Anti-interference transmission optimization method and system for building structure health monitoring signals

Abstract

The application relates to the technical field of anti-interference transmission, in particular to an anti-interference transmission optimization method and system for building structure health monitoring signals. The method collects building structure dynamic response signals through distributed wireless sensors and completes digital processing; a lightweight imitation learning model is deployed at a local control end to identify interference modes of a transmission channel of monitoring data and extract interference features; an adaptive anti-interference transmission strategy is generated according to the interference features, the monitoring data is stably transmitted to a central server to realize multi-node data fusion and health state evaluation; finally, the imitation learning model is updated in parameters according to a health evaluation result returned by the server, realizing high-reliability anti-interference transmission and adaptive optimization of building monitoring signals. The application improves the effective transmission rate and data integrity rate of building structure health monitoring signals in a complex interference environment, effectively improving signal transmission reliability and intelligent level.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of anti-interference transmission, in particular to an anti-interference transmission optimization method and system for building structure health monitoring signals. BACKGROUND

[0002] In related technologies, with the rapid development of smart cities and large-scale construction projects, building structure health monitoring has become an important technical means to ensure building safety and prolong service life. Such systems are usually composed of distributed wireless sensor networks, which are used to collect dynamic response signals of building structures in real time, such as strain, acceleration, displacement, and modal parameters, etc., and then transmit them to the central monitoring server through wireless channels. Through analysis and judgment, it can determine whether the structure has problems such as stiffness degradation, connection relaxation, or potential damage.

[0003] However, the existing building structure health monitoring method faces significant anti-interference challenges during the wireless transmission of signals. The construction site environment is complex, with multiple sources of electromagnetic interference, including construction machinery, elevator drive devices, industrial wireless device radiation, and multipath reflection effects. These factors can cause transmission signal frequency band conflicts, signal attenuation, synchronization deviations, and even data loss, which can seriously affect the real-time performance and accuracy of the monitoring system. In addition, traditional anti-interference strategies rely on centralized server-side computing and scheduling, with node-side only performing fixed power and frequency parameter configuration, which cannot adaptively adjust to dynamic interference environments, making it difficult for the system to balance real-time performance and energy consumption control under large-scale distributed deployment conditions.

[0004] In summary, the technical problems in related technologies need to be improved. SUMMARY

[0005] The main purpose of the embodiments of the present application is to propose an anti-interference transmission optimization method and system for building structure health monitoring signals, which can improve the effective transmission rate and data integrity rate of building structure health monitoring signals in complex interference environments, and effectively improve the signal transmission reliability and intelligent level.

[0006] To achieve the above purpose, one aspect of the embodiments of the present application proposes an anti-interference transmission optimization method for building structure health monitoring signals. The method is applied to a local control end, which is connected to multiple distributed wireless sensors to form a node. The control ends in multiple nodes are connected to a central server through a wireless network. The method includes the following steps:

[0007] Obtaining dynamic response signals of building structures collected by distributed wireless sensors in real time;

[0008] Converting the dynamic response signals into digital signals to form a building structure monitoring data set;

[0009] deploying a lightweight imitation learning model;

[0010] performing interference pattern recognition on a transmission channel corresponding to the building structure monitoring dataset based on the lightweight imitation learning model to generate an interference feature;

[0011] generating an anti-interference transmission strategy using the lightweight imitation learning model according to the interference feature;

[0012] transmitting the building structure monitoring dataset to a central server based on the anti-interference transmission strategy, so that the central server performs fusion analysis on building structure monitoring datasets of multiple nodes to obtain a building structure health evaluation result;

[0013] performing parameter updating on the lightweight imitation learning model based on the building structure health evaluation result returned by the central server.

[0014] In some embodiments, the dynamic response signal includes a strain signal, a vibration acceleration signal, a displacement signal, and a modal parameter signal generated during operation of the building structure.

[0015] In some embodiments, the lightweight imitation learning model is constructed based on a long short-term memory network and is trained at the end side using a historical interference index as a supervision signal.

[0016] In some embodiments, the interference pattern recognition on the transmission channel corresponding to the building structure monitoring dataset based on the lightweight imitation learning model to generate an interference feature includes:

[0017] measuring the received signal strength indication and the noise power of each frequency band in the building structure monitoring dataset one by one with a preset detection period as a window to obtain a target time domain signal;

[0018] converting the target time domain signal into a frequency domain energy spectrum through fast Fourier transform;

[0019] identifying a candidate interference frequency point in the frequency domain energy spectrum;

[0020] performing interference persistence discrimination on the candidate interference frequency point, and calculating an interference index based on the interference persistence discrimination result;

[0021] inputting the interference index into the lightweight imitation learning model to output an interference feature representing a current channel state; the interference feature includes an interference frequency distribution, an interference intensity level, and an interference duration interval.

[0022] In some embodiments, the interference index is calculated through a weighted summation formula as follows:

[0023] ;

[0024] wherein, is the interference duration of the frequency band is the total duration of the current detection period; represents the signal strength; M represents the total number of available frequency bands.

[0025] In some embodiments, the generating an anti-interference transmission strategy based on the interference feature using the lightweight imitation learning model comprises:

[0026] dynamically adjusting the transmission power based on the interference feature, the current battery power, and the transmission task priority to generate a transmission power parameter;

[0027] optimizing the initial frequency hopping sequence based on the interference frequency distribution corresponding to the interference feature to generate a frequency hopping sequence;

[0028] determining a selective retransmission mechanism based on the transmission quality coefficient corresponding to the interference feature;

[0029] outputting an anti-interference transmission strategy based on the joint optimization of the transmission power parameter, the frequency hopping sequence, and the retransmission mechanism using the lightweight imitation learning model.

[0030] In some embodiments, the determining a selective retransmission mechanism based on the transmission quality coefficient corresponding to the interference feature comprises:

[0031] calculating a transmission quality coefficient of the current transmission channel based on the interference feature;

[0032] determining a retransmission threshold based on the transmission quality coefficient, and triggering the selective retransmission mechanism when the transmission quality is lower than the retransmission threshold;

[0033] dividing the data to be transmitted into a plurality of data blocks of fixed length, and configuring a check code for each data block based on the selective retransmission mechanism;

[0034] performing integrity verification on the data blocks respectively, and if any data block fails the verification, performing retransmission only on the data block;

[0035] In the retransmission process, dynamically adjusting the retransmission proportion and the retransmission priority of the selective retransmission mechanism according to the transmission quality coefficient.

[0036] In some embodiments, the transmission quality coefficient is calculated based on the signal-to-interference ratio, the channel noise coefficient, and the channel delay spread parameter.

[0037] ​In some embodiments, the retransmission priority includes a high-level retransmission priority, a medium-level retransmission priority, or a low-level retransmission priority, the high-level retransmission priority is set based on a modal parameter related data block, the medium-level retransmission priority is set based on a strain and acceleration data block, and the low-level retransmission priority is set based on an environmental quantity or auxiliary data block.

[0038] To achieve the above object, another aspect of the embodiment of the present application provides an anti-interference transmission optimization system for building structure health monitoring signals. The system is arranged in a local control end. The local control end is connected with a plurality of distributed wireless sensors to form a node. The control ends in the plurality of nodes are connected with a central server through a wireless network. The system comprises:

[0039] A data acquisition module is configured to acquire dynamic response signals of a building structure in real time based on the distributed wireless sensors to form a building structure monitoring data set.

[0040] A first construction module is configured to deploy a lightweight imitation learning model.

[0041] An interference identification module is configured to identify interference patterns of transmission channels corresponding to the building structure monitoring data set based on the lightweight imitation learning model to generate interference features.

[0042] An anti-interference transmission strategy generation module is configured to generate an anti-interference transmission strategy based on the interference features and the lightweight imitation learning model.

[0043] A transmission control module is configured to transmit the building structure monitoring data set to the central server based on the anti-interference transmission strategy, so that the central server performs fusion analysis on the building structure monitoring data sets of the plurality of nodes to obtain a building structure health evaluation result.

[0044] A model update module is configured to update parameters of the lightweight imitation learning model based on the building structure health evaluation result returned by the central server.

[0045] To achieve the above object, another aspect of the embodiment of the present application provides an electronic device. The electronic device comprises a memory and a processor. The memory stores a computer program. The processor implements the above method when executing the computer program.

[0046] To achieve the above object, another aspect of the embodiment of the present application provides a computer readable storage medium. The computer readable storage medium stores a computer program. The computer program is executed by a processor to implement the above method.

[0047] To achieve the above object, another aspect of the embodiment of the present application provides a computer program product comprising a computer program which, when executed by a processor, implements the above method.

[0048] The embodiment of the present application at least has the following beneficial effects: the present application provides an anti-interference transmission optimization method and system for building structure health monitoring signals, which can adaptively extract interference features in different environments through interference mode identification of the transmission channel, and generate targeted anti-interference transmission strategies based on a lightweight imitation learning model, thereby effectively suppressing interference problems such as channel noise, spectral overlap and multipath fading, and significantly improving the transmission quality and stability of building structure monitoring signals. The lightweight imitation learning model is deployed locally at the wireless sensor node, enabling the node to have real-time learning and parameter updating capabilities. The model can dynamically optimize its own parameters according to the building structure health evaluation results, thereby continuously improving the adaptability and generalization ability of the anti-interference strategy, realizing a distributed self-optimizing transmission mechanism and reducing the dependence on the central server. Through fusion analysis of the target transmission signal, the monitoring results of multiple nodes can be integrated to reduce data distortion and packet loss problems caused by interference, ensuring the integrity and accuracy of the health evaluation data and providing high-confidence support for rapid judgment of the safety state of the building structure. The imitation learning model structure is lightweight and suitable for local inference and strategy generation by resource-constrained wireless nodes, without the need for complex cloud computing and high-bandwidth communication, significantly reducing node energy consumption and system maintenance costs, and facilitating stable operation in long-term building structure monitoring scenarios. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 is a flowchart of an anti-interference transmission optimization method for building structure health monitoring signals provided by the embodiment of the present application;

[0050] Figure 2 is Figure 1 a flowchart of step S5 in

[0051] Figure 3 is a module schematic diagram of an anti-interference transmission optimization system for building structure health monitoring signals provided by the embodiment of the present application. DETAILED DESCRIPTION

[0052] In order to make the purposes, technical solutions and advantages of the present application clearer, further described are the present application in combination with the accompanying drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and not intended to limit the present application. When the following description refers to the accompanying drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with embodiments of the present application. They are merely examples of apparatuses and methods consistent with some aspects of the embodiments of the present application as detailed in the appended claims.

[0053] It can be understood that the terms “first”, “second” and the like used in the present application can be used herein to describe various concepts, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another concept. For example, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as the first information, without departing from the scope of the embodiments of the present application. Depending on the context, the word “if” as used herein can be interpreted as “when” or “upon determination” or “in response to determining”.

[0054] The terms “at least one”, “multiple”, “each”, “any” and the like used in the present application include one, two or more than two, multiple includes two or more than two, each refers to each of the corresponding multiple, and any refers to any one of the multiple.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0056] The embodiment of the application provides an anti-interference transmission optimization method and system for building structure health monitoring signals, which can adaptively extract interference characteristics in different environments through interference mode identification of a transmission channel, and generate a targeted anti-interference transmission strategy based on a lightweight imitation learning model, thereby effectively suppressing interference problems such as channel noise, spectral overlap and multipath fading, and significantly improving the transmission quality and stability of building structure monitoring signals. A lightweight imitation learning model is deployed locally in the wireless sensor node, enabling the node to have real-time learning and parameter updating capabilities. The model can dynamically optimize its parameters according to the building structure health evaluation results, thereby continuously improving the adaptability and generalization ability of the anti-interference strategy, implementing a distributed self-optimizing transmission mechanism, and reducing the dependence on the central server. Through fusion analysis of the target transmission signal, the monitoring results of multiple nodes can be integrated to reduce data distortion and packet loss caused by interference, ensuring the integrity and accuracy of the health evaluation data and providing high-confidence support for rapid judgment of the safety state of the building structure. The lightweight design of the imitation learning model structure is suitable for resource-constrained wireless nodes for local reasoning and strategy generation, without the need for complex cloud computing and high-bandwidth communication, significantly reducing node energy consumption and system maintenance costs, and facilitating stable operation in long-term building structure monitoring scenarios.

[0057] The anti-interference transmission optimization method for building structure health monitoring signals provided by the embodiment of the application relates to the field of anti-interference transmission. The anti-interference transmission optimization method for building structure health monitoring signals provided by the embodiment of the application can be applied in a terminal, can be applied in a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted terminal, and the like, but is not limited thereto; the server end can be configured as a standalone physical server, can be configured as a server cluster or a distributed system composed of multiple physical servers, can be configured as a cloud server providing basic cloud computing services such as cloud service, cloud database, cloud computing, cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN, and big data and artificial intelligence platform, and can also be a node server in a blockchain network; the software can be an application that implements an anti-interference transmission optimization method for building structure health monitoring signals, and the like, but is not limited to the above forms.

[0058] The application is operable with numerous general purpose or special purpose computer system environments or configurations. Examples of well-known computing systems, environments, and / or configurations that can be suitable for use with the application include personal computers, server computers, handheld or laptop devices, tablet devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like. The application can be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and the like, that perform particular tasks or implement particular abstract data types. The application can also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules can be located in both local and remote computer storage media including memory storage devices.

[0059] In the embodiment of the application, the embodiment can be applied to the control end of a distributed wireless sensor node, the control end is connected with a plurality of wireless sensors in the distributed wireless sensor node respectively, and the plurality of control ends are connected with a center server. Specifically, the wireless sensor in the embodiment is used to collect a dynamic response signal of a building structure in a corresponding node, each control end is pre-deployed with a lightweight imitation learning model, each node corresponding control end generates an interference feature based on interference mode recognition of a building structure monitoring data set in the node through the lightweight imitation learning model, and generates an anti-interference transmission strategy based on the interference feature, and then transmits the building structure monitoring data set to the center server wirelessly based on the anti-interference transmission strategy. The center server fuses and analyzes the building structure monitoring data set sent by the plurality of node corresponding control ends to obtain a building structure health evaluation result.

[0060] It can be understood that the lightweight imitation learning model deployed by each control end in the embodiment can perform incremental learning based on the real-time acquired dynamic response signal of the building structure to update the lightweight imitation learning model parameters, and can also perform iterative learning based on the building structure health evaluation result fed back by the center server to update the lightweight imitation learning model parameters.

[0061] Exemplarily, taking a multi-story commercial building as an example, the embodiment can deploy a control terminal in each floor, distribute a plurality of sensors in the floor corresponding to the control terminal, and connect the plurality of sensors with the corresponding control terminal to form a node. The control terminals in the plurality of nodes are connected with the center server. The control terminal in each node can pre-deploy a lightweight imitation learning model. After the lightweight imitation learning model in the control terminal generates an anti-interference transmission strategy based on the data collected by the distributed sensors, the control terminal sends the data collected by the distributed sensors to the center server based on the anti-interference transmission strategy. The center server analyzes the data of the plurality of nodes to obtain the building structure health evaluation result.

[0062] Taking the control terminal in each node as an example, Figure 1 is an optional flowchart of a building structure health monitoring signal anti-interference transmission optimization method provided by the embodiment of the present application, Figure 1 The method in the embodiment can include, but is not limited to, steps S1 to S7:

[0063] S1: Obtain the dynamic response signal of the building structure collected by the distributed wireless sensor in real time; wherein the dynamic response signal includes strain signal, vibration acceleration signal, displacement signal and modal parameter signal generated in the operation process of the building structure.

[0064] In the embodiment, in the large building structure monitoring scene, the distributed wireless sensor node is installed in the high sensitive area such as beam end, column foot, floor shear wall core tube, steel structure node connection position, etc. Strain gauge, three-axis MEMS accelerometer, displacement meter and modal feature auxiliary extraction unit are built-in each node; the node collects the dynamic response signal of the building structure under the conditions of natural vibration, environmental load, wind load, operation disturbance and accidental impact at a sampling rate of ≥1 kHz in real time.

[0065] The dynamic response signal includes:

[0066] Strain signal: used to reflect the bending and stress distribution change of the member, and is the data most easily reflecting the connection relaxation or local crack initiation;

[0067] Vibration acceleration signal: used to reflect the dynamic response and energy transmission process of the overall structure;

[0068] Displacement signal: used to represent the macro deformation state of the member, especially the interlayer lateral displacement and flexibility change;

[0069] Modal parameter signal: including natural frequency, mode shape vector, damping ratio, etc., obtained online in a fast Fourier transform (FFT) and multi-point time-space cross-correlation fusion manner.

[0070] S2: Convert the dynamic response signal into a digital signal to form a building structure monitoring data set;

[0071] In this embodiment, the sensor node performs high-precision digitization processing on the original analog signal. The analog-to-digital conversion unit (ADC) built-in each node adopts a 24-bit resolution sampler to convert the continuous analog waveform signal into a standardized digital signal stream. The system organizes the collected signals into a unified format data frame through a hierarchical caching mechanism (ring buffer + queue caching), generating a building structure monitoring data set.

[0072] The data frame includes a frame header (node ID, timestamp, signal type, priority identifier), a frame body (multidimensional signal sequence), and a check code (CRC16). The data is stored in the node local cache after light compression encoding and enters the subsequent interference recognition and anti-interference transmission decision-making process.

[0073] S3: deploying a lightweight imitation learning model; wherein the lightweight imitation learning model is constructed based on a long short-term memory network and trained at the end side by taking the historical interference index as a supervision signal; the long short-term memory network includes a unidirectional long short-term memory network unit and a fully connected output layer.

[0074] In this embodiment, a lightweight imitation learning model is embedded on the wireless sensor node side. The model takes a simplified long short-term memory network as the main body, only uses a unidirectional LSTM unit and a fully connected output layer, and the size of the hidden layer is not more than 50 nodes, so that the model can be directly run on a low-power MCU (typical 32-bit low-frequency embedded microprocessor), without relying on central server computing resources or requiring excessive edge computing collaboration hardware.

[0075] The training mode of the lightweight imitation learning model adopts a "imitation learning" method, that is, the training target is not to directly classify interference, but to learn the decision-making rules of the central expert model fitting the transmission strategy under different interference scenarios; therefore, the model is not to judge whether the interference is strong or weak, but to directly predict the strategy reaction trend under the interference state, so that the end-side node has the ability of "recognition and decision-making", thereby significantly shortening the anti-interference strategy formation path.

[0076] At the input side of the lightweight imitation learning model, the channel state features detected by the node in real time are read, including channel energy distribution, interference frequency band coverage, interference duration characteristics, sudden energy information, and other channel time sequence variables, and an interference index and a strategy reaction trend are formed in a short time. The model continuously learns the central feedback data in an incremental online manner at the node end, so that it continuously strengthens the adaptability to the specific building site-specific interference environment (such as construction electric welder, curtain wall external elevator, elevator operation cycle interference, etc.) locally.

[0077] Because the model is deployed on the node, the inference process has millisecond-level latency, enabling policy decisions to be completed within the instantaneous period of interference occurrence. This transforms the generation of anti-interference policies from a centrally pushed model to a node-adaptive generation model, avoiding the accumulation of anti-interference latency caused by central processing / network round trips in traditional solutions. Simultaneously, the model's lightweight structure and small number of state parameters enable nodes to possess stable, continuous, and low-power local anti-interference policy generation capabilities during long-term operation, providing a feasible guarantee for continuous online operation of large-scale building monitoring.

[0078] Specifically, the historical interference data used in training the lightweight imitation learning model includes the interference frequency distribution, interference intensity, and interference duration of the channel in different time periods. Parameter optimization is performed by minimizing the mean square error (MSE) between the predicted and actual interference indices. The specific training objective function is as follows:

[0079] ;

[0080] in, This represents the loss function value during the training phase of a lightweight imitation learning model; The total number of training samples, The model predicts the first Interference index at each moment This represents the actual interference index. Regularization coefficient ( ≥0), For the model number The model has several weight parameters; its computational complexity is low, making it suitable for the real-time requirements of low-power sensor nodes.

[0081] S4: Based on a lightweight imitation learning model, interference pattern recognition is performed on the transmission channel corresponding to the building structure monitoring dataset to generate interference features;

[0082] Among them, interference pattern recognition is performed on the transmission channel corresponding to the building structure monitoring dataset based on a lightweight imitation learning model to generate interference features, including:

[0083] S41: Using a preset detection period as a window, measure the received signal strength indication and noise power of each frequency band one by one in the building structure monitoring dataset to obtain the target time domain signal;

[0084] S42: Convert the target time-domain signal into a frequency-domain energy spectrum using Fast Fourier Transform;

[0085] S43: Identify candidate interference frequency points in the frequency domain energy spectrum;

[0086] S44: Determine the interference persistence of candidate interference frequency points and calculate the interference index based on the interference persistence determination results;

[0087] The interference index is input into the lightweight imitation learning model, and an interference feature representing the current channel state is output; wherein the interference feature includes interference frequency distribution, interference intensity level and interference time interval.

[0088] In this embodiment, the distributed wireless sensor node performs interference mode recognition on the wireless transmission channel corresponding to the building structure monitoring data set based on the aforementioned lightweight imitation learning model, and generates a quantized interference feature representing the current channel interference state, which is used for subsequent anti-interference strategy generation.

[0089] Firstly, the node locally performs multi-band signal strength analysis (MSSA). The control end in the node measures the received signal strength indication (RSSI) and noise power of each frequency band in the current available monitoring frequency band set in the building structure monitoring data set with a fixed detection period (e.g. 500 ms) as the window. Subsequently, the time domain signal is converted into frequency domain energy spectrum by fast Fourier transform (FFT) to separate the energy distribution of different frequency components, realizing preliminary frequency spectrum characterization of external electromagnetic interference. For the frequency band with energy mutation or abnormal power peak, the system automatically marks it as a "candidate interference frequency point".

[0090] Secondly, the node discriminates the interference persistence of the candidate interference frequency point and calculates the interference index. When the signal power of a certain frequency band continuously exceeds the preset intensity threshold, the node records the interference duration of the frequency band and calculates the comprehensive interference index according to the interference amplitude and time length. This process considers different types of characteristics such as narrowband interference, wideband interference, multi-point interference and periodic interference. Through the time weighting window mechanism, the node can distinguish between burst interference and persistent interference, preventing short-term signal jitter from being misjudged as a high-intensity interference event.

[0091] Then, the node inputs the real-time calculated interference index into the lightweight imitation learning model as an input variable. The model internally extracts the variation law of the interference signal in the time series dimension and the spatial distribution characteristics of the interference intensity through imitation learning of historical interference samples. In the reasoning stage, the model no longer performs complex multi-layer convolution or feature decomposition, but directly outputs an interference feature vector through a lightweight time series memory structure. The vector comprehensively represents the current channel state, including:

[0092] Interference frequency distribution: describes the distribution form and concentration degree of interference in the frequency spectrum;

[0093] Interference intensity level: represents the influence degree of interference signal on channel quality in the form of quantized level (such as low, medium and high);

[0094] Interference time interval: used to distinguish between continuous interference, intermittent interference and burst interference, supporting subsequent dynamic power control and frequency hopping strategy decision.

[0095] Meanwhile, the model also records the difference between the interference characteristics and the internal historical memory state during each inference process, forming a lightweight local interference sample cache to provide raw support data for subsequent model parameter adaptive updates in S7.

[0096] Specifically, in the preset frequency band set , ( , for each frequency band ; if exceeds the preset threshold , it is determined as an interference frequency band and is added to the set ; the interference index is calculated by the weighted summation formula as follows:

[0097] ;

[0098] wherein is the interference duration of the frequency band , and is the total duration of the current detection period; represents the signal strength, and M represents the total number of available frequency bands. This method quantifies the dynamic characteristics of the interference frequency band, providing accurate input basis for subsequent frequency hopping sequence optimization.

[0099] In terms of interference types, this method mainly optimizes frequency domain detectable interference (such as narrowband interference and multi-frequency point interference). Its detection mechanism is based on the spectrum analysis of MSSA and FFT, which can effectively identify the strength and duration of the interference frequency band. However, for wideband burst interference and inter-symbol interference caused by multipath effects, this method indirectly represents it by introducing a time weighting factor in the interference index , and the correction formula is as follows:

[0100] ;

[0101] wherein is the burst interference strength factor (obtained by short-time energy mutation detection), is the empirical weight. This extension illustrates the applicability of the method in a multi-class interference environment, while avoiding excessive generalization technical assumptions.

[0102] S5: According to the interference characteristics, a lightweight imitation learning model is used to generate an anti-interference transmission strategy;

[0103] ​In this embodiment, to ensure the adaptability of the signal transmission process to structural monitoring, the method introduces a monitoring data priority parameter in the frequency hopping optimization and selective retransmission mechanism :

[0104] ;

[0105] It should be noted that 1 is low priority, and the data type is environmental quantity or auxiliary data block; 2 is medium priority, and the data type is strain and acceleration data block; 3 is high priority, and the data type is modal parameter related data block.

[0106] As shown in the reference Figure 2 , according to the interference characteristics, a lightweight imitation learning model is used to generate an anti-interference transmission strategy, including:

[0107] S51: Based on the interference characteristics, the current battery power and the transmission task priority are combined to dynamically adjust the transmission power, and a transmission power parameter is generated;

[0108] In this embodiment, first, according to the interference intensity level and interference duration interval extracted from the interference characteristics, combined with the current battery power state and transmission task priority of the node, adaptive adjustment of the transmission power is performed.

[0109] When high intensity or persistent interference is detected, the model adjusts the transmission power parameter according to the interference intensity weight factor to enhance the signal penetration capability; on the contrary, in a low interference or short time interference environment, the model automatically reduces the transmission power to save energy.

[0110] In addition, according to different task priorities (modal parameter data > strain and acceleration data > environmental quantity data), a differentiated power allocation strategy is executed:

[0111] For high-priority modal parameter data blocks, maintain a high transmission power to ensure that critical monitoring information can still be stably received by the central end under high interference conditions;

[0112] For medium-priority strain and acceleration data blocks, adopt medium power output to balance reliability and energy efficiency;

[0113] For low-priority environmental auxiliary data blocks, actively delay or reduce power transmission when interference is obvious to save node power resources.

[0114] Through this task-adaptive power adjustment strategy, interference suppression and energy consumption balance are achieved under resource-constrained conditions.

[0115] S52: Based on the interference frequency distribution corresponding to the interference characteristics, the initial frequency hopping sequence is optimized to generate a frequency hopping sequence;

[0116] In this embodiment, the original frequency hopping sequence is optimized according to the interference frequency distribution identified in the interference feature.

[0117] First, the frequency interval identified as the interference frequency band is removed from the available frequency band set to generate a preliminary interference-free frequency hopping set;

[0118] Then, the node generates an independent disturbance offset according to its ID and timestamp, and performs offset mapping on the interference-free frequency hopping set through a hash disturbance function to form a unique frequency hopping sequence of the node. This optimization process not only avoids collision of multiple nodes in the same frequency band, but also realizes frequency hopping peak avoidance in the time axis, improving the overall network spectrum utilization.

[0119] The lightweight imitation learning model also combines the spatial distribution characteristics of interference intensity in the frequency hopping optimization stage to dynamically determine the frequency hopping interval length and switching rate: for high-density interference scenarios, the model automatically shortens the frequency hopping residence time to quickly avoid interference frequency points; for low-density or intermittent interference scenarios, the model moderately extends the frequency hopping residence period to reduce switching energy consumption.

[0120] S53: determining a selective retransmission mechanism based on the transmission quality coefficient corresponding to the interference feature;

[0121] S54: outputting an anti-interference transmission strategy based on the joint optimization of the lightweight imitation learning model according to the transmission power parameter, the frequency hopping sequence and the retransmission mechanism.

[0122] Determining a selective retransmission mechanism based on the transmission quality coefficient corresponding to the interference feature includes:

[0123] S531: calculating the transmission quality coefficient of the current transmission channel based on the interference feature; the transmission quality coefficient is calculated based on the signal-to-interference ratio, the channel noise coefficient and the channel delay spread parameter;

[0124] S532: determining a retransmission threshold according to the transmission quality coefficient, and triggering the selective retransmission mechanism when the transmission quality is lower than the retransmission threshold;

[0125] S533: dividing the data to be transmitted into multiple data blocks of fixed length based on the selective retransmission mechanism, and configuring a check code for each data block;

[0126] S534: performing integrity check on the data blocks respectively, and if any data block fails the check, only the data block is retransmitted;

[0127] S535: In the retransmission process, the retransmission ratio and retransmission priority of the selective retransmission mechanism are dynamically adjusted according to the transmission quality coefficient; the retransmission priority includes high-level retransmission priority, medium-level retransmission priority or low-level retransmission priority, the high-level retransmission priority is set based on the modal parameter related data block, the medium-level retransmission priority is set based on the strain and acceleration data block, and the low-level retransmission priority is set based on the environmental quantity or auxiliary data block.

[0128] Specifically, the anti-interference transmission strategy includes dynamically adjusting the transmission power , optimizing the frequency hopping sequence , and the selective retransmission mechanism Three core parameters, the calculation formulas of which are as follows:

[0129] ;

[0130] Among them, is the reference transmission power, is the power adjustment coefficient , is the current channel interference index.

[0131] ;

[0132] Among them, is the initial frequency hopping sequence, is the set of available frequency bands, is the set of detected interference frequency bands, is a single frequency band element selected from the set of available frequency bands .

[0133] ;

[0134] ;

[0135] Among them, is the maximum retransmission ratio threshold , is the retransmission adjustment coefficient , is the current transmission quality coefficient, is the data priority weight coefficient, which is used to dynamically adjust the retransmission ratio according to the importance of the monitored data;

[0136] ;

[0137] Among them, SIR is the signal-to-interference ratio, is the channel noise coefficient.

[0138] Transmission quality coefficient The calculation also incorporates the channel delay spread (CDE) parameter, whose correction formula is as follows:

[0139] ;

[0140] in, For delay spread weighting coefficients ( CDE is the channel delay spread value; if Below the preset threshold Then the transmit power is triggered. The second adjustment is performed using the following formula:

[0141] ;

[0142] in, The transmission power at the previous moment. For power adjustment gain coefficient ( This correction mechanism can effectively address the impact of channel delay spread on signal integrity and further improve the reliability of anti-interference transmission.

[0143] It should be noted that in the step of dynamically adjusting the transmission power, the sensor node adjusts the power based on the current battery level. Priority of transmission tasks Joint adjustment The coefficients, and the specific rules are as follows:

[0144] when ( (This is the critical value of battery charge) and When it is high, The coefficient was adjusted to ( This refers to the power adjustment factor in low-battery mode. Task priority weight, 1≤ ≤2);

[0145] when and When it is low, The coefficient was adjusted to = ×( )( This is the power adjustment coefficient under normal power mode. ≥ );

[0146] This rule extends the battery life of sensor nodes while ensuring transmission quality by quantifying the relationship between battery status and task priority.

[0147] It should be noted that in generating frequency hopping sequences A node ID mapping and conflict avoidance mechanism is introduced. Specifically:

[0148] ;

[0149] in, For the first The final optimized frequency hopping sequence for the nth sensor node is the nth Each node uses a frequency band sequence for signal transmission, and this sequence is unique for each node.

[0150] It is the initial frequency hopping sequence, a basic frequency band sequence shared by all sensor nodes, and a benchmark template for frequency hopping optimization;

[0151] For a single available frequency band, it refers to a specific frequency band that has not been interfered with, selected from the set of available frequency bands;

[0152] This is the set of available frequency bands, representing the entire range of frequency bands that sensor nodes can use (such as preset industrial wireless monitoring frequency bands).

[0153] For the first The hash perturbation function for each node ID. This is the "union" operation for sets;

[0154] First, start with the set of available frequency bands. In the process, select those frequency bands that do not belong to the interference frequency band set. single frequency band (i.e., avoiding external interference frequency bands);

[0155] All filtered With the initial frequency hopping sequence Perform a "union" operation to merge them into a more comprehensive basic interference-free frequency hopping sequence—this sequence only addresses the situation of avoiding external interference, and has not yet resolved the frequency band conflict between nodes.

[0156] Furthermore, a node ID hash perturbation is added to generate a unique frequency hopping sequence for each node, avoiding frequency band conflicts between multiple nodes:

[0157] For the The unique identifier of each node Such as the hardware number of the sensor node, Substitute into the hash perturbation function This yields a personalized frequency band offset value;

[0158] The basic interference-free frequency hopping sequence generated in the first step is offset by the offset value of this frequency band, so that each frequency band position of the basic sequence undergoes a node-specific offset—ultimately yielding the first... of the nodes , different nodes have different offset results, and frequency band peak shifting is achieved.

[0159] The frequency hopping sequence is offset, so that the frequency hopping sequences of different nodes have a low overlap rate in the time and frequency domains, ensuring the adaptability of the scheme in a large-scale monitoring system.

[0160] It should be noted that the selective retransmission mechanism includes: dividing the data into multiple fixed-length data blocks , and calculating a check code (CRC) for each data block; if the receiving end detects a CRC error of a certain data block, only the data block is retransmitted, rather than full retransmission; at the same time, the data block size and the transmission bandwidth are introduced as constraint conditions in the calculation:

[0161] ;

[0162] wherein is the maximum allowed time for a single retransmission; this mechanism reduces the amount of redundant transmission data, reduces channel congestion and energy consumption caused by retransmission, while ensuring the integrity of critical data.

[0163] S6: transmitting the building structure monitoring data set to the center server based on the anti-interference transmission strategy, so that the center server performs fusion analysis on the building structure monitoring data sets of multiple nodes to obtain a building structure health evaluation result;

[0164] In this embodiment, after obtaining the anti-interference transmission strategy, the wireless sensor node performs wireless transmission management based on end-side adaptive control according to the current transmission channel state. The node dynamically adjusts the transmission power, channel frequency point selection, coding rate configuration and data packet sending priority according to the strategy instructions, and establishes a lightweight data integrity detection and real-time recovery mechanism at the node side to ensure that the target transmission signal still has high reliability output capability in a complex interference environment.

[0165] Specifically, the node configures multiple levels of cache queues in the transmission link, performs differentiated scheduling according to the importance level of the data (critical modal data > high dynamic change area data > steady state background data), and preferentially allocates critical data channels to candidate frequency points with the lowest interference degree, and automatically triggers frequency hopping, automatic retransmission or redundancy coding strategy under the strategy requirement. For non-critical data packets, low-power and low-frequency transmission is adopted to reduce the transmission resource occupation caused by additional interference.

[0166] The node-side wireless transmission module supports a real-time link monitoring mechanism. When sudden fluctuations occur in link status indicators, the edge-side rapid decision-making module immediately switches to a backup anti-interference strategy without relying on commands from the main control center, thereby reducing the impact of external network latency on the stability of real-time structural monitoring. Simultaneously, the transmission module utilizes CRC checksums, error frame discarding, and packet sequence number tracking to ensure data integrity.

[0167] Finally, the target transmission signal, after anti-interference optimization, is output in the form of timestamp alignment, node location identification, and data type label. It is then collected by the aggregation node and sent to the main control center, where it enters the subsequent signal fusion and building structure health evaluation stages.

[0168] In this embodiment, after receiving target transmission signals from multiple distributed wireless sensor nodes, the central monitoring server first performs precise data registration based on node ID, timestamp, and component location label to ensure consistent alignment of data from multiple nodes in both the time-series and component domains. Subsequently, it performs multi-source reliability judgment on the target transmission signals that have been optimized for anti-interference at the node side. Based on the node's own transmission stability, historical bit error rate, and node interference sensitivity level, it constructs weighting factors and integrates data from different nodes with differentiated reliability.

[0169] During the fusion phase, deep structural perception fusion processing is performed based on component distribution correlation, adjacent component correlation constraints, and spatial topology information to achieve cross-type joint analysis of strain, vibration acceleration, displacement, and modal parameter data. For local anomalies or short-period fluctuation signals appearing in the data, they are compared with the component's historical baseline behavior pattern library to identify potential health hazards such as local damage characteristics, signs of loose connections, local stiffness decay trends, and local abnormal modal drift, and the abnormal features are marked in the abnormal node list.

[0170] After completing the fusion calculation, a comprehensive operational status representation of the current building structure is constructed, resulting in a structural health assessment. This result includes information such as the global health level of the building structure, risk location of suspected abnormal areas, the trend of heterogeneity index changes, and suggestions for future risk evolution trends. This structural health assessment result not only serves as a real-time early warning output but also as an input reference for subsequent edge parameter updates of the imitation learning model, enabling dynamic disturbance adaptive optimization and continuous improvement of intelligent structural health perception.

[0171] S7: Update the parameters of the lightweight imitation learning model based on the building structure health assessment results returned by the central server.

[0172] In this embodiment, after completing the central end fusion analysis and health evaluation, the system will return the feedback data set of the interference environment information, node side transmission behavior, final evaluation accuracy result and data recovery confidence index in this period to each node. The node end updates the local lightweight imitation learning model according to the feedback data set, and the update is performed in a two-level linkage mode of "short period fast adaptation + long period global calibration".

[0173] In the short period update (such as every 1 hour), the node only updates the decision output layer and the feature weight mapping relationship, so that the node can respond to the interference event as soon as it occurs in the future, and make more efficient and controllable anti-interference strategy response without waiting for the central end instruction; This mechanism enables the node to gradually form "scene memory and individualized interference response habit", so that the node can converge and strengthen locally for specific buildings, specific floors, specific working conditions and specific electromagnetic interference, rather than starting from the initial reasoning every time.

[0174] In the long period update (such as every day or every week), the central server re-trains the model based on the wide area network running history and the structure health evaluation result, and distributes the new backbone weight file to all nodes, so as to avoid the risks of deviation accumulation caused by online incremental learning of the node end, early scene memory overfitting, local optimal lock, etc., and maintain the unified algorithm bottom logic of the whole network on a macro scale.

[0175] Through this mixed update mechanism of short period local fast adaptation and long period central unified calibration, the model not only can continuously strengthen its anti-interference strategy expression ability under specific environmental interference conditions, but also can gradually improve its overall generalization ability in complex multi-type interference combination scenes over time, so that the system anti-interference strategy presents the characteristics of sustainable iteration, cumulative enhancement and long-term evolution, forming a complete closed-loop self-enhancing anti-interference system.

[0176] Please refer to Figure 3 The embodiment of the application also provides an anti-interference transmission optimization system for building structure health monitoring signals. The system is arranged at a local control end. The local control end is connected with a plurality of distributed wireless sensors to form a node. The control ends in the plurality of nodes are connected with a central server through a wireless network. The system comprises:

[0177] A data acquisition module is configured to acquire dynamic response signals of a building structure in real time based on the distributed wireless sensors to form a building structure monitoring data set.

[0178] A first construction module is configured to deploy a lightweight imitation learning model.

[0179] The interference identification module is configured to identify an interference mode of a transmission channel corresponding to the building structure monitoring data set based on the lightweight imitation learning model, and generate an interference feature;

[0180] The anti-interference transmission strategy generation module is configured to generate an anti-interference transmission strategy based on the interference feature and the lightweight imitation learning model.

[0181] The transmission control module is configured to transmit the building structure monitoring data set to the central server based on the anti-interference transmission strategy, so that the central server performs fusion analysis on the building structure monitoring data sets of the multiple nodes to obtain a building structure health evaluation result.

[0182] The model update module is configured to update the lightweight imitation learning model based on the building structure health evaluation result returned by the central server.

[0183] It can be understood that the contents in the above method embodiments are applicable to the present system embodiment, the present system embodiment specifically implements the same functions as the above method embodiments, and achieves the same beneficial effects as the above method embodiments.

[0184] The present application embodiment further provides an electronic device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the above method when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a vehicle-mounted computer, etc.

[0185] It can be understood that the contents in the above method embodiments are applicable to the present device embodiment, the present device embodiment specifically implements the same functions as the above method embodiments, and achieves the same beneficial effects as the above method embodiments.

[0186] The present application embodiment further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the above method.

[0187] It can be understood that the contents in the above method embodiments are applicable to the present storage medium embodiment, the present storage medium embodiment specifically implements the same functions as the above method embodiments, and achieves the same beneficial effects as the above method embodiments.

[0188] The present application embodiment further provides a computer program product, which includes a computer program, and the computer program is executed by a processor to implement the above method.

[0189] It can be understood that the contents in the above method embodiments are all applicable to the program product embodiments, the program product embodiments specifically implement the same functions as the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0190] The memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include a high-speed random access memory and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory that is remotely arranged relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0191] The embodiment of the present application provides an anti-interference transmission optimization method and system for building structure health monitoring signals. The scheme can adaptively extract interference characteristics in different environments through interference mode identification of the transmission channel, and generate a targeted anti-interference transmission strategy based on a lightweight imitation learning model, thereby effectively suppressing interference problems such as channel noise, spectral overlap, and multipath fading, and significantly improving the transmission quality and stability of building structure monitoring signals. A lightweight imitation learning model is deployed locally on a wireless sensor node, enabling the node to have real-time learning and parameter updating capabilities. The model can dynamically optimize its parameters according to the building structure health evaluation results, thereby continuously improving the adaptability and generalization ability of the anti-interference strategy, implementing a distributed self-optimization transmission mechanism, and reducing the dependence on a central server. Through fusion analysis of the target transmission signal, the monitoring results of multiple nodes can be integrated to reduce data distortion and packet loss problems caused by interference, ensuring the integrity and accuracy of the health evaluation data and providing high-confidence support for rapid judgment of the safety state of the building structure. The lightweight structure design of the imitation learning model is suitable for resource-constrained wireless nodes for local inference and strategy generation, without the need for complex cloud computing and high-bandwidth communication, significantly reducing node energy consumption and system maintenance costs, and facilitating stable operation in long-term building structure monitoring scenarios.

[0192] The embodiments described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0193] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and can include more or fewer steps than the figures, or combine certain steps, or different steps.

[0194] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separated, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments.

[0195] Those skilled in the art can understand that all or some steps in the above disclosed method, the functional modules / units in the system and the device can be implemented as software, firmware, hardware and their appropriate combinations.

[0196] The terms "first", "second", "third", "fourth" and the like used in the description of the present application and the above-described drawings, if any, are used to distinguish similar objects, and do not necessarily have to describe a particular order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0197] The preferred embodiments of the embodiments of the present application are described above with reference to the accompanying drawings, and are not limited to the scope of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the embodiments of the present application.

Claims

1. A method for optimizing the anti-interference transmission of building structural health monitoring signals, characterized in that, The method is applied to a local control terminal, which is connected to multiple distributed wireless sensors to form a node. The control terminals within these nodes are all connected to a central server via a wireless network. The method includes the following steps: The system acquires dynamic response signals of a building structure in real time from distributed wireless sensors; these dynamic response signals include strain signals, vibration acceleration signals, displacement signals, and modal parameter signals generated during the building structure's operation. The dynamic response signal is converted into a digital signal to form a building structure monitoring dataset; Deploy a lightweight imitation learning model; Based on the lightweight imitation learning model, interference pattern recognition is performed on the transmission channel corresponding to the building structure monitoring dataset to generate interference features; Based on the interference characteristics, the lightweight imitation learning model is used to generate an anti-interference transmission strategy; Based on the anti-interference transmission strategy, the building structure monitoring dataset is transmitted to the central server, so that the central server can perform fusion analysis on the building structure monitoring datasets of multiple nodes and obtain the building structure health assessment results. The parameters of the lightweight imitation learning model are updated based on the building structure health assessment results returned by the central server. The step of generating an anti-interference transmission strategy using the lightweight imitation learning model based on the interference characteristics includes: Based on the interference characteristics, the transmission power is dynamically adjusted in combination with the current battery level and transmission task priority to generate transmission power parameters; Based on the interference frequency distribution corresponding to the interference characteristics, the initial frequency hopping sequence is optimized to generate a frequency hopping sequence. A selective retransmission mechanism is determined based on the transmission quality coefficient corresponding to the interference characteristics. Based on the lightweight imitation learning model, an anti-interference transmission strategy is output through joint optimization of the transmit power parameters, frequency hopping sequence, and retransmission mechanism.

2. The method according to claim 1, characterized in that, The lightweight imitation learning model is built on a long short-term memory network and is trained on the edge using the historical interference index as a supervision signal.

3. The method according to claim 1, characterized in that, The step of performing interference pattern recognition on the transmission channel corresponding to the building structure monitoring dataset based on the lightweight imitation learning model to generate interference features includes: Using a preset detection period as a window, the received signal strength indication and noise power of each frequency band are measured one by one in the building structure monitoring dataset to obtain the target time domain signal; The target time-domain signal is converted into a frequency-domain energy spectrum using a fast Fourier transform. Candidate interference frequency points are identified in the frequency domain energy spectrum; The candidate interference frequency points are subjected to interference persistence determination, and the interference index is calculated based on the interference persistence determination results; The interference index is input into the lightweight imitation learning model, and the interference features characterizing the current channel state are output; the interference features include interference frequency distribution, interference intensity level and interference duration interval.

4. The method according to claim 3, characterized in that, The interference index The weighted summation formula is used to calculate the following: ； in, frequency band The duration of the interference, This represents the total duration of the current testing cycle; M represents the signal strength; M represents the total number of available frequency bands.

5. The method according to claim 1, characterized in that, The selective retransmission mechanism based on the transmission quality coefficient corresponding to the interference characteristics includes: The transmission quality coefficient of the current transmission channel is calculated based on the interference characteristics. A retransmission threshold is determined based on the transmission quality coefficient, and a selective retransmission mechanism is triggered when the transmission quality is lower than the retransmission threshold. Based on the selective retransmission mechanism, the data to be transmitted is divided into multiple fixed-length data blocks, and a checksum is configured for each data block. The integrity of each data block is checked. If the check of any data block fails, only the data block is retransmitted. During the retransmission process, the retransmission ratio and retransmission priority of the selective retransmission mechanism are dynamically adjusted based on the transmission quality coefficient.

6. The method according to claim 5, characterized in that, The transmission quality coefficient is calculated based on the signal-to-interference ratio, channel noise figure, and channel delay spread parameter.

7. The method according to claim 5, characterized in that, The retransmission priority includes high-level retransmission priority, medium-level retransmission priority, or low-level retransmission priority. The high-level retransmission priority is set based on modal parameter related data blocks, the medium-level retransmission priority is set based on strain and acceleration data blocks, and the low-level retransmission priority is set based on environmental quantities or auxiliary data blocks.

8. An anti-interference transmission optimization system for building structural health monitoring signals, characterized in that, The system is set at a local control terminal, which is connected to multiple distributed wireless sensors to form a node. The control terminals within each node are all connected to a central server via a wireless network. The system includes: The data acquisition module is used to collect dynamic response signals of the building structure in real time based on distributed wireless sensors, forming a building structure monitoring dataset; the dynamic response signals include strain signals, vibration acceleration signals, displacement signals and modal parameter signals generated during the operation of the building structure; The first building block is used to deploy lightweight imitation learning models; The interference identification module is used to perform interference pattern identification on the transmission channel corresponding to the building structure monitoring dataset based on the lightweight imitation learning model, and generate interference features. An anti-interference transmission strategy generation module is used to generate an anti-interference transmission strategy based on the interference characteristics and using the lightweight imitation learning model. The transmission control module is used to transmit the building structure monitoring dataset to the central server based on the anti-interference transmission strategy, so that the central server can perform fusion analysis on the building structure monitoring datasets of multiple nodes and obtain the building structure health assessment results. The model update module is used to update the parameters of the lightweight imitation learning model based on the building structure health assessment results returned by the central server. The step of generating an anti-interference transmission strategy using the lightweight imitation learning model based on the interference characteristics includes: Based on the interference characteristics, the transmission power is dynamically adjusted in combination with the current battery level and transmission task priority to generate transmission power parameters; Based on the interference frequency distribution corresponding to the interference characteristics, the initial frequency hopping sequence is optimized to generate a frequency hopping sequence. A selective retransmission mechanism is determined based on the transmission quality coefficient corresponding to the interference characteristics. Based on the lightweight imitation learning model, an anti-interference transmission strategy is output through joint optimization of the transmit power parameters, frequency hopping sequence, and retransmission mechanism.

Images