Anti-interference transmission method of wireless low-power consumption transducer

Abstract

The application relates to the technical field of wireless transmission, and discloses an anti-interference transmission method for a wireless low-power transmitter. The method comprises the following steps: collecting real-time electromagnetic interference signal frequency spectrum of an environment where the transmitter is located, and extracting multi-dimensional interference features; constructing an interference feature library based on the features, and dividing interference feature clusters according to feature distribution density and intensity change trend; calculating dynamic fluctuation indexes and frequency domain concentration indexes of each cluster, and obtaining peak frequency drift indexes and spectrum bandwidth variation indexes of each cluster; generating an environmental interference entropy value based on the above indexes; constructing a transmission signal sequence covariance matrix and performing feature decomposition to obtain eigenvalue distribution, concentration, principal component contribution degree and dispersion coefficient related to a discrete state; combining the covariance matrix row vector correlation, the environmental interference entropy value and the principal component contribution degree and dispersion coefficient, generating a signal transmission purity index, and performing anti-interference transmission control according to the index to improve the transmission stability of the transmitter in a complex electromagnetic environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wireless transmission, in particular to an anti-interference transmission method of a wireless low-power transmitter. BACKGROUND

[0002] In the fields of industrial automation, environmental monitoring, intelligent sensing, etc., as a key equipment for data acquisition and transmission, the transmission stability of a wireless low-power transmitter directly affects the operation efficiency of the entire system. With the wide application of various electronic devices, the electromagnetic environment in which the transmitter is located is becoming increasingly complex. Electromagnetic interference signals from different devices are superimposed on each other, resulting in frequent problems such as signal distortion, transmission delay, and even data loss during data transmission of the transmitter.

[0003] For the anti-interference solution of a wireless low-power transmitter, most of them are limited to a single interference suppression level. For example, some technologies optimize the hardware circuit design of the transmitter to enhance its shielding ability against interference signals of specific frequencies, but this method can only deal with interference signals of known frequency ranges. When new frequency interference signals appear in the environment, the anti-interference effect will be greatly reduced. Some other technologies use fixed signal filtering algorithms to process the transmission signals for noise reduction. However, such algorithms cannot adjust the filtering parameters in real time according to the dynamic changes of interference signals. In the scenario where the intensity and frequency of interference signals are constantly fluctuating, it is difficult to effectively filter the interference components, and at the same time, it may cause excessive attenuation to useful signals, affecting the accuracy of data transmission.

[0004] The existing technologies generally lack systematic analysis and evaluation mechanisms for environmental interference signals, and cannot fully grasp the characteristic distribution law of interference signals, resulting in a lack of scientific basis when formulating anti-interference strategies. For example, when multiple different types of interference signals exist at the same time, the existing technologies cannot accurately distinguish the influence degree of each interference signal, nor can they judge the trend of the interference signals, so as to take targeted anti-interference measures. Therefore, the transmission stability of the transmitter in a complex electromagnetic environment cannot be effectively guaranteed, which seriously restricts the application of wireless low-power transmitters in fields with high requirements for transmission reliability, such as industrial control, medical monitoring, etc. SUMMARY

[0005] The present application aims to provide an anti-interference transmission method of a wireless low-power transmitter to solve the problems raised in the background.

[0006] To achieve the above-mentioned purpose, the present application provides an anti-interference transmission method of a wireless low-power transmitter, which comprises:

[0007] Collecting the real-time electromagnetic interference signal spectrum of the environment in which the transmitter is located, and extracting multi-dimensional interference characteristics;

[0008] constructing an interference feature library based on multi-dimensional interference features, and dividing interference feature clusters according to the distribution density and intensity change trend of the features in the interference feature library;

[0009] calculating the dynamic fluctuation index and the frequency domain concentration index of each interference feature cluster;

[0010] obtaining the peak frequency drift index and the spectral bandwidth variation index of each interference feature cluster;

[0011] generating an environmental interference entropy value based on the peak frequency drift index, the spectral bandwidth variation index, the average value of the dynamic fluctuation index of all interference feature clusters, and the average value of the feature similarity between interference feature clusters;

[0012] constructing a covariance matrix of the current transmission signal sequence, performing eigenvalue decomposition on the covariance matrix to obtain each eigenvalue, and obtaining the principal component contribution degree and the dispersion coefficient of the eigenvalue sequence of the transmission signal sequence based on the distribution state, the concentration state, and the dispersion state of the eigenvalue;

[0013] generating a signal transmission purity index based on the correlation between each row vector in the covariance matrix, the environmental interference entropy value, the principal component contribution degree, and the dispersion coefficient, and performing anti-interference transmission control based on the signal transmission purity index.

[0014] Preferably, the interference feature cluster division method comprises:

[0015] using a spectral peak detection algorithm to identify all main peaks and secondary peaks in the electromagnetic interference signal spectrum, segmenting the spectrum sequence from each secondary peak to form a feature subsequence, and calculating the spectral energy range of each feature subsequence;

[0016] inputting the spectral energy range of all feature subsequences into a clustering algorithm to output each feature clustering cluster;

[0017] calculating the internal spectral energy mean of each feature clustering cluster, taking the feature clustering cluster with the largest internal spectral energy mean as the dominant interference cluster, and taking the spectral subsequence corresponding to the internal features of the dominant interference cluster as the interference feature cluster.

[0018] Preferably, the method for obtaining the dynamic fluctuation index comprises:

[0019] For each interference feature cluster, locate the maximum spectral intensity point in the interference feature cluster, and form a left spectral segment by combining the maximum spectral intensity point and the left spectral point, and form a right spectral segment by combining the right spectral point of the maximum spectral intensity point;

[0020] obtaining the first-order gradient sequence of the left spectral segment of each interference feature cluster, and generating a gradient transformation sequence by processing the first-order gradient sequence using a nonlinear transformation function;

[0021] The absolute value of the difference between the gradient transform sequence element and the absolute value of the first order gradient sequence length of the left side spectrum segment is calculated as the fluctuation intensity index of the left side spectrum segment;

[0022] The fluctuation intensity index of the right side spectrum segment of each interference characteristic cluster is obtained by using the same method;

[0023] The product of the absolute value of the difference between the fluctuation intensity indexes of the left side spectrum segment and the right side spectrum segment and the mean value is calculated as the dynamic fluctuation index of the interference characteristic cluster;

[0024] The variance of the spectral intensity of the left side spectrum segment and the variance of the spectral intensity of the right side spectrum segment of each interference characteristic cluster are calculated, and the mean value of the variances of the two sides is taken as the frequency domain concentration index.

[0025] Preferably, the method for obtaining the spectral bandwidth variation index comprises:

[0026] The main peak intensity values in all interference characteristic clusters are extracted, and a main peak intensity sequence is constructed by sorting the frequency point positions in the spectral sequence;

[0027] The frequency point positions of each main peak in the interference characteristic cluster are sorted by size to construct a main peak frequency point sequence;

[0028] The first order difference sequence of the main peak intensity sequence and the main peak frequency point sequence is obtained, the variance of the first order difference sequence of the main peak intensity sequence is taken as the peak frequency drift index, and the variance of the first order difference sequence of the main peak frequency point sequence is taken as the spectral bandwidth variation index.

[0029] Preferably, the method for generating the environmental interference entropy value comprises:

[0030] The mean value of the dynamic fluctuation indexes of all interference characteristic clusters is calculated as the average fluctuation index;

[0031] The mean value of the spectral similarity coefficients of each interference characteristic cluster and all other interference characteristic clusters is taken as the characteristic cluster similarity;

[0032] The peak frequency drift index, the spectral bandwidth variation index, the average fluctuation index, and the mean value of the characteristic cluster similarity are input into an environmental interference entropy value calculation model to generate the environmental interference entropy value.

[0033] Preferably, the method for obtaining the principal component contribution degree and the dispersion coefficient comprises:

[0034] All characteristic values are arranged in descending order to construct a characteristic value sequence;

[0035] The characteristic value sequence is input into an adaptive threshold segmentation algorithm, and a segmentation threshold value is output. The characteristic values greater than or equal to the segmentation threshold value are listed as principal characteristic values, and the characteristic values less than the segmentation threshold value are listed as secondary characteristic values;

[0036] The proportion of the number of main eigenvalues to the total number of eigenvalues is calculated as the principal component contribution degree;

[0037] The product of the mean value and the variance of the main eigenvalue is calculated as the main eigenvalue fluctuation index, and the secondary eigenvalue fluctuation index is calculated in the same way. The absolute value of the difference between the main eigenvalue fluctuation index and the secondary eigenvalue fluctuation index is taken as the dispersion coefficient.

[0038] Preferably, the signal transmission purity index generation method comprises:

[0039] The absolute value of the correlation coefficient of each row vector in the covariance matrix and other row vectors is calculated as the row correlation degree.

[0040] The mean value of all row correlation degrees of the covariance matrix is taken as the matrix autocorrelation degree.

[0041] The principal component contribution degree, the dispersion coefficient, the matrix autocorrelation degree, and the environmental interference entropy value are input into the purity calculation model to generate the signal transmission purity index.

[0042] Preferably, the anti-interference transmission control based on the signal transmission purity index comprises:

[0043] An adaptive frequency hopping interval is calculated according to the signal transmission purity index.

[0044] The carrier frequency of the transmitter is dynamically adjusted based on the adaptive frequency hopping interval.

[0045] The data to be transmitted is segmented into data block groups at the adjusted carrier frequency, and a unique cluster identification code and a hierarchical check code are attached to each data block to form a transmission cluster unit.

[0046] The transmission cluster units are sent in hierarchical order, and multi-level checking is performed at the receiving end.

[0047] Preferably, the multi-level checking comprises:

[0048] The receiving end reorganizes the data block groups according to the cluster identification code and extracts the hierarchical check code.

[0049] First-level parity checking is performed, and if the checking fails, second-level cyclic redundancy checking is activated.

[0050] If the cyclic redundancy checking fails, a check feature library is generated based on the feature distribution of the data block group.

[0051] The best fault-tolerant decoding strategy is matched according to the check feature library for data recovery.

[0052] Preferably, the matching of the fault-tolerant decoding strategy comprises:

[0053] The current channel quality index and the remaining energy index of the transmitter are monitored.

[0054] The maximum allowed retransmission number is calculated according to the channel quality index, and the minimum power consumption transmission mode is calculated according to the residual energy index;

[0055] If the number of data recovery failures does not reach the maximum allowed retransmission number, initiate a local retransmission request in the minimum power consumption transmission mode;

[0056] Otherwise, select to discard or perform degraded decoding according to the priority mark of the data block group.

[0057] Compared with the prior art, the beneficial effects of the present application are:

[0058] The anti-interference transmission method of the wireless low-power transmitter can comprehensively and meticulously master the specific properties of the interference signals in the environment by collecting the real-time electromagnetic interference signal spectrum of the environment where the transmitter is located and extracting multi-dimensional interference features, breaking the limitation of traditional technologies that can only process single or fixed type interference. Based on the multi-dimensional interference features, an interference feature library is constructed, and the interference feature clusters are divided according to the distribution density and intensity change trend of the features, so that complex and diverse interference signals can be classified and arranged, and the distribution law and change trend of different interference signals are clearly presented, providing comprehensive information support for subsequent formulation of accurate anti-interference strategies.

[0059] The calculation of the dynamic fluctuation index and the frequency domain concentration index of each interference feature cluster, as well as the acquisition of the peak frequency drift index and the spectral bandwidth variation index, can quantize the dynamic change characteristics and frequency domain distribution features of the interference signals from multiple dimensions. Through these quantitative indicators, the stability, frequency change range and bandwidth fluctuation of the interference signals can be accurately judged. Compared with the deficiency of traditional technologies that cannot quantize the features of interference signals, this method can more deeply analyze the essential properties of interference signals, thereby providing accurate judgment basis for subsequent adjustment of anti-interference measures.

[0060] The generation of the environmental interference entropy value based on the above multi-dimensional quantitative indicators can comprehensively reflect the complexity and uncertainty of the interference signals in the environment. The introduction of the environmental interference entropy value can convert the overall interference environment which is difficult to measure into a quantifiable indicator, making the evaluation of the interference environment more objective and scientific, and avoiding the problem that the anti-interference measures are not targeted due to the inability to accurately evaluate the interference degree in the traditional technology when facing complex interference environment.

[0061] The construction of the covariance matrix of the current transmission signal sequence and the feature decomposition can obtain the distribution state, concentration state and dispersion state of the eigenvalues, and then obtain the principal component contribution degree and dispersion coefficient, which can comprehensively analyze the characteristics of the transmission signal itself. Through these signal characteristic indicators, the distribution of the effective components in the transmission signal and the dispersion degree of the signal can be clearly understood, which provides a direct basis for judging the degree of influence of the signal by the interference, and also provides a reference direction for subsequent adjustment of signal transmission parameters.

[0062] The signal transmission purity index is generated based on the correlation between each row vector in the covariance matrix, the environmental interference entropy value, the principal component contribution degree and the dispersion coefficient, and interference-resistant transmission control is performed according to the index, so that the dynamic and accurate interference-resistant measures are realized. Compared with the fixed interference-resistant strategy of the traditional technology, the method can flexibly adjust the interference-resistant parameters and strategies according to the real-time signal transmission purity index, timely responds and takes corresponding processing measures when the interference signal changes, effectively filters the interference components, reduces the influence on the useful signals, and guarantees the stability and accuracy of data transmission.

[0063] The whole method does not need to rely on complex hardware upgrade, but through the analysis of the interference signal and the evaluation of the signal characteristics, combined with the dynamically adjusted interference-resistant control strategy, efficient interference-resistant transmission can be realized, the power consumption of the transmitter is effectively controlled while the interference-resistant effect is guaranteed, which meets the application requirements of the wireless low-power transmitter, expands its application range in various complex electromagnetic environments, and is especially suitable for scenes with high requirements for transmission stability and device power consumption. BRIEF DESCRIPTION OF DRAWINGS

[0064] Figure 1 The working principle diagram of the wireless low-power transmitter interference-resistant transmission method is provided.

[0065] Figure 2 The method flowchart for interference feature cluster division is provided.

[0066] Figure 3 The method flowchart for obtaining the spectral bandwidth variation index and the peak frequency drift index is provided.

[0067] Figure 4 The method flowchart for obtaining the principal component contribution degree and the dispersion coefficient is provided. DETAILED DESCRIPTION

[0068] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0069] Please refer to Figure 1 The present application provides a wireless low-power transmitter interference-resistant transmission method, which comprises:

[0070] The real-time electromagnetic interference signal spectrum of the environment where the acquisition transmitter is located is collected, and multi-dimensional interference features are extracted; an interference feature library is constructed based on the multi-dimensional interference features, and interference feature clusters are divided according to the distribution density and intensity change trend of the features in the interference feature library; the dynamic fluctuation index and the frequency domain concentration index of each interference feature cluster are calculated; the peak frequency drift index and the spectral bandwidth variation index of each interference feature cluster are obtained; based on the peak frequency drift index, the spectral bandwidth variation index, the average value of the dynamic fluctuation index of all interference feature clusters, and the average value of the feature similarity between interference feature clusters, an environmental interference entropy value is generated; a covariance matrix of the current transmission signal sequence is constructed, the covariance matrix is decomposed to obtain each eigenvalue, and based on the distribution state, concentration state and dispersion state of the eigenvalue, the principal component contribution degree and the dispersion coefficient of the eigenvalue sequence of the transmission signal sequence are obtained; a signal transmission purity index is generated based on the correlation between the row vectors in the covariance matrix, the environmental interference entropy value, the principal component contribution degree and the dispersion coefficient; and anti-interference transmission control is performed based on the signal transmission purity index.

[0071] Embodiment 1: see Figure 2 In the operating environment of a wireless low-power transmitter, the process of collecting the electromagnetic interference signal spectrum of its surroundings in real time is completed by a built-in spectrum analysis module. This module captures the radio frequency energy distribution within a specific frequency band at a fixed sampling period and outputs a digital sequence composed of frequency points and intensity values. This sequence is the original electromagnetic interference signal spectrum. Multi-dimensional interference features are extracted from it, including but not limited to the peak intensity, average energy, zero-crossing points, spectral centroid, and spectral roll-off points. These features collectively provide a multi-angle quantitative description of the current electromagnetic environment.

[0072] Based on these multi-dimensional interference features, the system constructs a dynamically updated interference feature library. This library not only stores the feature vectors at the current time, but also retains historical features within a certain time window to analyze the evolution trend of the features. The division of interference feature clusters relies on the analysis of the data distribution in the feature library. Specifically, a spectrum peak detection algorithm is used to process the original electromagnetic interference signal spectrum. This algorithm scans the entire spectrum sequence, identifies all frequency points that meet the local maximum condition and have an intensity that exceeds the global average intensity by a certain percentage, and marks these points as primary or secondary peaks. The distinction between primary and secondary peaks is based on their relative intensity and their prominence in the frequency domain.

[0073] From the identified peak positions, the continuous spectrum sequence is segmented into multiple feature subsequences. The boundary of each feature subsequence is determined by the adjacent peak frequency or the start and end points of the spectrum. The spectral energy range of each feature subsequence is calculated, which is the difference between the maximum and minimum spectral intensity values in the subsequence. This energy range reflects the dynamic variation amplitude of the interference signal in the local frequency band. The calculated spectral energy range values of all feature subsequences are input into a clustering algorithm for processing. The clustering algorithm analyzes the distribution of the energy range values and classifies the energy ranges with similar values into the same set, thereby forming several feature clustering clusters. Each clustering cluster represents a local spectrum region with similar energy fluctuation characteristics.

[0074] The average internal spectral energy of each feature clustering cluster is calculated, which is the arithmetic mean of the spectral energy ranges of all feature subsequences in the cluster. The internal average of all feature clustering clusters is compared, and the clustering cluster with the largest internal average is determined as the dominant interference cluster. This means that the spectrum region represented by this cluster exhibits the strongest energy variation characteristics, usually corresponding to the main source of interference in the environment. Finally, all feature subsequences belonging to this dominant interference cluster are extracted from the original spectrum, and these sequences are formally defined as the interference feature cluster required for subsequent analysis. For each interference feature cluster, further analysis of its dynamic characteristics is required. First, locate the frequency point with the maximum spectral intensity in the interference feature cluster, i.e., the maximum spectral intensity point. Divide the spectrum sequence of the interference feature cluster into two parts at this point: the left spectrum band consists of the maximum intensity point and all frequency points to its left; the right spectrum band consists of all frequency points to the right of the maximum intensity point. This division helps to separately examine the different behaviors of the interference signal on both sides of the peak frequency.

[0075] The left side spectrum segment is analyzed to calculate the rate of change of intensity values of adjacent frequency points in the segment, i.e. the first order gradient sequence. This sequence reflects the speed and direction of the change of intensity when transitioning from the left side of the spectrum to the peak. A non-linear transformation function is applied to the first order gradient sequence, which can amplify significant gradient changes while suppressing minor fluctuations, thus obtaining a transformed gradient sequence. The sum of absolute values of all elements of the transformed sequence is calculated, and then the absolute value of the difference between this value and the total length of the first order gradient sequence of the left side spectrum segment is calculated. This result is defined as the fluctuation intensity index of the left side spectrum segment, which quantifies the severity of signal change in the left side frequency band. The same analysis process is applied to the right side spectrum segment. The first order gradient sequence of the right side is calculated, the same non-linear transformation is applied, and the sum of absolute values of elements of the transformed sequence and the absolute value of the difference between this value and the length of the sequence are calculated to obtain the fluctuation intensity index of the right side spectrum segment. The absolute value of the difference between the left side fluctuation intensity index and the right side fluctuation intensity index is calculated, and the arithmetic mean of the two indices is calculated. The product of the above absolute difference value and the mean value is the dynamic fluctuation index of the interference feature cluster. This index comprehensively reflects the asymmetry of the change of the interference signal on both sides of the peak and the overall fluctuation level.

[0076] The variances of intensity values of all frequency points in the left side spectrum segment and the variances of intensity values of all frequency points in the right side spectrum segment are calculated respectively. Variance measures the degree of dispersion of data around its mean value. The arithmetic mean of the two variances is calculated, and this value is taken as the frequency domain concentration index of the interference feature cluster, which represents the degree of concentration of the distribution of interference energy around the peak in the frequency domain. For each interference feature cluster divided, a set of quantitative indicators describing its dynamic characteristics and concentration characteristics is obtained: dynamic fluctuation index and frequency domain concentration index. These indicators provide a data basis for subsequent comprehensive evaluation of the complexity and intensity of the electromagnetic environment. The entire implementation process gradually extracts key features from the original spectrum data through multi-stage processing of the spectrum signal, and finally converts them into quantifiable indicator values for guiding anti-interference transmission decisions.

[0077] Example 2: see Figure 3After the completion of the division of each interference feature cluster and the extraction of basic parameters, the implementation process enters the quantitative analysis stage of the macroscopic time-varying characteristics of the interference spectrum. The core of this stage is to extract key features from the identified interference feature clusters and calculate a set of comprehensive indices that can represent the complexity and instability of the entire electromagnetic environment. The object of processing is all the interference feature clusters divided previously. Each interference feature cluster corresponds to a local energy concentration area in the spectrum, and it contains a dominant peak inside. Extract the intensity value of the main peak from each interference feature cluster, which is a scalar value representing the peak energy level of the interference signal in the cluster. Then, according to the order of the frequency point positions of these main peaks in the original global spectrum, sort all the main peak intensity values to construct an ordered main peak intensity sequence. The order of elements in this sequence reflects the distribution of different interference sources in the frequency domain, and the change of its numerical value represents the difference in intensity of each interference source.

[0078] Similarly, for all these interference feature clusters, extract the specific frequency point position value corresponding to the main peak in each cluster. Arrange these frequency point position values in ascending or descending order according to their numerical value to construct a main peak frequency point sequence. This sequence describes the specific distribution of the main interference sources in the environment on the frequency axis. In order to quantify the instability of these main peak characteristics over time or environmental changes, it is necessary to analyze the first-order difference characteristics of the above two sequences. Calculate the first-order difference sequence of the main peak intensity sequence, that is, calculate the difference between each element and its next element in the sequence. This difference sequence reflects the change amplitude and direction of the intensity level between adjacent main peaks. Calculate the variance of this first-order difference sequence, and define the variance value as the peak frequency drift index. The size of this index reflects the fluctuation of the intensity level of the main interference sources in the environment, and the larger the value, the more intense and unstable the change in interference intensity.

[0079] Calculate the first-order difference sequence of the main peak frequency point sequence, that is, calculate the difference in frequency point position between adjacent main peaks. This difference sequence reflects the change in distribution density of interference sources in the frequency domain and the possible frequency shift phenomenon. Calculate the variance of this first-order difference sequence, and define the variance value as the spectral bandwidth variation index. This index quantifies the degree of change in the distribution of interference sources on the frequency axis, and the increase in the value indicates that the appearance position of the interference source is more dispersed or moves more frequently, meaning that the spectrum occupation is more complex. After obtaining the dynamic fluctuation index that characterizes the dynamic characteristics of a single interference feature cluster, it is necessary to evaluate the fluctuation level of the overall interference from a global perspective. Calculate the arithmetic mean of the dynamic fluctuation indices of all interference feature clusters, and this average value is the average fluctuation index, which represents the average intensity of the dynamic change of interference signals in the current electromagnetic environment.

[0080] Further analysis of the interrelation between different interference sources. For each interference feature cluster, the similarity coefficient between its spectrum waveform and the spectrum waveform of each other interference feature cluster is calculated. This coefficient is derived by comparing the spectrum shape, distribution, etc. of the two clusters. For each interference feature cluster, the average of the similarity coefficients of this cluster with all other clusters is calculated, obtaining the feature cluster similarity of this cluster. Subsequently, the arithmetic average of the feature cluster similarities of all interference feature clusters is calculated, obtaining the average feature cluster similarity. This value reflects the similarity degree between different interference sources in the environment, and a high average value indicates that the types of interference sources may be relatively single, and a low average value indicates that the interference sources are diverse and complex.

[0081] The four key indicators calculated above: peak frequency drift index, spectrum bandwidth variation index, average fluctuation index, and average feature cluster similarity, are input into an environmental interference entropy value calculation model. According to the different weights and interrelationships of these indicators, the model generates a scalar output value, i.e. the environmental interference entropy value, through a comprehensive mathematical operation. This entropy value is a highly summarized indicator that integrates interference intensity stability, frequency domain distribution characteristics, overall fluctuation level, and interference source diversity, thus comprehensively quantifying the interference complexity and unpredictability of the current wireless environment. A higher environmental interference entropy value means that the transmission environment is poor, and more aggressive anti-interference strategies need to be taken; while a lower entropy value indicates that the environment is relatively stable and simple. This entropy value provides a crucial environmental state input for subsequent adaptive transmission decisions. The entire implementation process converts raw spectrum data into comprehensive environmental evaluation parameters with clear physical meaning and decision value through layer-by-layer analysis and calculation.

[0082] Example 3: see Figure 4 In the signal processing flow of a wireless low-power transmitter, the internal structure analysis of the current signal sequence to be transmitted is the key link to generate the signal transmission purity index. This implementation starts with constructing a mathematical expression representing the internal relationship of the signal sequence. Specifically, the system collects the discrete signal samples to be sent by the transmitter, which are arranged in time sequence to form a multi-dimensional signal vector. Based on this signal vector, the covariance matrix is calculated. This matrix is a symmetric matrix, whose diagonal elements represent the variances of each dimension signal, and the non-diagonal elements represent the covariance relationship between different dimension signals, reflecting the linear correlation degree between the components of the signal sequence.

[0083] After the covariance matrix is constructed, it is subjected to an eigen decomposition operation. Eigen decomposition is a matrix decomposition technique that aims to decompose the original matrix into a set of eigenvectors and corresponding eigenvalues. In this process, a numerically stable algorithm is employed to solve all eigenvalues of the matrix, which are essentially scaling factors of the linear transformation represented by the covariance matrix, and their magnitudes are directly related to the energy or importance of the principal components in the signal sequence. All solved eigenvalues are sorted in descending order of their numerical values to form an ordered eigenvalue sequence. The distribution pattern of this sequence contains key information about the intrinsic structure of the signal.

[0084] To quantify the distribution characteristics of the eigenvalue sequence, it is input into an adaptive threshold segmentation algorithm. This algorithm automatically analyzes the numerical distribution of the sequence and finds an optimal segmentation point. The principle of determining this segmentation point is to maximize the inter-class difference of the two sub-sequences after segmentation. The elements in the eigenvalue sequence that are greater than or equal to the segmentation threshold are classified as principal eigenvalues, while the elements that are less than the segmentation threshold are classified as secondary eigenvalues. Principal eigenvalues represent the dominant, high-energy main components in the signal sequence, while secondary eigenvalues correspond to low-energy, secondary components or noise.

[0085] The proportion of the number of principal eigenvalues to the total number of eigenvalues is calculated, which is defined as the principal component contribution degree of the signal sequence. The higher this proportion value, the more concentrated the signal energy is on a few principal components, and the clearer the signal structure. Conversely, a low proportion value means that the signal energy is dispersed, possibly with more interference or complex structure. Further analysis of the statistical characteristics of the principal eigenvalue and secondary eigenvalue sets is performed to calculate the arithmetic mean of the principal eigenvalues and the variance of the values in the set. The product of the mean of the principal eigenvalues and the variance of the set is defined as the principal eigenvalue fluctuation index. This index combines the average energy level of the principal components and their internal fluctuation degree. The same calculation method is used to process the secondary eigenvalue set to calculate the product of the mean and variance of the secondary eigenvalues, obtaining the secondary eigenvalue fluctuation index. Finally, the absolute value of the difference between the principal eigenvalue fluctuation index and the secondary eigenvalue fluctuation index is calculated, which is defined as the dispersion coefficient. The size of the dispersion coefficient reflects the difference in energy distribution and stability between the main components and the secondary components in the signal sequence. The larger the difference, the higher the dispersion coefficient.

[0086] In the analysis of internal correlation of signal sequence, the correlation strength between the row vectors of the covariance matrix needs to be investigated. The Pearson correlation coefficient of each row vector in the matrix with all other row vectors is calculated. The Pearson correlation coefficient measures the degree of linear correlation between two vectors. For each row, the absolute values of the correlation coefficients with all other rows are calculated, and the arithmetic mean of these absolute values is taken to obtain the row correlation degree of the row. The higher the row correlation degree, the stronger the linear dependence between the signal dimension represented by the row and other dimensions. Then, the arithmetic mean of the row correlation degrees of all rows of the covariance matrix is calculated to obtain the matrix autocorrelation degree. The matrix autocorrelation degree is a global indicator that comprehensively represents the average linear correlation strength between the dimensions of the entire signal sequence. A higher matrix autocorrelation degree usually means that the internal structure of the signal sequence is tight, and the dimensions have a greater mutual influence.

[0087] The four key parameters, environmental interference entropy value (from previous analysis), principal component contribution of signal sequence, dispersion coefficient, and matrix autocorrelation degree, have been obtained. These parameters are input into a pre-set purity calculation model. The model is a trained multi-layer perceptron neural network structure. Before input, all parameters need to be normalized to eliminate dimensional differences and fall within the numerical range that the model can handle. The normalized parameters are used as the input layer node values of the neural network. The neural network contains one or more hidden layers, each composed of several neuron nodes. The nodes communicate information through connections with weights. Each neuron performs weighted summation on its inputs and generates output through a nonlinear activation function (such as ReLU or Sigmoid). Finally, the output layer nodes are processed by the activation function to generate a scalar output value, which is the signal transmission purity index. This index is a comprehensive quantitative evaluation result that combines the complexity of external electromagnetic environmental interference (environmental interference entropy value) and the structural characteristics of internal signal sequence (principal component contribution, dispersion coefficient, and matrix autocorrelation degree), ultimately providing a numerical basis for judging the purity of signal transmission under current channel conditions. The entire implementation process converts raw signal data and environmental interference information into a purity index with clear guiding significance through rigorous mathematical operations and model calculations, laying the foundation for subsequent adaptive anti-interference transmission control decisions. The formula is as follows:

[0088]

[0089] Where: represents the optimal segmentation threshold determined by the adaptive threshold segmentation algorithm; represents the variable that traverses all possible segmentation points of the eigenvalue sequence; represents the proportion of the number of principal eigenvalues to the total number of eigenvalues when is the segmentation point; represents the proportion of the number of principal eigenvalues to the total number of eigenvalues when The ratio of the number of secondary eigenvalues to the total number of eigenvalues when the split point is taken as The arithmetic mean of the primary eigenvalues when the split point is taken as The arithmetic mean of the primary eigenvalues when the split point is taken as The arithmetic mean of the secondary eigenvalues when the split point is taken as The arithmetic mean of the secondary eigenvalues when the split point is taken as. The goal of this formula is to find a split point such that the inter-class variance between the two subsets generated by the split (the primary eigenvalue set and the secondary eigenvalue set) is maximized.

[0090] In the anti-interference transmission control of the wireless low-power transducer, after the calculation of the signal transmission purity index is completed, the system executes dynamic transmission strategy adjustment according to the index value. The signal transmission purity index is a normalized value between 0 and 1, and the lower the value, the more serious the current environmental interference or the more unfavorable the signal structure for pure transmission, and more active anti-interference measures are needed; on the contrary, the higher the value, the better the transmission conditions.

[0091] The system internally presets a frequency hopping interval mapping table, which defines the adaptive frequency hopping interval values corresponding to different intervals of the signal transmission purity index. This mapping relationship is set based on the balance principle of channel quality and frequency hopping overhead. When the purity index is low, it means that the channel environment is poor or the signal is susceptible to interference, and more frequent frequency switching is needed to avoid interference, so the allocated frequency hopping interval is smaller. When the purity index is high, the channel conditions are good, and the frequency hopping interval can be appropriately increased to reduce unnecessary frequency switching overhead and potential channel discovery delay. The following table shows a typical mapping relationship:

[0092] Table 1: Mapping relationship between signal transmission purity index and adaptive frequency hopping interval.

[0093] Signal transmission purity index interval Adaptive frequency hopping interval (ms) Applicable scenario description [0.0,0.2) 5 Strong environmental interference, unstable signal structure [0.2,0.4) 10 Strong environmental interference, significant interference source exists [0.4,0.6) 25 Medium interference level, multiple interference sources exist [0.6,0.8) 50 Light interference, relatively stable channel conditions [0.8,1.0] 100 Weak interference, excellent channel conditions

[0094] The system looks up this mapping table according to the real-time calculated signal transmission purity index to determine the adaptive frequency hopping interval value to be used currently. After receiving this frequency hopping interval instruction, the frequency synthesizer module of the transducer starts the dynamic frequency adjustment mechanism. This mechanism includes a frequency hopping sequence generator and an accurate timer. The frequency hopping sequence is usually generated based on a pre-shared key or a specific algorithm to ensure synchronization between the transmitter and the receiver. The timer counts down according to the set adaptive frequency hopping interval, and when the count reaches zero, the frequency synthesizer immediately switches the carrier frequency of the transducer to the next predetermined frequency point in the frequency hopping sequence. This dynamic adjustment enables the transmission signal to actively avoid the current frequency band with severe interference and jump to a possibly cleaner frequency point for communication.

[0095] After switching to the new carrier frequency, the transmitter starts processing the raw data to be transmitted. In order to enhance the robustness of the transmission, the raw data stream is segmented into a series of fixed-length data blocks. The length of the data block is usually determined by a combination of factors such as the physical layer frame structure, modulation scheme, and expected bit error rate. The segmented data blocks are arranged in order to form a data block group. In order to facilitate identification, reorganization, and verification at the receiving end, the system attaches two key pieces of information to each data block: a unique cluster identification code and a hierarchical verification code.

[0096] The unique cluster identification code is an incremental sequence number that explicitly identifies the position of the data block in the raw data stream and the transmission batch to which it belongs. The hierarchical verification code contains two levels of verification information: the first level is a simple parity check code used to quickly detect whether an odd number of bit errors has occurred during data block transmission; the second level is a more powerful cyclic redundancy check code whose generating polynomial is selected based on the data block length and the expected error detection capability, used to detect more complex error patterns. The data block with the attached cluster identification code and hierarchical verification code constitutes a complete transmission cluster unit.

[0097] The transmitter strictly transmits in the hierarchical order of the transmission cluster units. The transmission order is based on the incremental order of the cluster identification code, ensuring that the receiving end can receive in order. During transmission, each transmission cluster unit is modulated and transmitted as an independent transmission frame structure. At the receiving end, the processing process starts with the binary data stream obtained after demodulating the received wireless signal. The receiving end first identifies and sorts the received data blocks based on the cluster identification code information contained in the transmission cluster unit frame structure. By analyzing the cluster identification code, the receiving end can determine the position of each data block in the original data block group and reorganize them in the correct order to restore the original data block group structure. After reorganization, the receiving end extracts the hierarchical verification code attached to each data block and starts the multi-level verification mechanism. First, perform the first level of verification, i.e. parity check. The receiving end recalculates the parity bit based on the received data block content and compares it with the received parity check code. If they are consistent, it is determined that the data block passes the parity check and the data transmission is likely correct. If the parity check fails, it indicates that at least one bit error has occurred in the data block during transmission, at which point the receiving end automatically activates the second level of verification mechanism, i.e. cyclic redundancy check.

[0098] The receiving end uses the same cyclic redundancy check generating polynomial as the sending end to recalculate the check code for the received data block content and compare it with the received cyclic redundancy check code. If the two match, the data block passes the cyclic redundancy check and the data is considered complete despite the parity check failure. If the cyclic redundancy check also fails, the data block is confirmed to have experienced an error during transmission that cannot be corrected by the check code. For data blocks that fail the cyclic redundancy check, the receiving end does not immediately discard or request retransmission of the entire data, but attempts data recovery. At this point, the receiving end analyzes the characteristic distribution of the entire data block group currently received. These characteristics include: the cluster identification code position of the error data block, the error pattern (such as consecutive bit error or random bit error), the number of error data blocks and their distribution within the group. Based on this characteristic information, the receiving end builds a check feature library.

[0099] According to the content of the check feature library, the receiving end matches and selects the best fault-tolerant decoding strategy. For example, if the number of error data blocks is small and the position is scattered, a forward error correction algorithm can be used to try to repair it; if the error is concentrated in a specific area, an interpolation or context-based data reconstruction algorithm can be used. The selected fault-tolerant decoding strategy is applied to the error data block to try to recover its original content. If the recovery is successful, the data block group is reconstructed; if the recovery fails, the data block group check failure is recorded and it is decided whether to request retransmission according to the subsequent process. The entire receiving end processing flow aims to maximize the use of received information, reduce unnecessary retransmission, and improve transmission efficiency.

[0100] In the fault-tolerant processing flow of the wireless low-power transmitter system, when the receiving end fails to recover the data through the fault-tolerant decoding strategy, the system enters a decision-making phase that takes into account the channel state and device energy consumption. The implementation of this phase relies on continuous monitoring of the current communication link quality and the energy reserves of the transmitter. The system monitors two key operating parameters in real time. The first parameter is the current channel quality indicator. This indicator is derived by analyzing the data of the receiving end physical layer and link layer, and its calculation combines the stability of the received signal strength indicator and the success rate of link layer frame reception. The received signal strength indicator reflects the degree of signal attenuation in space propagation, while the frame reception success rate directly reflects the reliability of data transmission. These two sub-indicators are weighted and fused to form a value that represents the current wireless channel transmission conditions. The lower the value, the worse the channel conditions, the higher the bit error rate, and the less reliable the transmission. The second parameter monitored is the remaining energy indicator of the transmitter. This indicator is provided by the power management unit built into the transmitter. The power management unit measures the battery voltage, discharge current and estimates the remaining capacity through a precision circuit, and finally outputs a value that reflects the current available energy level of the device. This value is directly related to the time the transmitter can continue to work and the number of additional communication operations it can support.

[0101] Based on the real-time acquired channel quality indicator, the system calculates the maximum number of data retransmissions allowed under the current channel condition. The calculation process refers to the pre-set mapping relationship between channel quality and maximum retransmission number. This mapping relationship is usually set as the worse the channel quality, the lower the maximum number of retransmissions allowed. This is because in poor channel conditions, the success probability of multiple retransmissions is still very low, which will consume valuable energy resources and increase transmission delay. The mapping relationship can be implemented in the form of a lookup table or a simple function. For example, when the channel quality indicator is below a certain threshold, the maximum number of retransmissions allowed is set to zero or a very small value; as the channel quality indicator improves, the maximum number of retransmissions allowed also increases accordingly.

[0102] According to the real-time acquired residual energy indicator, the system calculates and determines the minimum power consumption transmission mode to be adopted. The core goal of the minimum power consumption transmission mode is to maximize the energy consumption of a single communication operation. The implementation usually includes selecting the lowest transmission power level supported by the device, and adopting the simplest modulation and demodulation scheme with the lowest requirement for hardware processing capability. Although reducing the transmission power may shorten the communication distance or reduce the signal quality, and adopting a simplified modulation method may reduce the data transmission rate, in the energy-limited scenario, it is crucial to prioritize the endurance capability of the device.

[0103] The system maintains a counter inside to record the cumulative number of times that the fault-tolerant decoding strategy has tried to recover data but ultimately failed for the same data block group that is currently being processed. This failure number reflects the difficulty of recovering the data block group and the severity of the current channel condition. The decision logic is as follows: if the recorded number of fault-tolerant decoding failures has not reached the upper limit of the maximum number of retransmissions allowed calculated according to the current channel quality, the system will choose to initiate a local data retransmission request. This retransmission request is strictly executed in the minimum power consumption transmission mode determined in the preceding. Local retransmission means that the receiving end only requests to resend those data blocks that have failed the check and have not been successfully recovered by fault-tolerant decoding, rather than the entire original data block group. The request information is sent to the transmitter through the feedback channel. After receiving the retransmission request, the transmitter only re-modulates and sends the specified failed data blocks. After receiving the retransmitted data blocks, the receiving end again attempts to perform parity check, cyclic redundancy check, and attempts fault-tolerant decoding when the check fails. At the same time, the fault-tolerant decoding failure counter is updated according to the recovery result of this time.

[0104] If the number of failed fault-tolerant decoding has reached or exceeded the maximum allowed retransmission number, the system will not initiate a new retransmission request, but instead move on to the final processing decision. At this point, the system needs to decide on the subsequent operation based on the priority tag assigned to the data block group when it was segmented and encapsulated. The priority tag is usually attached to each data block group when the original data is segmented at the sending end, and is used to identify the importance level of the data group. For example, high priority can correspond to critical control instructions or sensor data with high real-time requirements; low priority can correspond to non-critical configuration information or historical data that can be delayed in transmission.

[0105] According to the priority tag, the system performs different operations: if the data block group is tagged as high priority, the system will attempt to perform degraded decoding. Degraded decoding is a lossy data reconstruction method that aims to recover as much usable information as possible in the case where complete and accurate data cannot be obtained. Specific strategies can include: using the information of other successfully received data blocks within the data block group for interpolation or extrapolation; using the correlation between data blocks for speculation; or simply extracting the passing bits of the received data, ignoring the error part. The result of degraded decoding may be incomplete or reduced accuracy data, but retains some information value. If the data block group is tagged as low priority, the system may choose to discard the data block group directly, without any further recovery attempt, to avoid further resource consumption. The discard operation will release the related buffer and may report a data loss event to the upper layer application.

[0106] The entire implementation process monitors the channel state and device energy dynamically, and calculates the key parameters (maximum allowed retransmission number, minimum power consumption mode) according to the preset strategy, and finally selects between retransmission request, degraded decoding or direct discard according to the number of failed fault-tolerant recovery and the importance of the data. This mechanism aims to seek a balance between data transmission reliability and device energy consumption life under the dual constraints of energy limitation and channel instability.

[0107] It should be noted that the relational terms such as first and second, and the like, are used solely to distinguish one from another entity or action, without necessarily requiring or implying any such actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0108] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely exemplary of the principles and application of the present application. Numerous modifications and adaptions can be effected without departing from the spirit and scope of the present application, which is not limited to the exact construction and arrangement described. It is intended, therefore, to cover all modifications and adaptions that fall within the scope of the claims and their equivalents.

Claims

1. A wireless low-power-consumption transmitter anti-interference transmission method, characterized in that, The method comprises the following steps: Collecting real-time electromagnetic interference signal spectrum of the environment where the transmitter is located, and extracting multi-dimensional interference features; Building an interference feature library based on the multi-dimensional interference features, and dividing interference feature clusters according to the distribution density and intensity variation trend of the features in the interference feature library; Calculating the dynamic fluctuation index and the frequency domain concentration index of each interference feature cluster; Obtaining the peak frequency drift index and the spectral bandwidth variation index of each interference feature cluster; Generating an environmental interference entropy value based on the peak frequency drift index, the spectral bandwidth variation index, the average value of the dynamic fluctuation index of all interference feature clusters, and the average value of the feature similarity between interference feature clusters; Building a covariance matrix of the current transmission signal sequence, performing eigenvalue decomposition on the covariance matrix to obtain each eigenvalue, and obtaining the principal component contribution degree and the dispersion coefficient of the eigenvalue sequence of the transmission signal sequence based on the distribution state, the concentration state and the dispersion state of the eigenvalue; Generating a signal transmission purity index based on the correlation between the row vectors in the covariance matrix, the environmental interference entropy value, the principal component contribution degree and the dispersion coefficient; Performing anti-interference transmission control based on the signal transmission purity index; The method for obtaining the dynamic fluctuation index comprises: For each interference feature cluster, locating the maximum spectral intensity point in the interference feature cluster, and forming a left spectral segment by combining the maximum spectral intensity point and the left spectral point, and forming a right spectral segment by combining the right spectral point of the maximum spectral intensity point; Obtaining the first-order gradient sequence of the left spectral segment of each interference feature cluster, and generating a gradient transformation sequence by processing the first-order gradient sequence using a nonlinear transformation function; Calculating the absolute value of the difference between the sum of the elements of the gradient transformation sequence and the length of the first-order gradient sequence of the left spectral segment as the fluctuation intensity index of the left spectral segment; Using the same method to obtain the fluctuation intensity index of the right spectral segment of each interference feature cluster; Calculating the product of the absolute value of the difference between the fluctuation intensity indexes of the left spectral segment and the right spectral segment and the average value as the dynamic fluctuation index of the interference feature cluster; Calculating the variance of the spectral intensity of the left spectral segment and the variance of the spectral intensity of the right spectral segment, and taking the average value of the variances of the two sides as the frequency domain concentration index; The method for obtaining the spectral bandwidth variation index comprises: Extracting the main peak intensity values in all interference feature clusters, and constructing a main peak intensity sequence by sorting the frequency point positions in the spectral sequence; Constructing a main peak frequency point sequence by sorting the frequency point positions of the main peaks in the interference feature clusters in size; Obtaining the first-order difference sequences of the main peak intensity sequence and the main peak frequency point sequence, and taking the variance of the first-order difference sequence of the main peak intensity sequence as the peak frequency drift index and the variance of the first-order difference sequence of the main peak frequency point sequence as the spectral bandwidth variation index; The method for generating the environmental interference entropy value comprises: Calculating the average value of the dynamic fluctuation indexes of all interference feature clusters as the average fluctuation index; Obtaining the average value of the spectral similarity coefficients between each interference feature cluster and all other interference feature clusters as the average feature cluster similarity; Inputting the peak frequency drift index, the spectral bandwidth variation index, the average fluctuation index and the average feature cluster similarity into an environmental interference entropy value calculation model to generate the environmental interference entropy value; The method for obtaining the principal component contribution degree and the dispersion coefficient comprises: Ranking all eigenvalues in descending order to construct an eigenvalue sequence; Inputting the eigenvalue sequence into an adaptive threshold segmentation algorithm to output a segmentation threshold, and listing eigenvalues greater than or equal to the segmentation threshold as main eigenvalues and eigenvalues less than the segmentation threshold as secondary eigenvalues; Calculating the proportion of the number of main eigenvalues in the total number of eigenvalues as the main component contribution degree; Calculating the product of the mean and variance of the main eigenvalues as the main eigenvalue fluctuation index, and calculating the secondary eigenvalue fluctuation index in the same way, and taking the absolute value of the difference between the main eigenvalue fluctuation index and the secondary eigenvalue fluctuation index as the dispersion coefficient; The method for generating the signal transmission purity index comprises: Calculating the absolute value of the correlation coefficient of each row vector in the covariance matrix and other row vectors as the row correlation degree; Taking the mean of all row correlation degrees of the covariance matrix as the matrix autocorrelation degree; Inputting the main component contribution degree, the dispersion coefficient, the matrix autocorrelation degree, and the environmental interference entropy value into a purity calculation model to generate the signal transmission purity index.

2. The anti-jamming transmission method of a wireless low-power consumption transducer according to claim 1, wherein, The method for dividing the interference feature cluster comprises: Using a spectrum peak detection algorithm to identify all main peaks and secondary peaks in the electromagnetic interference signal spectrum, segmenting the spectrum sequence from each secondary peak to form a feature subsequence, and calculating the spectrum energy range of each feature subsequence; Inputting the spectrum energy range of all feature subsequences into a clustering algorithm to output each feature clustering cluster; Calculating the internal spectrum energy mean of each feature clustering cluster, taking the feature clustering cluster with the largest internal spectrum energy mean as the dominant interference cluster, and taking the spectrum subsequence corresponding to the internal features of the dominant interference cluster as the interference feature cluster.

3. The anti-jamming transmission method of claim 1, wherein, The anti-interference transmission control based on the signal transmission purity index comprises: Calculating an adaptive frequency hopping interval according to the signal transmission purity index; Dynamically adjusting the transmitter carrier frequency based on the adaptive frequency hopping interval; Dividing the data to be transmitted into data block groups at the adjusted carrier frequency, adding a unique cluster identification code and a hierarchical check code to each data block to form a transmission cluster unit; Sending the transmission cluster units in hierarchical order, and performing multi-level checking at the receiving end.

4. The anti-jamming transmission method of claim 3, wherein, The multi-level checking comprises: The receiving end reorganizes the data block groups according to the cluster identification code and extracts the hierarchical check code; Performing a first-level parity check, and activating a second-level cyclic redundancy check if the check fails; If the cyclic redundancy check fails, generating a check feature library based on the feature distribution of the data block groups; According to the check feature library, the best fault-tolerant decoding strategy is matched to recover the data.

5. The anti-jamming transmission method of claim 4, wherein, The matching of the fault-tolerant decoding strategy comprises: Monitoring the current channel quality index and the transmitter residual energy index; According to the channel quality index, the maximum allowed retransmission times are calculated, and according to the residual energy index, the minimum power consumption transmission mode is calculated; If the number of failed data recovery times does not reach the maximum allowed retransmission times, initiate a local retransmission request using the minimum power consumption transmission mode; Otherwise, according to the priority mark of the data block group, select to discard or perform a degraded decoding.

Images