An anti-interference automatic synchronization frequency hopping method applicable to wireless ad hoc networks

Abstract

The invention relates to the field of wireless communication technology, and specifically discloses an anti-interference automatic synchronous frequency hopping method suitable for a wireless ad hoc network. By real-time acquisition of signal strength, signal-to-noise ratio and bit error rate data of communication frequencies, and by using advanced signal processing technologies such as empirical mode decomposition, fast Fourier transform and Haar wavelet transform, combined with machine learning algorithms such as K-means clustering and random forest model, key characteristic indicators reflecting the state of the frequency are calculated, a comprehensive characteristic vector is constructed, and it is predicted whether the frequency is interfered with in the future. When it is detected that the current frequency is interfered with, the system selects an uninterrupted frequency from a preset communication frequency pool for replacement, and continuously monitors the new frequency. If it is still interfered with for multiple times, it switches to a spare frequency hopping sequence. The invention not only improves the scientificity and accuracy of frequency selection, but also avoids interference, enhances the synchronization adjustment capability of the entire network, and improves the overall performance and anti-interference capability of the wireless ad hoc network.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and particularly to an anti-interference automatic synchronization frequency hopping method applicable to wireless ad hoc networks. Background Art

[0002] With the rapid development of wireless communication technologies, wireless ad hoc networks have been widely used in fields such as emergency response, military applications, and the Internet of Things due to their characteristic of enabling autonomous communication between nodes without relying on fixed infrastructure. However, the limited wireless communication spectrum resources are vulnerable to external interference, making it a great challenge to maintain high-quality communication in complex electromagnetic environments. Especially in high-density user scenarios or when there are multiple coexisting devices in the same frequency band, the signal interference problem is particularly prominent, seriously affecting the stability and reliability of the network. To solve these problems, researchers have developed various anti-interference technologies and frequency hopping mechanisms, aiming to avoid interference sources by dynamically adjusting communication frequency points to ensure the stability of the communication link. However, how to efficiently and accurately select and switch to the optimal frequency point and maintain the overall coordination and stability of the network during this process remains a key problem to be solved urgently.

[0003] The existing technologies have the following deficiencies:

[0004] The existing technologies still have many deficiencies in meeting the anti-interference requirements in wireless ad hoc networks. Firstly, traditional manual or simple rule-based frequency point selection methods lack scientific basis and forward-looking evaluation capabilities, making it difficult to effectively identify potential interference and make timely adjustments, resulting in frequent communication interruptions or quality problems. Secondly, existing dynamic adjustment mechanisms often fail to fully utilize advanced signal processing technologies and machine learning algorithms for multi-dimensional data analysis, resulting in inaccurate judgment of frequency point states and inability to comprehensively reflect the actual interference situation of frequency points. In addition, there are also defects in the whole-network synchronous adjustment of existing technologies. When a certain node detects interference and switches the frequency point, other nodes may not be able to obtain updated information in time, resulting in network fragmentation or data transmission failure. Finally, there is a lack of an effective intelligent feedback calibration mechanism to verify the effectiveness of the new frequency hopping sequence in real time, making it difficult to ensure continuous high-quality communication during network operation. These deficiencies limit the performance of wireless ad hoc networks in complex electromagnetic environments, and there is an urgent need for a more intelligent and efficient solution to improve their anti-interference capabilities and overall reliability. Summary of the Invention

[0005] The purpose of the present invention is to provide an anti-interference automatic synchronization frequency hopping method applicable to wireless ad hoc networks to solve the problems in the above background.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] An anti-interference automatic synchronization frequency hopping method applicable to wireless ad hoc networks, comprising the following steps:

[0008] S1: Real-time collect the signal strength data of the current communication frequency point, and judge whether the current communication frequency point is interfered according to the signal strength change amplitude and change frequency;

[0009] S2: According to the judgment result, divide the current communication frequency point into: an interfered communication frequency point and an un-interfered communication frequency point;

[0010] S3: Based on the interfered communication frequency point, the system selects an un-interfered communication frequency point from a pre-set communication frequency point pool to replace the interfered communication frequency point, and continuously monitors the replaced communication frequency point. If the replaced frequency point is determined to be an interfered communication frequency point for M consecutive times, automatically switch to the backup frequency hopping sequence;

[0011] S4: Based on the un-interfered communication frequency point, extract the signal-to-noise ratio feature and bit error rate feature of the current communication frequency point, and construct them into a comprehensive feature vector as the input of the comprehensive scoring model. According to the comprehensive scoring model, predict whether the communication frequency point will be interfered in the future for a period of time, and preferentially select the frequency point that remains un-interfered for a long time in the future as the new communication frequency point.

[0012] As a further solution of the present invention: The judgment of whether the current communication frequency point is interfered specifically includes:

[0013] Real-time collect the signal strength of the current communication frequency point, calculate the amplitude coefficient and frequency coefficient respectively according to the signal strength change amplitude and change frequency, perform normalization calculation processing on the amplitude coefficient and frequency coefficient to obtain the interference coefficient, and judge whether the interference coefficient of the current communication frequency point is greater than or equal to the threshold. If so, mark it as an interfered communication frequency point, otherwise mark it as an un-interfered communication frequency point.

[0014] As a further solution of the present invention: The acquisition process of the amplitude coefficient is:

[0015] Real-time collect the signal strength data of the current communication frequency point according to the time series;

[0016] Perform empirical mode decomposition on the collected signal strength data to obtain a series of intrinsic mode functions and a residual term; based on the energy level of each intrinsic mode function, select the intrinsic mode function with the highest energy as the main component reflecting the signal strength change;

[0017] For the selected main intrinsic mode function, calculate the difference between the maximum value and the minimum value of the intrinsic mode function to obtain the signal strength change amplitude, and calculate the ratio of the signal strength change amplitude to the average signal strength within the time window to obtain the amplitude coefficient.

[0018] As a further solution of the present invention: The process of obtaining the frequency coefficient is as follows:

[0019] Collect the signal strength data of the current communication frequency point in real time according to the time series;

[0020] Apply the Haar wavelet transform to the collected signal strength data and decompose it into approximation coefficients and detail coefficients at different scales;

[0021] By calculating the average value after summing the squares of all detail coefficients at each scale, obtain the energy level of the detail coefficients at each scale;

[0022] Multiply the energy level of each scale by its corresponding scale number and then divide by the sum of the energy levels at all scales to obtain the frequency coefficient.

[0023] As a further solution of the present invention: If the replaced frequency point is continuously determined to be an interfered communication frequency point for M times, it will automatically switch to the standby frequency hopping sequence, which specifically includes:

[0024] If the communication frequency point is continuously determined to be interfered for M times, immediately select an undisturbed frequency point from the preset communication frequency point pool as a replacement;

[0025] During the replacement process, the system not only updates the local frequency point configuration, but also broadcasts the new frequency point information to adjacent nodes through the frequency hopping channel to ensure that the entire network can be synchronized and adjusted;

[0026] If the performance of the new sequence does not meet the standard, immediately adjust and regenerate the alternative sequence until the performance of the new sequence meets the standard.

[0027] As a further solution of the present invention: Predicting whether the communication frequency point will be interfered in the next period of time specifically includes:

[0028] Based on the undisturbed communication frequency points, obtain the signal-to-noise ratio data of the current communication frequency point according to the time series, calculate the difference between the signal-to-noise ratio of the current communication frequency point and the preset threshold, calculate the signal-to-noise ratio anomaly coefficient according to the change degree of the signal-to-noise ratio difference, obtain the bit error rate data of the current communication frequency point according to the time series, calculate the bit error rate anomaly coefficient according to the change rate of the bit error rate data within the time series, construct the signal-to-noise ratio anomaly coefficient and the bit error rate anomaly coefficient into a comprehensive feature vector as the input of the comprehensive scoring model, and predict whether the communication frequency point will be interfered in the next period of time according to the output of the model.

[0029] As a further solution of the present invention: The process of obtaining the signal-to-noise ratio anomaly coefficient is as follows:

[0030] Based on the undisturbed communication frequency points, use a wireless network analyzer to collect the signal-to-noise ratio data of the communication frequency point in real time at a frequency of once per second to form a time series data;

[0031] Perform a fast Fourier transform on the collected signal-to-noise ratio time series data to convert it to the frequency domain for analysis; based on the fast Fourier transform results, calculate the squared amplitude of each frequency component to obtain the power spectral density of each frequency component.

[0032] Calculate the mean and variance of the signal-to-noise ratio from the time domain data. According to the calculated mean of the signal-to-noise ratio, calculate the difference between the mean of the signal-to-noise ratio and a preset threshold. According to the degree of change of the signal-to-noise ratio difference, calculate the ratio of the absolute value of the difference between the mean of the signal-to-noise ratio and the preset threshold to the variance to obtain the signal-to-noise ratio anomaly coefficient.

[0033] As a further solution of the present invention: the process of obtaining the bit error rate anomaly coefficient is as follows:

[0034] Based on the undisturbed communication frequency points, use a wireless network analyzer to collect the bit error rate data of the communication frequency points in real time at a frequency of once per second to form a time series data.

[0035] For each time point, calculate the difference between the current bit error rate and the bit error rate of the previous time point, and calculate the ratio with the time interval to obtain the change rate of the bit error rate.

[0036] Take the change rate of the bit error rate as the key feature of the bit error rate data and use it as input data for K-means clustering analysis. Select the number of clusters G and initialize the centroid positions. Perform clustering by minimizing the squared distance between all points and the centroids of their respective clusters. Assign the data points to the nearest cluster and update the positions of the cluster centroids until convergence.

[0037] Based on the K-means clustering results, identify the abnormal points. For the change rate of the bit error rate at each time point, calculate its distance to the nearest cluster center and calculate the ratio with the maximum value of all distances to obtain the bit error rate anomaly coefficient.

[0038] As a further solution of the present invention: the process of constructing the comprehensive scoring model is as follows:

[0039] Based on each undisturbed communication frequency point, obtain the signal-to-noise ratio anomaly coefficient and the bit error rate anomaly coefficient, construct a comprehensive feature vector as the input of the comprehensive scoring model. Take minimizing the error between predicting whether the communication frequency point will be disturbed in the future period and whether the communication frequency point is actually disturbed in the future period as the training objective, train the comprehensive scoring model. According to the trained comprehensive scoring model, the output of the comprehensive scoring model is whether the current communication frequency point is a disturbed communication frequency point or an undisturbed communication frequency point, and the comprehensive scoring model is a random forest model.

[0040] The beneficial effects of the present invention:

[0041] (1) The present invention realizes the precise quantification and intelligent prediction of the communication frequency point status by integrating real-time data acquisition and advanced signal processing technologies, including empirical mode decomposition, fast Fourier transform, and Haar wavelet transform, as well as machine learning algorithms such as K-means clustering and random forest models. Specifically, the system first uses a wireless network analyzer to collect the signal strength, signal-to-noise ratio, and bit error rate data of each frequency point in real time at a second-level frequency, and analyzes the main components of the signal strength change through EMD, converts the signal-to-noise ratio to the frequency domain through FFT to reveal its internal pattern, and captures the high-frequency fluctuation characteristics in the bit error rate through Haar wavelet transform. Subsequently, these multi-dimensional features are constructed into a comprehensive feature vector and fed into a trained random forest model as input to predict whether the frequency point will be interfered with in the future for a period of time. This method not only significantly improves the scientificity and accuracy of frequency point selection, but also can prospectively evaluate potential interference situations, ensuring that the network provides stable and high-quality communication services during long-term future operation. Compared with traditional manual or simple rule-based frequency point selection methods, the present invention, with its highly intelligent data processing ability and accurate prediction model, can more effectively identify and avoid interference sources, greatly improving communication quality and network reliability. Especially in a complex and changeable electromagnetic environment, it provides strong anti-interference ability and optimization performance guarantee for wireless ad hoc networks. Therefore, the present invention shows significant advantages in improving communication efficiency and reducing the impact of interference, and is a key technological innovation for realizing efficient and reliable wireless communication.

[0042] (2) The present invention introduces an advanced dynamic adjustment mechanism. When it detects that the current communication frequency point is interfered with, the system can intelligently select an undisturbed frequency point from a pre-set communication frequency point pool for replacement and continuously monitor the new frequency point. If the new frequency point still shows interference in consecutive detections, it will automatically switch to a standby frequency hopping sequence to ensure the stability and reliability of the communication link. This process not only relies on real-time data acquisition and analysis, but also combines an intelligent feedback calibration mechanism and a sliding window technique to verify the effectiveness of the new frequency hopping sequence, so as to still maintain high-efficiency and stable communication quality in a high-interference environment. Through this mechanism, the wireless ad hoc network can show excellent adaptability and robustness in a complex and changeable electromagnetic environment, significantly reducing communication interruptions or quality problems caused by interference. More importantly, this method supports full-network synchronous adjustment, broadcasts the updated frequency point information through the frequency hopping channel, and ensures that all nodes can be synchronized in time, enhancing the overall coordination and stability of the network. This highly intelligent dynamic adjustment strategy not only improves the anti-interference ability of individual nodes, but also optimizes the performance of the entire network, enabling the wireless ad hoc network to have stronger self-repair and self-optimization capabilities when facing the challenges of a complex electromagnetic environment, providing a solid technical guarantee for realizing seamless and high-quality data transmission. Brief Description of the Drawings

[0043] The present invention will be further described below with reference to the accompanying drawings.

[0044] Figure 1 It is a flow chart of the specific steps of an anti-interference automatic synchronization frequency hopping method applicable to wireless ad hoc networks according to the present invention. Specific embodiments

[0045] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0046] Please refer to Figure 1 As shown, the present invention is an anti-interference automatic synchronization frequency hopping method applicable to wireless ad hoc networks, including the following steps:

[0047] S1: Real-time collect the signal strength data of the current communication frequency point, and judge whether the current communication frequency point is interfered according to the change amplitude and change frequency of the signal strength;

[0048] S2: According to the judgment result, divide the current communication frequency point into: interfered communication frequency points and non-interfered communication frequency points;

[0049] S3: Based on the interfered communication frequency points, the system selects non-interfered communication frequency points from a pre-set communication frequency point pool to replace the interfered communication frequency points, and continuously monitors the replaced communication frequency points. If the replaced frequency points are continuously determined to be interfered communication frequency points for M times, automatically switch to the standby frequency hopping sequence;

[0050] S4: Based on the non-interfered communication frequency points, extract the signal-to-noise ratio characteristics and bit error rate characteristics of the current communication frequency point, and construct them into a comprehensive feature vector as the input of the comprehensive scoring model. According to the comprehensive scoring model, predict whether the communication frequency point will be interfered in the next period of time, and preferentially select the frequency points that remain non-interfered for a long time in the future as the new communication frequency points.

[0051] In S1, real-time collect the signal strength data of the current communication frequency point, and judge whether the current communication frequency point is interfered according to the change amplitude and change frequency of the signal strength, specifically including:

[0052] Use a wireless network analyzer with signal strength measurement function, configure it to monitor a specific communication frequency point, ensure that the device settings include the correct center frequency, bandwidth and the required resolution bandwidth to meet the requirements of signal capture, set an automatic collection script through its supporting software development kit, and collect the signal strength data of the specific communication frequency point in real time at a frequency of once per second.

[0053] Real-time collect the signal strength of the current communication frequency point. According to the change amplitude and change frequency of the signal strength, calculate the amplitude coefficient and frequency coefficient respectively. Perform normalization calculation processing on the amplitude coefficient and frequency coefficient to obtain the interference coefficient. Determine whether the interference coefficient of the current communication frequency point is greater than or equal to the threshold. If so, mark it as an interfered communication frequency point; if not, mark it as an un-interfered communication frequency point;

[0054] The process of obtaining the amplitude coefficient is as follows:

[0055] Collect the signal strength data of the current communication frequency point in real time according to the time series;

[0056] Perform empirical mode decomposition on the collected signal strength data to obtain a series of intrinsic mode functions and a residual term; Based on the energy level of each intrinsic mode function, select the intrinsic mode function with the highest energy as the main component reflecting the signal strength change;

[0057] Among them, the calculation expression of empirical mode decomposition is: ;

[0058] In the formula, represents the signal strength value measured at time , represents the time acquisition point, represents the number of intrinsic mode functions, represents the total number of intrinsic mode functions, represents the residual term, represents the th intrinsic mode function;

[0059] Among them, the calculation expression of the energy level of each intrinsic mode function is: ;

[0060] In the formula, represents the start time point of the acquisition time period, represents the end time point of the acquisition time period, represents the energy level of the th intrinsic mode function;

[0061] For the selected main intrinsic mode function, calculate the difference between the maximum value and the minimum value of the intrinsic mode function to obtain the signal strength change amplitude. Calculate the ratio of the signal strength change amplitude to the average signal strength within the time window to obtain the amplitude coefficient.

[0062] It should be noted that: Through the empirical mode decomposition technology, the components that can best reflect the signal fluctuation characteristics can be extracted from the complex signal environment, and the amplitude coefficient can be calculated accordingly to help the system accurately judge the frequency point state. This method improves the anti-interference ability and enhances the adaptability and stability of the wireless ad hoc network in complex environments.

[0063] The process of obtaining the frequency coefficient is as follows:

[0064] Collect the signal strength data of the current communication frequency point in real time according to the time series;

[0065] Apply the Haar wavelet transform to the collected signal strength data and decompose it into approximation coefficients and detail coefficients at different scales;

[0066] Calculate the energy level of the detail coefficients at each scale, and the calculation expression is: ;

[0067] In the formula, represents the th scale, represents the number of detail coefficients, represents the th detail coefficient at the th scale, represents the th energy level of the detail coefficients at the th scale, represents the total number of detail coefficients at the th scale, represents the

[0068] th detail coefficient at the

[0069] It should be noted that: By applying the Haar wavelet transform to perform multi-resolution analysis on the signal strength data, the high-frequency components in the signal can be captured more precisely, and the frequency coefficient can be calculated accordingly. The frequency coefficient not only reflects the frequency characteristics of the signal strength changing with time, but also provides an important basis for evaluating whether the communication frequency point is interfered.

[0070] The calculation process of the interference coefficient is as follows:

[0071] ;

[0072] In the formula, represents the interference coefficient, represents the amplitude coefficient, represents the frequency coefficient, and represent preset coefficients, and and are both greater than 0.

[0073] In S3, based on the interfered communication frequency points, the system selects non-interfered communication frequency points from a pre-set communication frequency point pool to replace the interfered communication frequency points, and continuously monitors the replaced communication frequency points. If the replaced frequency points are determined to be interfered communication frequency points for M consecutive times, the system automatically switches to a backup frequency hopping sequence, which specifically includes:

[0074] When it is detected that the current communication frequency point is interfered, the system selects a non-interfered communication frequency point from the pre-set communication frequency point pool for replacement and continuously monitors the replaced communication frequency point; specifically, it includes the following steps: First, the system evaluates the interference status of each communication frequency point in real time. Once it is found that the currently used communication frequency point is determined to be interfered for M consecutive times, a non-interfered frequency point is immediately selected from the pre-set communication frequency point pool as a replacement.

[0075] During the replacement process, the system not only updates the local frequency point configuration but also broadcasts the new frequency point information to adjacent nodes through the frequency hopping channel to ensure that the entire network can be adjusted synchronously. In addition, for the replaced communication frequency point, the system sets a continuous monitoring period. If the frequency point is determined to be interfered in M consecutive checks, the backup frequency hopping sequence switching mechanism is triggered, and the intelligent feedback calibration mechanism, that is, dynamically adjusts the frequency hopping sequence according to the actual communication quality feedback from the receiving end, and uses the sliding window technology to verify the effectiveness of the new sequence to ensure high-quality data transmission even in a high-interference environment. This method greatly enhances the ability of the wireless ad hoc network to cope with complex electromagnetic environments and improves the overall network performance.

[0076] The automatic switching to the backup frequency hopping sequence includes:

[0077] If the communication frequency point is determined to be interfered for M consecutive times, a non-interfered frequency point is immediately selected from the pre-set communication frequency point pool as a replacement;

[0078] During the replacement process, the system not only updates the local frequency point configuration but also broadcasts the new frequency point information to adjacent nodes through the frequency hopping channel to ensure that the entire network can be adjusted synchronously;

[0079] If the performance of the new sequence does not meet the standard, it is immediately adjusted and a new alternative sequence is generated until the performance of the new sequence meets the standard.

[0080] In S4, based on the non-interfered communication frequency points, the signal-to-noise ratio feature and bit error rate feature of the current communication frequency point are extracted and constructed into a comprehensive feature vector as the input of the comprehensive scoring model. According to the comprehensive scoring model, it is predicted whether the communication frequency point will be interfered in the next period of time, and the frequency points that remain non-interfered for a long time in the future are preferentially selected as the new communication frequency points, which specifically includes:

[0081] Based on the undisturbed communication frequency points, obtain the signal-to-noise ratio (SNR) data of the current communication frequency point according to the time series, calculate the difference between the SNR of the current communication frequency point and the preset threshold, calculate the SNR anomaly coefficient according to the change degree of the SNR difference, obtain the bit error rate (BER) data of the current communication frequency point according to the time series, calculate the BER anomaly coefficient according to the change rate of the BER data within the time series, construct the SNR anomaly coefficient and the BER anomaly coefficient into a comprehensive feature vector as the input of the comprehensive scoring model, and the output of the model is whether the communication frequency point will be disturbed in a future period of time.

[0082] The process of obtaining the SNR anomaly coefficient is as follows:

[0083] Based on the undisturbed communication frequency points, use a wireless network analyzer to collect the SNR data of the communication frequency point in real time at a frequency of once per second to form a time series data;

[0084] Perform a fast Fourier transform on the collected SNR time series data to convert it to the frequency domain for analysis; based on the fast Fourier transform result, calculate the power spectral density of each frequency component, and the calculation formula is: ;

[0085] In the formula, represents the number of frequency components, represents the th frequency component, represents the th frequency component's complex value, represents the th frequency component's power spectral density;

[0086] Calculate the mean and variance of the SNR from the time domain data, calculate the difference between the SNR mean and the preset threshold according to the calculated SNR mean, calculate the ratio of the absolute value of the difference between the SNR mean and the preset threshold to the variance according to the change degree of the SNR difference, and obtain the SNR anomaly coefficient.

[0087] It should be noted that: by collecting the SNR data of the communication frequency point in real time and applying the fast Fourier transform for frequency domain analysis, not only can the SNR anomaly situation be accurately identified, but also a quantitative index can be provided to guide subsequent operations. The system can effectively identify and quantify the SNR anomaly situation, ensure communication quality, and improve the overall network performance.

[0088] The process of obtaining the BER anomaly coefficient is as follows:

[0089] Use a wireless network analyzer to collect the BER data of a specific communication frequency point in real time at a frequency of once per second to form a time series data;

[0090] For each time point, calculate the difference between the current bit error rate and the bit error rate of the previous time point, and calculate the ratio with the time interval to obtain the change rate of the bit error rate;

[0091] Take the change rate of the bit error rate as the key feature of the bit error rate data, use it as input data for K-means clustering analysis, select the number of clusters G and initialize the centroid position, perform clustering by minimizing the squared distance between all points and the centroid of their respective clusters, assign the data points to the nearest cluster, and update the position of the cluster centroid until convergence;

[0092] Based on the K-means clustering results, identify the outlier points. For the change rate of the bit error rate at each time point, calculate its distance to the nearest cluster center, and calculate the ratio with the maximum value of all distances to obtain the bit error rate anomaly coefficient.

[0093] It should be noted that: by collecting the bit error rate data of communication frequency points in real time, using the K-means clustering algorithm to identify the abnormal patterns in the bit error rate, and calculating the bit error rate anomaly coefficient, the refined evaluation of the communication link quality is realized.

[0094] The construction process of the comprehensive scoring model is as follows:

[0095] Based on each undisturbed communication frequency point, obtain the signal-to-noise ratio anomaly coefficient and the bit error rate anomaly coefficient, construct a comprehensive feature vector, use it as the input of the comprehensive scoring model, take minimizing the error between predicting whether the communication frequency point will be disturbed in the next period of time and whether the communication frequency point is actually disturbed in the next period of time as the training objective, train the comprehensive scoring model. According to the trained comprehensive scoring model, the output of the comprehensive scoring model is whether the current communication frequency point is a disturbed communication frequency point or an undisturbed communication frequency point. The comprehensive scoring model is a random forest model;

[0096] The training process of the random forest model is as follows:

[0097] Collect historical data to construct a training set. Each sample contains a comprehensive feature vector and its corresponding label (disturbed or undisturbed). If the communication frequency point is determined to be disturbed in the next period of time, the label is set to 1; otherwise, it is set to 0. Use the random forest algorithm to train the training set. The random forest samples the training data to generate multiple sub-sample sets, and trains a decision tree on each sub-sample set; the final prediction result is determined by the voting of all decision trees;

[0098] According to the trained comprehensive scoring model, input the comprehensive feature vector of the current communication frequency point. The model outputs whether the current communication frequency point is a disturbed communication frequency point or an undisturbed communication frequency point. Prioritize selecting the frequency points that remain undisturbed in the long term in the future as new communication frequency points.

[0099] It should be noted that: by collecting the signal-to-noise ratio (SNR) and bit error rate (BER) data of communication frequency points in real time, calculating the corresponding anomaly coefficients, and constructing them into a comprehensive feature vector as the input, a random forest model is trained to predict whether the communication frequency points will be interfered in a future period of time. This method can not only accurately identify potential interference situations in the communication link, but also provide quantitative indicators to guide subsequent operations.

[0100] The working principle of the present invention: By collecting the signal strength data of the current communication frequency point in real time, calculating the interference coefficient according to the signal strength change amplitude and frequency, and judging whether the frequency point is interfered. Subsequently, the communication frequency points are divided into two categories: interfered and non-interfered, and the interfered frequency points are replaced. When it is detected that the current frequency point is interfered, the system selects an un-interfered frequency point from the pre-set communication frequency point pool for replacement, and continuously monitors the replaced frequency point; if it is determined to be interfered for M consecutive times, the standby frequency hopping sequence switching mechanism is triggered. Further, based on the un-interfered frequency points, SNR and BER features are extracted and constructed into a comprehensive feature vector as the input of the random forest model to predict whether the frequency points will be interfered in a future period of time. Specifically, a wireless network analyzer is used to collect SNR and BER data in real time, calculate the SNR anomaly coefficient and the BER anomaly coefficient respectively, and combine the fast Fourier transform and the K-means clustering algorithm to achieve accurate quantification. The trained random forest model outputs the prediction result of the frequency point status, and preferentially selects the frequency points that remain un-interfered in the future for a long time as the new communication frequency points. This method not only improves the scientificity and accuracy of frequency point selection, but also enhances the network's ability to cope with interference, ensures high-quality data transmission, and significantly improves the overall performance and stability of the wireless ad hoc network. In addition, the effectiveness of the new sequence is verified through an intelligent feedback calibration mechanism and a sliding window technique, and high-efficiency and stable communication quality can be maintained even in a high-interference environment.

[0101] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0102] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired or wireless (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that contains one or more collections of available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0103] It should be understood that the term "and / or" in this document is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context before and after.

[0104] It should be understood that in various embodiments of the present application, the magnitudes of the sequence numbers of the above processes do not imply the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0105] The above has described in detail one embodiment of the present invention, but the content described is only the preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of the application of the present invention should still fall within the scope covered by the patent of the present invention.

Claims

1. An anti-interference automatic synchronous frequency hopping method suitable for wireless ad hoc networks, characterized in that: The following steps are involved: S1: collect the signal strength data of the current communication frequency in real time, and judge whether the current communication frequency is interfered according to the amplitude and frequency of signal strength change; Wherein, the amplitude coefficient and the frequency coefficient are calculated respectively according to the amplitude and frequency of signal strength change; The process of obtaining the amplitude coefficient is as follows: Collect signal strength data of current communication frequency in real time according to time series; Performing empirical mode decomposition on the collected signal strength data to obtain a series of intrinsic mode functions and a residual term; based on the energy level of each intrinsic mode function, selecting the intrinsic mode function with the highest energy as the main component reflecting the change of signal strength; For the selected main intrinsic mode function, the maximum value and the minimum value of the intrinsic mode function are calculated to obtain the signal strength change amplitude, and the signal strength change amplitude is calculated to obtain the amplitude coefficient by calculating the ratio of the signal strength change amplitude to the average signal strength in the time window; The process of obtaining the frequency coefficient is as follows: Collect signal strength data of current communication frequency in real time according to time series; Apply Haar wavelet transform to the acquired signal strength data to decompose it into approximate coefficients and detail coefficients at different scales; The energy level of the detail coefficient at each scale is obtained by calculating the square sum of all detail coefficients at each scale and taking the average value; The frequency coefficient is obtained by multiplying the energy level of each scale by its corresponding scale number and then dividing it by the sum of the energy levels at all scales; S2: According to the judgment result, the current communication frequency point is divided into: an interfered communication frequency point and an uninterfered communication frequency point; S3: Based on the interfered communication frequency, the system selects an uninterfered communication frequency from a preset communication frequency pool to replace the interfered communication frequency, and continuously monitors the replaced communication frequency. If the replaced frequency is determined to be an interfered communication frequency for M consecutive times, it automatically switches to the backup frequency hopping sequence; S4: Based on the undisturbed communication frequency, the signal-to-noise ratio characteristics and bit error rate characteristics of the current communication frequency are extracted and constructed into a comprehensive feature vector as the input of the comprehensive scoring model. According to the comprehensive scoring model, it is predicted whether the communication frequency will be interfered in the future, and the frequency that will remain undisturbed for a long time in the future is preferentially selected as the new communication frequency.

2. The anti-interference automatic synchronous frequency hopping method applicable to a wireless ad hoc network according to claim 1, characterized in that: The determining whether the current communication frequency point is interfered with specifically includes: The amplitude coefficient and the frequency coefficient are normalized and calculated to obtain the interference coefficient, and it is determined whether the interference coefficient of the current communication frequency is greater than or equal to the threshold. If so, it is marked as an interfered communication frequency, otherwise it is marked as an undisturbed communication frequency.

3. The anti-interference automatic synchronous frequency hopping method applicable to a wireless ad hoc network according to claim 1, characterized in that: If the replaced frequency point is determined to be the interfered communication frequency point for M consecutive times, automatically switching to the standby frequency hopping sequence specifically includes: If a communication frequency is determined to be interfered with for M consecutive times, a non-interfered frequency is immediately selected from the preset communication frequency pool as a replacement; During the replacement process, the system not only updates the local frequency configuration, but also broadcasts the new frequency information to adjacent nodes through the frequency hopping channel to ensure that the entire network can be adjusted synchronously; If the performance of the new sequence does not meet the standard, the alternative sequence is immediately adjusted and regenerated until the performance of the new sequence meets the standard.

4. The anti-interference automatic synchronous frequency hopping method applicable to wireless ad hoc networks according to claim 1, characterized in that: The prediction of whether the communication frequency point in the future period is interfered specifically includes: Based on the undisturbed communication frequency, the signal-to-noise ratio data of the current communication frequency is obtained according to the time series, the difference between the signal-to-noise ratio of the current communication frequency and the preset threshold is calculated, and the signal-to-noise ratio anomaly coefficient is calculated according to the degree of change of the signal-to-noise ratio difference. The bit error rate data of the current communication frequency is obtained according to the time series, and the bit error rate anomaly coefficient is calculated according to the change rate of the bit error rate data in the time series. The signal-to-noise ratio anomaly coefficient and the bit error rate anomaly coefficient are constructed into a comprehensive feature vector as the input of the comprehensive scoring model. According to the output of the model, it is predicted whether the communication frequency will be interfered in the future.

5. The anti-interference automatic synchronous frequency hopping method applicable to wireless ad hoc networks according to claim 4, characterized in that: The process of obtaining the signal-to-noise ratio anomaly coefficient is as follows: Based on the undisturbed communication frequency point, a wireless network analyzer is used to collect the signal-to-noise ratio data of the communication frequency point in real time at a frequency of once per second to form a time series data; Perform fast Fourier transform on the collected signal-to-noise ratio time series data to convert it into the frequency domain for analysis; based on the fast Fourier transform results, calculate the square of the amplitude of each frequency component to obtain the power spectrum density of each frequency component; The mean and variance of the signal-to-noise ratio are calculated from the time domain data. According to the calculated mean signal-to-noise ratio, the difference between the mean signal-to-noise ratio and the preset threshold is calculated. According to the degree of change of the signal-to-noise ratio difference, the ratio of the absolute value of the difference between the mean signal-to-noise ratio and the preset threshold to the variance is calculated to obtain the signal-to-noise ratio anomaly coefficient.

6. The anti-interference automatic synchronous frequency hopping method applicable to wireless ad hoc networks according to claim 4, characterized in that: The process of obtaining the bit error rate abnormality coefficient is as follows: Based on the undisturbed communication frequency point, a wireless network analyzer is used to collect the bit error rate data of the communication frequency point in real time at a frequency of once per second to form a time series data; For each time point, the difference between the current bit error rate and the bit error rate at the previous time point is calculated, and the ratio is calculated with the time interval to obtain the rate of change of the bit error rate; The bit error rate change rate is used as the key feature of the bit error rate data and as input data for K-means clustering analysis. The number of clusters G is selected and the centroid position is initialized. Clustering is performed by minimizing the square distance between all points and the centroid of their clusters. Data points are assigned to the nearest cluster and the position of the cluster centroid is updated until convergence. Based on the K-means clustering results, outliers are identified. For the rate of change of the bit error rate at each time point, the distance to the nearest cluster center is calculated, and the ratio is calculated with the maximum value of all distances to obtain the bit error rate anomaly coefficient.

7. The anti-interference automatic synchronous frequency hopping method applicable to wireless ad hoc networks according to claim 4, characterized in that: The construction process of the comprehensive scoring model is as follows: Based on each undisturbed communication frequency, the signal-to-noise ratio anomaly coefficient and the bit error rate anomaly coefficient are obtained, and a comprehensive feature vector is constructed as the input of the comprehensive scoring model. The comprehensive scoring model is trained with the minimization of the error between the prediction of whether the communication frequency in the future period is interfered with and the actual communication frequency in the future period being interfered as the training goal. According to the trained comprehensive scoring model, the output of the comprehensive scoring model is whether the current communication frequency is an interfered communication frequency or an undisturbed communication frequency. The comprehensive scoring model is a random forest model.

Images