptive adjustment to environmental characteristics. Adaptive adjustments are made based on the network environment fingerprint characteristics of "Env004": Since the signal strength of "Env004" is relatively stable (the coefficient of variation is less than 0.02), the RSSI weight is increased: w_RSSI_adjusted = w_RSSI_init × 1.05 = 0.40 × 1.05 = 0.42; then other weights are adjusted to ensure that the total is 1: Total adjustment = 0.42 - 0.40 = 0.02; w_SINR_adjusted = 0.30 - 0.02 × (0.30 / 0.60) = 0.30 - 0.01 = 0.29; w_Delay_adjusted = 0.20 - 0.02 × (0.20 / 0.60) = 0.20 - 0.0067 = 0.1933 ≈ 0.19; w_Loss_adjusted = 0.10 - 0.02 × (0.10 / 0.60) = 0.10 - 0.0033 = 0.0967 ≈ 0.10. The final initial parameter set is: Params_init_Env004 = {"SQI_weights": {"w_RSSI": 0.42, "w_SINR": 0.29, "w_Delay": 0.19, "w_Loss": 0.10}, "Decision_th": 0.60, "Monitor_level": 1, "TTL_factors": {"RSSI_factor": 1.006, "SINR_factor": 0.911, "NetParams_factor": 0.867}}.

[0114] 4.2. Optimize environment-specific parameters. "Env003" has a certain historical transmission record, and parameters can be optimized based on these records: 4.2.1. Initial parameter learning rate setting. Get the initial parameter learning rate: Base learning rate: Base_learning_rate = 0.05; parameter uncertainty index: Uncertainty_index = 0.15; initial learning rate: Learning_rate = Base_learning_rate × (1 + Uncertainty_index) = 0.05 × (1 + 0.15) = 0.0575.

[0115] 4.2.2 Parameter Update. Assume the statistics for the most recent 10 transmission results: ActualSuccess = 0.78; PredictedSuccess = 0.85; TargetRate = 0.82; FeatureValues = [0.62, 0.40, 0.75, 0.88] (RSSI_norm, SINR_norm, Delay_norm, Loss_norm). Update SQI weights: Calculate the error: Error = ActualSuccess - PredictedSuccess = 0.78 - 0.85 = -0.07; Weight adjustment formula: SQI_weights_new = SQI_weights_old + Learning_rate × Error × FeatureValues. Apply to each weight separately: w_RSSI_new = 0.40 + 0.0575 × (-0.07) × 0.62 = 0.40 - 0.0025 = 0.3975 ≈ 0.40; w_SINR_new = 0.30 + 0.0575 × (-0.07) × 0.40 = 0.30 - 0.0016 = 0.2984 ≈ 0.30; w_Delay_new = 0.20 + 0.0575 × (-0.07) × 0.75 = 0.20 - 0.0030 = 0.1970 ≈ 0.20; w_Loss_new = 0.10 + 0.0575 × (-0.07) × 0.88 = 0.10 - 0.0035 =0.0965 ≈ 0.10. Decision threshold update: Decision_th_new = Decision_th_old + Learning_rate × (TargetRate - ActualRate) = 0.65 + 0.0575 × (0.82 - 0.78) = 0.65 + 0.0575 × 0.04 = 0.65 + 0.0023 = 0.6523 ≈ 0.65.

[0116] 4.3. Transmission Timing Decision. The transmission decision score is calculated based on the current environment ID, the comprehensive signal quality index, and the priority of the data to be transmitted: Tscore = SQI × SQI_weight + Priority × Priority_weight; where SQI is the comprehensive signal quality index, with a value of 0.5922; SQI_weight is the signal quality weight, set to 0.6; Priority is the data priority, with a value of 8; MaxPriority is the highest priority, with a value of 10; and Priority_weight is the priority weight, set to 0.4. Tscore = 0.6 × 0.5922 + 0.4 × 8 / 10 = 0.35532 + 0.32 = 0.67532. The transmission decision score is compared with the decision threshold: Transmission Decision Score: Tscore = 0.67532; Decision Threshold: Decision_th = 0.65. Since Tscore (0.67532) > Decision_th (0.65), the state is determined to be transmittable.

[0117] 4.4. Environmental feedback and model update. Assuming the transfer execution result is successful, record the result of this transfer attempt and update the environmental model: 4.4.1. Update the Environmental Stability Score. The environmental stability score is updated using the decay memory mechanism: Stab_score_new = Decay_factor × Stab_score_old + (1 - Decay_factor) × Current_stability. Decay_factor is the decay factor, set to 0.8; Stab_score_old is the old stability score, set to 1.0; Current_stability is the current transmission stability, set to 1.0 due to successful transmission. Stab_score_new = 0.8 × 1.0 + 0.2 × 1.0 = 0.8 + 0.2 = 1.0.

[0118] 4.4.2. Update the transmission success rate. Similarly, update the transmission success rate: SuccessRate_new = Decay_factor × SuccessRate_old + (1 - Decay_factor) × Current_result; where Decay_factor is the decay factor, set to 0.8; SuccessRate_old is the old success rate, 0.78; Current_result is the current transmission result, 1 for success and 0 for failure. SuccessRate_new = 0.8 × 0.78 + 0.2 × 1 = 0.624 + 0.2 = 0.824.

[0119] 4.4.3. Update the optimal transmission threshold. Assuming that the current actual transmission success rate is close to the target success rate, fine-tune the optimal transmission threshold: OptimalTh_new = OptimalTh_old + Learning_rate × (TargetRate - SuccessRate_new) = 0.65 + 0.0575 × (0.82 - 0.824) = 0.65 + 0.0575 × (-0.004) = 0.65 - 0.00023 = 0.64977 ≈ 0.65.

[0120] 4.5. Resource optimization and transmission execution.

[0121] 4.5.1 Transmission Parameter Selection. Select the optimal transmission parameters based on the current environment: Transmission power level: Power_level = 3 (medium); Modulation and Coding Scheme: MCS_index = 12 (higher throughput); Maximum number of retransmissions: Max_retries = 2; Message segment size: Segment_size = 1024 bytes.

[0122] 4.5.2 Data Prioritization and Packing. The data set to be transmitted contains five data packets, sorted by priority: Data packet 1: priority 8, size 500 bytes; Data packet 2: priority 8, size 300 bytes; Data packet 3: priority 6, size 800 bytes; Data packet 4: priority 5, size 1200 bytes; Data packet 5: priority 3, size 200 bytes. Packing is performed based on segment size and priority: Transmission unit 1: contains data packet 1 and data packet 2 (total size 800 bytes, priority 8); Transmission unit 2: contains data packet 3 (size 800 bytes, priority 6); Transmission unit 3: contains the first 1024 bytes of data packet 4 (priority 5); Transmission unit 4: contains the remaining 176 bytes of data packet 4 and data packet 5 (total size 376 bytes, priority 5).

[0123] 4.5.3. Transmission Execution. Send transmission units in order of priority and record the transmission execution results: Transmission unit 1: Success, delay 125ms; Transmission unit 2: Success, delay 142ms; Transmission unit 3: Success, delay 156ms; Transmission unit 4: Success, delay 118ms. Transmission execution results: All successful, average delay 135.25ms, no packet loss.

[0124] This embodiment details the implementation process of the communication control method for the secure access module of the smart energy unit. Through specific data and calculation processes, the implementation effect is demonstrated: the environmental segmentation management mechanism based on topology changes successfully achieves refined environmental segmentation, decomposes environmental changes into three-dimensional features, and effectively handles changes in complex network environments; the small sample rapid adaptation algorithm achieves rapid adaptation to the new environment through weighted fusion of similar environments, shortening the environmental adaptation time from 26.7 minutes of traditional methods to 5.3 minutes; the construction and matching of multi-dimensional network environment fingerprints provides accurate environmental representation, which improves the communication success rate from 75.8% to 87.0% compared to traditional methods; the differentiated indicator validity period management achieves efficient utilization of monitoring resources and reduces energy consumption, with average power consumption reduced from 165mW to 129.3mW.

[0125] This invention accurately quantifies and classifies network environment changes, enabling smart energy units to distinguish between noise changes caused by signal fluctuations and topological changes caused by actual location movement. This avoids the frequent mis-segmentation or overlooking of important changes that occur with traditional methods based solely on signal strength. In practical applications, the environmental segmentation error rate was reduced from 28.5% with traditional methods to 6.3%, while recognition accuracy was increased by 22.7%. This enables smart energy units to accurately perceive and adjust their communication strategies in complex and changing network environments, improving communication reliability and stability. This approach is particularly suitable for scenarios where smart energy units move between different network coverage areas. It also addresses the "cold start" problem of smart energy unit communication systems, enabling rapid adaptation to new environments with minimal or even zero data samples. In practical applications, the adaptation time to a new environment was shortened by 78% from an average of 26.7 minutes with traditional adaptive methods to 5.3 minutes. The initial communication failure rate was also reduced by 69.4% from 39.5% to 12.1%. This system improves the self-starting capability and environmental adaptability of the smart energy unit communication system, making it particularly suitable for initial deployment or deployment in new network environments. It effectively addresses the limitation of traditional methods that require a large number of samples to build effective models. The network environment fingerprint automatically adjusts the importance of each dimension based on the actual environmental characteristics, improving the accuracy and robustness of environmental identification. In environments with large signal fluctuations but relatively stable network topology, the system automatically reduces the weight of signal strength and increases the weight of topological features, and vice versa. Experimental results show that the environmental identification accuracy rate increased from 73.2% (using the traditional single-signal strength method) to 92.5%, an improvement of 19.3 percentage points. Even in areas with drastic signal strength fluctuations at the edge of coverage, the recognition accuracy rate remained above 85%, providing an accurate and reliable environmental characterization foundation for subsequent environmental segmentation determination and communication parameter optimization. Dynamic adaptive adjustment of the environmental segmentation threshold is achieved, addressing the difficulty of traditional fixed threshold methods in adapting to diverse network environments. By using the communication success rate and the frequency of environmental changes as opposing adjustment factors, a bidirectional closed-loop adjustment is achieved, linking segmentation sensitivity with environmental stability and communication quality. In practical applications, the over-segmentation and under-segmentation rates of environmental segmentation were reduced from 22.3% and 17.5% in traditional methods to 5.4% and 3.9%, respectively, representing reductions of 75.8% and 77.7%, improving the accuracy of environmental segmentation. Furthermore, the system maintained efficient operation both in edge coverage areas with severe signal fluctuations and in central coverage areas with stable signals, reducing the communication failure rate caused by environmental misjudgment and improving the overall adaptability and stability of the smart energy unit communication system. By analyzing the dynamic characteristics of the environment, differentiated validity period calculation formulas were designed for different monitoring indicators, achieving an optimal balance between monitoring resources and communication reliability.In practical applications, compared to traditional fixed-period monitoring methods, while maintaining the same communication success rate, this approach reduced monitoring operations by 43.6%, lowered average power consumption by 21.6% (from 165mW to 129.3mW), and extended battery life by 27.5%. This approach is particularly suitable for battery-powered smart energy units, significantly extending device operating time while maintaining communication reliability. This resolves the conflict between monitoring frequency and battery life in traditional methods, improving the sustainability and stability of the entire energy management system. It achieves closed-loop optimization of validity period management, resolving the issue of over- or under-adjustment caused by confusing different types of failure causes in traditional validity period adjustment mechanisms. This approach avoids over-adjustment caused by simple addition, enabling the system to implement more targeted adjustment strategies for different types of transmission failure causes. Experimental data shows that the transmission failure rate caused by indicator expiration was reduced from 15% to 4.2%, and the transmission failure rate caused by indicator changes was reduced from 8% to 2.5%. Furthermore, it effectively avoids over-monitoring and reduces the proportion of ineffective monitoring from 26.4% to 8.3%. It performs particularly well in complex and changing network environments, enabling the smart energy unit communication system to more accurately predict the optimal time to update indicators, maintain efficient operation under various network conditions, and improve the system's adaptability and resource utilization efficiency.

[0126] The preferred embodiments of the present invention are described in detail above. However, the present invention is not limited to the specific details in the above embodiments. Within the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all fall within the scope of protection of the present invention.

Claims

1. A communication control method for a smart energy unit secure access module, characterized in that: include: Obtain the currently connected base station information and adjacent cell information, and construct the network environment fingerprint and similarity function; Based on the similarity function and the current and historical network environment fingerprints, the degree of environmental change is calculated, the type of environmental change is determined, and environmental parameters are generated; Combine it with the priority of the data to be transmitted to determine the monitoring level, obtain a multi-dimensional monitoring indicator set, and calculate the comprehensive signal quality index; Based on environmental parameters, comprehensive signal quality index and priority of data to be transmitted, the transmission decision score is calculated, the transmission decision result is generated and the transmission is executed to obtain the transmission execution result.

2. The method according to claim 1, characterized in that Construct network environment fingerprint and similarity function, including: Obtain the ID, signal strength and quality of the currently connected base station to form a basic network parameter set; Read the neighboring network parameter set, including the list of neighboring cells that the device can perceive and their corresponding signal strength list; Construct a network environment fingerprint based on the basic network parameter set and the adjacent network parameter set; Based on the matching degree of the primary base station ID, the cosine similarity of the neighboring cell lists, and the signal strength difference, a similarity function is established to evaluate the similarity between network environment fingerprints.

3. The method according to claim 1, characterized in that Generate environment parameters, including: Compare the historical network environment fingerprints of the current and previous time periods, and calculate the environmental change degree, which comprehensively considers changes in signal strength, primary base station switching, and changes in adjacent cell topology. If the degree of change is greater than a preset threshold, it indicates a significant environmental change, and a new environmental segment is created and assigned an ID. Otherwise, it indicates a continuous environmental change, and the current environmental segment information is updated, forming an updated network environment database. Every predetermined period, the similarity function is used to calculate the similarity between the current network environment fingerprint and the network environment fingerprints in the network environment library, obtain the most similar environment ID and similarity score, and use this to judge whether the current environment belongs to a known environment or a new environment, and generate the current environment ID and environment stability score as environment parameters.

4. The method according to claim 1, wherein Calculates a comprehensive signal quality index, including: The initial monitoring level is determined based on the current environment ID and environment stability score in the environmental parameters. Based on this, the monitoring trigger function is called to decide whether to trigger a higher level of monitoring based on the current signal strength, environment stability score, and priority of the data to be transmitted, thereby obtaining the actual monitoring level. Differentiated validity periods are set for monitoring indicators based on the signal variation coefficient and environment change frequency. A multi-dimensional monitoring indicator set is obtained according to the monitoring level actually performed, and is normalized and weighted to obtain a comprehensive signal quality index.

5. The method according to claim 3, characterized in that Get the transfer execution results, including: When a new environment is identified, similar environments are searched in the network environment library and parameter sets are extracted. The initial parameter set is calculated based on similarity weighting. The signal quality assessment parameters and decision thresholds are optimized based on historical transmission result records. Calculate the transmission decision score based on the signal quality evaluation parameters, the comprehensive signal quality index, and the priority of the data to be transmitted, and compare it with the decision threshold to make a decision on whether to transmit or wait; If transmission is required, the optimal transmission parameters are selected based on the current environment characteristics and the data to be transmitted, the data to be transmitted is prioritized and appropriately packaged, the transmission is executed, and the transmission execution results are obtained.

6. The method according to claim 2, characterized in that Construct a network environment fingerprint, including: Read the basic network parameter set and the adjacent network parameter set, construct the original network dataset, extract key features from it, analyze the stability, and obtain the feature reliability weight; Based on key features and feature reliability weights, a structured network environment fingerprint is constructed, compressed, and optimized to generate an optimized fingerprint. Calculate the similarity between the optimized fingerprint and the historical network environment fingerprint within a predetermined period to verify the recognition effectiveness of the fingerprint; The optimized fingerprint that passes the verification is stored in the historical fingerprint set as the final network environment fingerprint.

7. The method according to claim 3, characterized in that Determining the type of environmental change also includes: The environmental change degree is decomposed into signal strength change components, main base station change flags, and cell topology change rates to construct an environmental change feature vector. This feature vector is combined with the historical environmental stability index to evaluate the actual impact of environmental changes on communications and obtain the environmental change significance. Based on historical records of environmental changes and communication success rate changes, the environmental segmentation threshold is calculated and compared with the significance of environmental changes. Environmental changes are divided into three categories: major changes, minor changes, and noise, and whether to trigger environmental segmentation is determined accordingly.

8. The method according to claim 4, characterized in that Set differentiated validity periods for monitoring indicators, including: Based on the historical network environment fingerprint set corresponding to the current environment ID, the short-term and long-term coefficients of variation of signal strength, the frequency of network topology changes, and the frequency of master base station switching are analyzed to form an environment dynamic characteristic vector. Based on this, different basic validity periods are calculated for signal strength, signal quality, and network parameters. Read the highest priority of the data to be transmitted and the current energy status of the device, adjust the basic validity period, and obtain the optimized validity period; Based on the transmission failure rate caused by indicator expiration and indicator changes during the validity period in historical transmission records, the comprehensive failure impact coefficient is calculated, and the optimized validity period is further adjusted to obtain the final validity period.

9. The method according to claim 5, characterized in that The initial parameter set is calculated based on similarity weighting, including: Using a similarity function, the current network environment fingerprint is compared with the historical network environment fingerprints in the network environment library. Environments with similarities exceeding a preset threshold are selected to form a similar environment set and parameter stability is verified. The reliability score and similarity score of each environment are calculated and multiplied to obtain a comprehensive weight. The communication parameter sets of each environment are extracted from the similar environment set, and weighted calculations are performed using the corresponding comprehensive weights to obtain the weighted parameter sets; they are adaptively adjusted according to the characteristics of the current network environment fingerprint to form the initial parameter set.

10. The method according to claim 7, characterized in that Calculates environmental segmentation thresholds, including: Read the basic threshold of the system configuration as the initial calculation basis for the environmental segmentation threshold; Calculate the average communication success rate within a predetermined time window based on historical communication records and generate a success rate adjustment factor; Analyze historical environmental change records, calculate the frequency of environmental changes per unit time, and generate a change frequency adjustment factor; Taking into account the stability, importance and typical scene characteristics of the current environment, an environment-specific adjustment factor is generated; The final adaptive environment segmentation threshold is calculated by multiplying the basic threshold by (1 + change frequency adjustment factor), (1 - success rate adjustment factor) and the environment-specific adjustment factor.

Images