Communication iron tower data acquisition monitoring system based on multi-sensor integration

Abstract

The invention relates to the technical field of big data analysis, and provides a communication iron tower data acquisition monitoring system based on multi-sensor integration, which comprises the following steps: acquiring temperature data, humidity data and infrared light signal data of a communication iron tower, and dividing detection intervals; calculating the temperature and humidity data unbalance degree and the infrared light harmonic abnormity of the monitoring interval, and obtaining the iron tower environment abnormity degree of the monitoring interval; calculating the peak period tendency and the noise interference degree of the monitoring interval; and obtaining a monitoring result of the communication iron tower data according to the iron tower environment anomaly degree and the noise interference degree of all the monitoring intervals. According to the invention, the accuracy of communication tower data monitoring can be improved.

Description

Technical Field

[0001] This invention relates to the field of big data analytics, specifically to a communication tower data acquisition and monitoring system based on multi-sensor integration. Background Technology

[0002] Communication towers are core infrastructure for signal transmission in mobile communications, broadcasting, and television, and their operational status directly affects the stability of signal coverage and the quality of communication services. Integrating various types of sensors onto communication towers, combined with remote data transmission and intelligent analysis platforms, enables refined, automated, and intelligent monitoring of tower status, ensuring the reliability of communication networks.

[0003] However, the frequency of fault and abnormal events is much lower than that of normal conditions. This leads to a serious data imbalance between the abnormal data and normal data detected by multiple sensors. When using classification decision algorithms to detect communication tower data, the model boundary of the classification decision algorithm is easily shifted to the abnormal data, which greatly increases the false negative rate and causes a few abnormal fault types to be missed. Summary of the Invention

[0004] This invention provides a communication tower data acquisition and monitoring system based on multi-sensor integration to solve the problem of decision boundary shift caused by the imbalance between abnormal and normal data, leading to misjudgment of abnormal multi-sensor data in communication towers. The specific technical solution adopted is as follows: One embodiment of the present invention provides a communication tower data acquisition and monitoring system based on multi-sensor integration, the system comprising the following modules: The data acquisition module is used to collect temperature data, humidity data, and infrared light signal data from the communication tower. The environmental anomaly assessment module is used to calculate the temperature and humidity data imbalance of the monitoring interval based on the time correlation and similarity differences of the ratio of temperature and humidity data in adjacent monitoring intervals. Based on all high-frequency components of the infrared light signal data in the frequency domain within the detection interval, and the differences in energy values ​​of different frequency components, it calculates the infrared light harmonic anomaly of the detection interval. This reflects the high-frequency energy dispersion and harmonic amplification of the infrared light signal data in the frequency domain within the monitoring interval. Based on the temperature and humidity data imbalance and the infrared light harmonic anomaly, it calculates the tower environmental anomaly degree of the monitoring interval. The noise interference evaluation module is used to calculate the peak periodicity trend of the monitoring interval based on the time interval between abrupt changes in the temperature data, humidity data, and infrared light signal data of the detection interval, as well as the changing trend of the abrupt changes. It also calculates the noise interference degree of the monitoring interval by combining the energy differences of adjacent harmonic frequencies in the frequency domain of the temperature data, humidity data, and infrared light signal data of the detection interval. The monitoring result acquisition module is used to acquire monitoring results of communication tower data based on the tower environment anomaly and noise interference levels in all monitoring intervals.

[0005] Furthermore, the method for obtaining the temperature and humidity data imbalance in the monitoring interval is as follows: The negative correlation result of the similarity between temperature data and humidity data in the detection interval is recorded as the negative correlation strength between temperature and humidity in adjacent monitoring intervals. Based on the temporal correlation of the ratio of temperature data to humidity data at different collection times within the detection interval, the temperature and humidity autocorrelation strength of the detection interval is calculated. The imbalance of temperature and humidity data in the monitoring interval is positively correlated with the negative correlation strength and autocorrelation strength of temperature and humidity in the monitoring interval, respectively.

[0006] Furthermore, the method for determining the temperature and humidity autocorrelation intensity within the detection range is as follows: The ratio of temperature data to humidity data at the time of acquisition is denoted as the temperature-humidity ratio at the time of acquisition. A temperature-humidity ratio sequence corresponding to the detection interval is established. The mean of the autocorrelation coefficients of the temperature-humidity ratio sequence corresponding to the detection interval at all lag orders is denoted as the temperature-humidity autocorrelation intensity of the detection interval.

[0007] Furthermore, the method for calculating the infrared harmonic anomaly of the detection interval is as follows: The interquartile range of all high-frequency components in the frequency domain of the detection interval is denoted as the high-frequency interquartile range of the detection interval. The sum of the differences between the energy values ​​of all harmonic frequencies and the energy values ​​of the fundamental frequency component in the frequency domain of the detection interval is recorded as the first sum of the detection interval. The infrared harmonic anomaly of the detection interval is calculated based on the high-frequency interquartile range and the first cumulative sum.

[0008] Furthermore, the environmental anomaly of the tower is positively correlated with the imbalance of temperature and humidity data and the anomaly of infrared harmonics in the monitoring interval.

[0009] Furthermore, the method for obtaining the peak periodic trend of the monitoring interval is as follows: Identify abrupt changes in temperature, humidity, and infrared light signal data within the detection range. Record the time interval between the abrupt change and the previous adjacent abrupt change in the same data as the abrupt change time interval and calculate the abrupt change interval information entropy of the monitoring range. Identify the changing trends of all abrupt changes in all temperature data, humidity data, and infrared light signal data within the monitoring interval, and calculate the second cumulative sum of the monitoring interval; The peak periodicity trend of the monitoring interval is calculated based on the information entropy of the mutation interval and the second cumulative sum.

[0010] Furthermore, the method for obtaining the information entropy of the mutation interval in the monitoring interval is as follows: The information entropy of the time interval between all mutation points identified in the monitoring interval is denoted as the mutation interval information entropy of the monitoring interval.

[0011] Furthermore, the method for obtaining the noise interference level in the monitoring interval is as follows: The sum of the normalized differences between the energy values ​​of all odd harmonic frequencies and the previous adjacent even harmonic frequency in the infrared light signal data of the detection interval is recorded as the infrared light sum of the detection interval. The temperature sum and humidity sum of the detection interval are also calculated. The spectral distortion of the detection interval is calculated based on the sum of infrared light, temperature, and humidity within the detection interval. The positive correlation between the peak periodicity trend and the spectral distortion of the monitoring interval is denoted as the noise interference degree of the monitoring interval.

[0012] Furthermore, the method for determining the spectral distortion of the detection interval is as follows: The sum of the infrared light accumulation, temperature accumulation, and humidity accumulation in the detection interval is denoted as the third accumulation of the detection interval. The positive correlation result of the third accumulation of the detection interval is denoted as the spectral distortion of the detection interval.

[0013] Furthermore, the specific steps for obtaining the monitoring results of communication tower data based on the tower environment anomaly and noise interference levels in all monitoring intervals are as follows: Based on the tower environment anomaly and noise interference of all monitoring intervals, the comprehensive score of the monitoring interval is obtained; the normalized value of the comprehensive score of the monitoring interval is recorded as the fault anomaly factor of the monitoring interval. The fault level of the communication tower in the detection range is assigned based on the fault anomaly factors in the monitoring range; the fault level is used as a label, and the fault level, temperature data, humidity data and infrared light signal data of the detection range are input into the anomaly classification model to obtain the anomaly classification results of the communication tower in the detection range.

[0014] The beneficial effects of this invention are: This application evaluates the correlation and temporal correlation of temperature and humidity data due to the decrease in autocorrelation strength caused by high-frequency fluctuations resulting from abnormal events, based on the characteristic that the autocorrelation strength of temperature and humidity data decreases due to high-frequency fluctuations caused by abnormal events. It obtains the temperature and humidity data imbalance degree within the monitoring interval, which reflects the disruption of the negative correlation balance between temperature and humidity data of the communication tower within the monitoring interval. Then, it evaluates the high-frequency energy dispersion and harmonic aberration degree of the infrared light sequence of the communication tower within the monitoring interval, obtaining the infrared harmonic anomaly of the detection interval. Furthermore, it calculates the tower environmental anomaly degree within the monitoring interval. A higher tower environmental anomaly degree indicates a greater likelihood of detecting abnormal numbers exhibiting fault characteristics within the monitoring interval. Furthermore, considering the issue of noise samples easily mixed into multi-sensor systems in complex natural environments, the peak periodicity trend of the monitoring interval is calculated based on the fluctuation degree of the abrupt change time interval of abnormal peaks in the multi-sensor data of the communication tower, as well as the long-term trend of the abrupt changes in the multi-sensor data. Combined with the degree of increase in energy values ​​corresponding to odd harmonic frequencies in the spectrum of the multi-sensor data collected by the tower within the detection interval, the noise interference degree of the monitoring interval is calculated. Finally, based on the tower environmental anomaly degree and noise interference degree of all monitoring intervals, the monitoring results of the communication tower data are obtained, resolving the decision boundary shift caused by the imbalance between abnormal and normal data, thus preventing misjudgment of anomalies in the multi-sensor data of the communication tower and improving the accuracy of communication tower data monitoring. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating a communication tower data acquisition and monitoring system based on multi-sensor integration, provided in one embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a communication tower data acquisition and monitoring system based on multi-sensor integration, provided in one embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 It shows a flowchart of a communication tower data acquisition and monitoring system based on multi-sensor integration according to an embodiment of the present invention. Figure 2 The diagram shows a schematic of a communication tower data acquisition and monitoring system based on multi-sensor integration provided by an embodiment of the present invention. The system includes: a data acquisition module, an environmental anomaly evaluation module, a noise interference evaluation module, and a monitoring result acquisition module.

[0019] The data acquisition module collects temperature data, humidity data, and infrared light signal data from the communication tower.

[0020] Temperature sensors, humidity sensors, and flame sensors are deployed on the surface of the communication tower base. Temperature, humidity, and infrared light signal data of the communication tower are collected using the temperature sensors, humidity sensors, and flame sensors, respectively.

[0021] Among them, temperature data, humidity data, and infrared light signal data are multi-sensor data; in this embodiment, the infrared light signal data is 760-1100nm infrared light signal data. In this embodiment, the acquisition frequency of temperature data, humidity data, and infrared light signal data is all set to 1Hz, and 4 hours is used as a detection interval; in this embodiment, the timestamps of temperature data, humidity data, and infrared light signal data are unified through a Beidou receiver or the IEEE 1588PTP protocol, and the analog signals of temperature data, humidity data, and infrared light signal data are converted into digital signals using an ADC0832 conversion chip.

[0022] This embodiment uses RS485 / Modbus protocol and 4G / 5G transmission to realize the acquisition and communication of data from temperature sensor, humidity sensor and flame sensor.

[0023] It should be noted that, for ease of calculation, all temperature, humidity, and infrared light signal data involved in the calculation in this embodiment have undergone data preprocessing to eliminate the influence of dimensions. This embodiment uses the Z-Score standard normalization method to perform dimensionless processing on the temperature, humidity, and infrared light signal data respectively. In practical applications, implementers can use other methods such as the maximum-minimum normalization method for dimensionless processing, which are not limited here.

[0024] The temperature data, humidity data, and infrared light signal data of all acquisition times within the same detection interval are arranged sequentially to obtain the temperature sequence, humidity sequence, and infrared light sequence of the detection interval.

[0025] At this point, the temperature sequence, humidity sequence, and infrared light sequence of the detection range have been obtained.

[0026] The environmental anomaly assessment module calculates the temperature and humidity data imbalance degree of the monitoring interval based on the time correlation and similarity differences of the ratio of temperature and humidity data in adjacent monitoring intervals. It also calculates the infrared harmonic anomaly of the monitoring interval based on all high-frequency components of the infrared light signal data in the frequency domain and the differences in energy values ​​of different frequency components. This reflects the high-frequency energy dispersion and harmonic amplification degree of the infrared light signal data in the frequency domain within the monitoring interval. Finally, it calculates the tower environmental anomaly degree of the monitoring interval based on the temperature and humidity data imbalance degree and the infrared harmonic anomaly.

[0027] In the process of data acquisition and monitoring of communication towers based on multi-sensor integration, data imbalance occurs due to the influence of physical characteristics and the occurrence patterns of fault and abnormal events. This means that the frequency of fault and abnormal events is much lower than under normal operating conditions, and the multi-sensor data of the communication tower structure is mostly in a steady state. If Support Vector Machine (SVM) is directly used to monitor communication tower data and identify abnormal data, the SVM model, under the principle of minimizing empirical risk (ERM), will tend to optimize the classification accuracy of the relatively high proportion of normal data in order to reduce overall error. This will cause the decision boundary to shift towards the region corresponding to abnormal data, resulting in a few abnormal fault data being ignored, and thus leading to serious fault detection problems.

[0028] The less significant the imbalance of abnormal fault data in the time dimension in multi-sensor data, the greater the possibility of abnormal fault data in the corresponding detection interval. The greater the significance of the disruption of the negative correlation steady-state balance maintained by temperature and humidity data under normal operating conditions, the lower the autocorrelation strength of temperature and humidity data will be due to high-frequency fluctuations caused by abnormal events.

[0029] The negative correlation results of the similarity between the temperature and humidity sequences in the detection interval are recorded as the negative correlation strength between temperature and humidity in adjacent monitoring intervals.

[0030] It is understood that negative correlation processing is applied to the similarity of the temperature and humidity sequences within the detection interval, ensuring that the similarity between the temperature and humidity sequences within the detection interval is negatively correlated with the negative correlation strength between temperature and humidity in adjacent monitoring intervals. It is understood that the negative correlation in this application refers to the relationship between the independent and dependent variables. The independent variable is the similarity between the temperature and humidity sequences within the detection interval, and the dependent variable is the negative correlation strength between temperature and humidity in adjacent monitoring intervals. The negative correlation means that the dependent variable decreases (increases) as the independent variable increases (decreases), and can be an inverse relationship, a subtraction relationship, etc.

[0031] Preferably, as an embodiment of this application, the absolute value of the similarity between the temperature sequence and the humidity sequence of the detection interval is recorded as the first similarity of the monitoring interval, and the difference between the number 1 and the first similarity of the monitoring interval is recorded as the negative correlation strength between temperature and humidity of adjacent monitoring intervals.

[0032] Some other embodiments of this application may involve using the similarity between the temperature and humidity sequences of the detection interval as the exponent of an exponential function with the natural constant e as the base, and taking the reciprocal of the calculation result of the exponential function as the negative correlation strength between temperature and humidity in adjacent monitoring intervals.

[0033] This embodiment uses cosine similarity to measure the similarity between temperature and humidity sequences.

[0034] The ratio of temperature data to humidity data at the time of acquisition is recorded as the temperature-humidity ratio at the time of acquisition. The temperature-humidity ratios at all acquisition times within the same detection interval are arranged sequentially to obtain the temperature-humidity ratio sequence. The autocorrelation function ACF is used to obtain the autocorrelation coefficient of the temperature-humidity ratio sequence at all lag orders. The mean of the autocorrelation coefficients of all lag orders corresponding to the detection interval is recorded as the temperature-humidity autocorrelation intensity of the detection interval.

[0035] The smaller the autocorrelation strength of temperature and humidity in the detection range, the more obvious the impact of faults and abnormal events on the communication tower within the monitoring range will be, causing a more significant disruption to the steady-state balance of the temperature and humidity ratio at different collection times within the monitoring range, and the more obvious the high-frequency fluctuations in the temperature and humidity data of the communication tower will be.

[0036] The positive correlation result of the difference between the temperature and humidity autocorrelation strength of the monitoring interval and the negative correlation strength of temperature and humidity in adjacent monitoring intervals is recorded as the temperature and humidity data imbalance degree of the monitoring interval.

[0037] It is understood that the difference in the negative correlation strength of temperature and humidity between adjacent monitoring intervals and the autocorrelation strength of temperature and humidity within the monitoring intervals are positively correlated. This ensures that both the difference in the negative correlation strength of temperature and humidity between adjacent monitoring intervals and the autocorrelation strength of temperature and humidity within the monitoring intervals are positively correlated with the temperature and humidity data imbalance of the monitoring intervals. It is understood that the positive correlation in this application refers to the relationship between the independent and dependent variables. The independent variables are the difference in the negative correlation strength of temperature and humidity between adjacent monitoring intervals and the autocorrelation strength of temperature and humidity within the monitoring intervals. The dependent variable is the temperature and humidity data imbalance of the monitoring intervals. The positive correlation means that the dependent variable increases (decreases) as the independent variable increases (decreases), and can be an additive or multiplicative relationship.

[0038] Preferably, as an embodiment of this application, the difference in the negative correlation strength between temperature and humidity between the latter and former monitoring intervals in adjacent monitoring intervals is recorded as the first difference between the former and latter monitoring intervals. The first difference between the monitoring intervals is used as the exponent of an exponential function with the natural constant e as the base, and the calculation result of the exponential function is recorded as the first exponential value of the monitoring interval. The product of the negative correlation strength between temperature and humidity in the monitoring interval and the sum of the preset adjustment parameters and the first exponential value of the monitoring interval is recorded as the temperature and humidity data imbalance degree of the monitoring interval.

[0039] The purpose of adjusting the parameter is to prevent the negative correlation strength between temperature and humidity in the monitoring interval from being 0, which would result in the temperature and humidity data imbalance in the monitoring interval being 0. In this embodiment, the value of the adjusting parameter is 1.5.

[0040] The imbalance of temperature and humidity data in the monitoring interval is used to reflect the disruption of the negative correlation balance between temperature and humidity data of communication towers within the monitoring interval. The more significant the imbalance between temperature and humidity data, the more slight the data imbalance of communication towers within the monitoring interval, and the more likely there are abnormal temperature and humidity data within the monitoring interval.

[0041] When an abnormal heat source occurs within the time period corresponding to the monitoring interval, the more significant the high-frequency energy dispersion of the infrared light sequence in the monitoring interval, the more pronounced the harmonic component enhancement in the infrared light sequence. Therefore, the high-frequency energy dispersion and harmonic enhancement degree of the infrared light sequence of the communication tower within the monitoring interval are evaluated.

[0042] The infrared light sequence of the detection interval was transformed to the frequency domain using FFT Fourier transform. All frequency components of the infrared light sequence of the monitoring interval were extracted using Hilbert transform. The frequency corresponding to the highest energy value in the frequency domain was defined as the fundamental frequency, and the frequency components corresponding to the fundamental frequency were taken as the fundamental frequency components. All integer multiples of the fundamental frequency were taken as harmonic frequencies of the infrared light sequence, and the harmonic frequency with an energy value of 0 was taken as the maximum harmonic frequency of the fundamental frequency. The frequencies corresponding to the highest energy values ​​among all frequency components of the infrared light sequences of all monitoring intervals were extracted and formed into a dataset. The Otsu's inter-class variance method was used to calculate the segmentation threshold of the dataset, and all components corresponding to frequencies greater than or equal to the segmentation threshold were classified as high-frequency components.

[0043] Among them, the FFT Fourier transform, Hilbert transform, and maximal inter-class variance method are all well-known techniques and will not be elaborated further.

[0044] The interquartile range of all high-frequency components in the frequency domain of the detection interval is denoted as the high-frequency interquartile range of the detection interval. The sum of the differences between the energy values ​​of all harmonic frequencies and the energy values ​​of the fundamental frequency components in the frequency domain of the detection interval is denoted as the first sum of the detection interval. The normalized value of the ratio of the high-frequency interquartile range of the detection interval to the first sum is denoted as the infrared harmonic anomaly of the detection interval.

[0045] In the process of calculating the ratio, in order to avoid the denominator being zero, a preset value needs to be added to the denominator. In this example, the preset value is 0.01.

[0046] In this embodiment, the sigmoid function is used to calculate the normalized value. The sigmoid function is a well-known technique and will not be described in detail here. As other implementation methods, implementers can use other methods of the prior art, such as the tanh function.

[0047] Infrared harmonic anomalies are used to reflect the high-frequency energy dispersion and harmonic amplification of infrared light signal data in the frequency domain within the monitoring interval. When the likelihood of heat source anomalies in the environment where the communication tower is located is higher, the energy distribution range of the high-frequency components of the infrared light signal in the frequency domain is wider, i.e., the high-frequency interquartile range of the detection interval is larger. Simultaneously, the difference between the harmonic frequency energy and the fundamental frequency energy of the infrared light signal in the frequency domain is smaller, i.e., the first cumulative sum of the detection interval is smaller. In this case, the infrared harmonic anomaly of the detection interval is greater.

[0048] When abnormal data indicating faults are more likely to be detected within the monitoring range, the imbalance between temperature and humidity data in the communication tower environment becomes more significant. At the same time, the distribution range of high-frequency energy in the infrared light signal is wider, and the increase of harmonic frequency components is more obvious.

[0049] Therefore, the positive correlation between the temperature and humidity data imbalance and the infrared harmonic anomaly in the monitoring interval is recorded as the tower environment anomaly degree in the monitoring interval.

[0050] Preferably, as an embodiment of this application, the product of the temperature and humidity data imbalance degree of the monitoring interval and the infrared light harmonic anomaly degree is recorded as the tower environment anomaly degree of the monitoring interval.

[0051] The greater the anomaly of the tower environment in the monitoring range, the more likely abnormal data indicating a fault will be detected within the monitoring range.

[0052] At this point, the anomaly level of the tower environment within the monitoring range is obtained.

[0053] The noise interference evaluation module is used to calculate the peak periodicity trend of the monitoring interval based on the time interval between abrupt changes in the temperature data, humidity data, and infrared light signal data of the detection interval, as well as the changing trend of the abrupt changes. It also calculates the noise interference degree of the monitoring interval by combining the energy differences of adjacent harmonic frequencies in the frequency domain of the temperature data, humidity data, and infrared light signal data of the detection interval.

[0054] In the data monitoring process of communication towers based on multi-sensor integration, relying solely on the imbalance between temperature and humidity data and the anomalies of infrared harmonics as criteria for classification decision models has significant drawbacks. Specifically, this approach ignores the problem of noise samples easily mixed into multi-sensor systems in complex natural environments and lacks targeted analysis of how noise samples cause anomalies in multi-sensor data. This can lead to the misclassification of mixed noise data as anomaly data corresponding to the fault when there is an imbalance between abnormal and normal data, resulting in frequent false alarms. This not only increases the difficulty of communication tower operation and maintenance but also poses a threat to the safety of tower operation.

[0055] Specifically, when the anomalies in the multi-sensor data collected within the monitoring interval are more likely to be caused by noise samples rather than actual faults, the following characteristics will be observed: the periodic intensity of the abnormal peak values ​​in the multi-sensor data is low, and the long-term trend is weak; due to noise interference, the spectral distortion of the multi-sensor data is more significant, specifically manifested as an abnormal increase in the energy of odd harmonics in the frequency domain.

[0056] The BG sequence segmentation algorithm is used to obtain abrupt changes in the temperature, humidity, and infrared light sequences within the detection interval. The time interval between an abrupt change and its preceding adjacent abrupt change is denoted as the abrupt change interval. The information entropy of the abrupt change intervals of all abrupt changes identified within the monitoring interval is denoted as the abrupt change interval information entropy of the monitoring interval. The sum of the Hurst exponents of the sequences composed of all abrupt changes identified within the temperature, humidity, and infrared light sequences within the monitoring interval is denoted as the second cumulative sum of the monitoring interval. The ratio of the abrupt change interval information entropy to the second cumulative sum is denoted as the peak periodicity trend of the monitoring interval.

[0057] In particular, the first mutation point has no preceding adjacent mutation point, therefore, the mutation time interval is not calculated for the first mutation point; in the process of calculating the ratio, in order to avoid the denominator being zero, a preset value needs to be added to the denominator, and the preset value in this embodiment is 0.01.

[0058] The methods used in the BG sequence segmentation algorithm to obtain mutation points, calculate information entropy, and calculate the Hurst exponent of the sequence are all well-known techniques and will not be elaborated further.

[0059] When the peak periodicity of the monitoring interval is greater, the time interval of abrupt changes in abnormal peaks in the multi-sensor data of the communication tower is more volatile, the long-term trend of abrupt changes in the multi-sensor data is weaker, and the anomalies in the multi-sensor data within the monitoring interval are more likely to be caused by noise sample mixing.

[0060] The sum of the normalized differences between the energy values ​​of all odd harmonic frequencies and the previous adjacent even harmonic frequency in the infrared light sequence of the detection interval is denoted as the infrared light sum of the detection interval.

[0061] In this embodiment, the sigmoid function is used to calculate the normalized value. The sigmoid function is a well-known technique and will not be described in detail here. As other implementation methods, implementers can use other methods of the prior art, such as the tanh function.

[0062] The same method can be used to process the temperature and humidity sequences within the detection range to obtain the cumulative sum of temperature and humidity within the detection range.

[0063] The sum of the infrared light accumulation, temperature accumulation, and humidity accumulation in the detection interval is denoted as the third accumulation of the detection interval. The positive correlation result of the third accumulation of the detection interval is denoted as the spectral distortion of the detection interval.

[0064] Preferably, as an embodiment of this application, the third cumulative sum of the detection interval is taken as the exponent value of an exponential function with the natural constant e as the base, and the calculation result of the exponential function is recorded as the spectral distortion of the detection interval.

[0065] The greater the spectral distortion in the detection range, the more pronounced the increase in energy values ​​corresponding to odd harmonic frequencies in the spectrum of the multi-sensor data collected by the communication tower within the detection range. In this case, the anomalies in the multi-sensor data within the monitoring range are more likely to be caused by noise sample mixing.

[0066] The positive correlation between the peak periodicity trend and the spectral distortion of the monitoring interval is denoted as the noise interference degree of the monitoring interval.

[0067] Preferably, as an embodiment of this application, the normalized value of the product of the peak periodicity trend and the spectral distortion of the monitoring interval is denoted as the noise interference degree of the monitoring interval.

[0068] In this embodiment, the sigmoid function is used to calculate the normalized value. The sigmoid function is a well-known technique and will not be described in detail here. As other implementation methods, implementers can use other methods of the prior art, such as the tanh function.

[0069] The greater the noise interference in the monitoring interval, the more likely the anomalies in the multi-sensor data within the monitoring interval are caused by noise samples being mixed in.

[0070] Understandably, the less likely the anomalies in the multi-sensor data within the monitoring range are caused by noise samples, and the more likely abnormal data indicating a fault are to be detected within the monitoring range, the greater the likelihood of a fault event occurring in the communication tower within the monitoring range, and the more severe the fault event occurring in the communication tower within the monitoring range.

[0071] At this point, the noise interference level of the monitoring range is obtained.

[0072] The monitoring result acquisition module is used to acquire monitoring results of communication tower data based on the tower environment anomaly and noise interference levels in all monitoring intervals.

[0073] The CIM comprehensive index method was used to process the tower environmental anomaly and noise interference levels in all monitoring intervals, obtaining a comprehensive score for each monitoring interval. The comprehensive score of each monitoring interval is a positively correlated result of the tower environmental anomaly and a negatively correlated result of the noise interference. The weights of the tower environmental anomaly and noise interference levels were determined using the entropy weight method. The normalized value of the comprehensive score for each monitoring interval was denoted as the fault anomaly factor for that interval.

[0074] The CIM comprehensive index method for obtaining the comprehensive score and the entropy weight method for determining the weights are both well-known techniques and will not be elaborated further. This embodiment uses the maximum-minimum value normalization method to calculate the normalized value. In practical applications, implementers may use other existing methods, such as the sigmoid function, to calculate the normalized value, and this is not limited here.

[0075] The system presets a first, second, third, and fourth classification thresholds, and classifies the fault level of the communication tower within the monitoring interval based on the fault anomaly factors in the monitoring interval.

[0076] Specifically, when the fault anomaly factor in the monitoring interval is less than or equal to the first classification threshold, the fault level of the communication tower in the detection interval is recorded as 1; when the fault anomaly factor in the monitoring interval is greater than the first classification threshold and less than or equal to the second classification threshold, the fault level of the communication tower in the detection interval is recorded as 2; when the fault anomaly factor in the monitoring interval is greater than the second classification threshold and less than or equal to the third classification threshold, the fault level of the communication tower in the detection interval is recorded as 3; when the fault anomaly factor in the monitoring interval is greater than the third classification threshold and less than or equal to the fourth classification threshold, the fault level of the communication tower in the detection interval is recorded as 4; and when the fault anomaly factor in the monitoring interval is greater than the fourth classification threshold, the fault level of the communication tower in the detection interval is recorded as 5.

[0077] The fault level of the communication tower in the detection range is used as a label for the temperature sequence, humidity sequence, and infrared light sequence of the communication tower in the detection range. A model training set is constructed based on the temperature sequence, humidity sequence, and infrared light sequence of the communication tower in the detection range. An SVM support vector machine with Gaussian kernel function is used to train the model training set to obtain an anomaly classification model of multi-sensor data of the communication tower.

[0078] The training process for SVM (Support Vector Machine) is a well-known technique and will not be described in detail here.

[0079] The fault level of the communication tower in the detection range is used as a label for the temperature sequence, humidity sequence, and infrared light sequence of the communication tower in the detection range. The fault level, temperature sequence, humidity sequence, and infrared light sequence of the detection range are input into the anomaly classification model to obtain the anomaly classification results of the communication tower in the detection range.

[0080] The abnormal classification results of communication towers within the detection range are stored using a MySQL database for easy subsequent querying.

[0081] At this point, the monitoring results of the communication tower data have been obtained.

[0082] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A communication tower data acquisition and monitoring system based on multi-sensor integration, characterized in that, The system includes the following modules: The data acquisition module is used to collect temperature data, humidity data, and infrared light signal data from the communication tower. The environmental anomaly assessment module is used to calculate the temperature and humidity data imbalance of the monitoring interval based on the time correlation and similarity differences of the ratio of temperature and humidity data in adjacent monitoring intervals. Based on all high-frequency components of the infrared light signal data in the frequency domain within the detection interval, and the differences in energy values ​​of different frequency components, it calculates the infrared light harmonic anomaly of the detection interval. This reflects the high-frequency energy dispersion and harmonic amplification of the infrared light signal data in the frequency domain within the monitoring interval. Based on the temperature and humidity data imbalance and the infrared light harmonic anomaly, it calculates the tower environmental anomaly degree of the monitoring interval. The noise interference evaluation module is used to calculate the peak periodicity trend of the monitoring interval based on the time interval between abrupt changes in the temperature data, humidity data, and infrared light signal data of the detection interval, as well as the changing trend of the abrupt changes. It also calculates the noise interference degree of the monitoring interval by combining the energy differences of adjacent harmonic frequencies in the frequency domain of the temperature data, humidity data, and infrared light signal data of the detection interval. The monitoring result acquisition module is used to acquire monitoring results of communication tower data based on the tower environment anomaly and noise interference levels in all monitoring intervals.

2. The communication tower data acquisition and monitoring system based on multi-sensor integration according to claim 1, characterized in that, The method for obtaining the temperature and humidity data imbalance in the monitoring interval is as follows: The negative correlation result of the similarity between temperature data and humidity data in the detection interval is recorded as the negative correlation strength between temperature and humidity in adjacent monitoring intervals. Based on the temporal correlation of the ratio of temperature data to humidity data at different collection times within the detection interval, the temperature and humidity autocorrelation strength of the detection interval is calculated. The imbalance of temperature and humidity data in the monitoring interval is positively correlated with the negative correlation strength and autocorrelation strength of temperature and humidity in the monitoring interval, respectively.

3. The communication tower data acquisition and monitoring system based on multi-sensor integration according to claim 2, characterized in that, The method for determining the temperature and humidity autocorrelation intensity within the detection range is as follows: The ratio of temperature data to humidity data at the time of acquisition is denoted as the temperature-humidity ratio at the time of acquisition. A temperature-humidity ratio sequence corresponding to the detection interval is established. The mean of the autocorrelation coefficients of the temperature-humidity ratio sequence corresponding to the detection interval at all lag orders is denoted as the temperature-humidity autocorrelation intensity of the detection interval.

4. The communication tower data acquisition and monitoring system based on multi-sensor integration according to claim 1, characterized in that, The method for calculating the infrared harmonic anomaly of the detection interval is as follows: The interquartile range of all high-frequency components in the frequency domain of the detection interval is denoted as the high-frequency interquartile range of the detection interval. The sum of the differences between the energy values ​​of all harmonic frequencies and the energy values ​​of the fundamental frequency component in the frequency domain of the detection interval is recorded as the first sum of the detection interval. The infrared harmonic anomaly of the detection interval is calculated based on the high-frequency interquartile range and the first cumulative sum.

5. The communication tower data acquisition and monitoring system based on multi-sensor integration according to claim 1, characterized in that, The environmental anomaly of the tower is positively correlated with the imbalance of temperature and humidity data and the anomaly of infrared harmonics in the monitoring interval.

6. The communication tower data acquisition and monitoring system based on multi-sensor integration according to claim 1, characterized in that, The method for obtaining the peak periodic trend of the monitoring interval is as follows: Identify abrupt changes in temperature, humidity, and infrared light signal data within the detection range. Record the time interval between the abrupt change and the previous adjacent abrupt change in the same data as the abrupt change time interval and calculate the abrupt change interval information entropy of the monitoring range. Identify the changing trends of all abrupt changes in all temperature data, humidity data, and infrared light signal data within the monitoring interval, and calculate the second cumulative sum of the monitoring interval; The peak periodicity trend of the monitoring interval is calculated based on the information entropy of the mutation interval and the second cumulative sum.

7. The communication tower data acquisition and monitoring system based on multi-sensor integration according to claim 6, characterized in that, The method for obtaining the entropy of the mutation interval information in the monitoring interval is as follows: The information entropy of the time interval between all mutation points identified in the monitoring interval is denoted as the mutation interval information entropy of the monitoring interval.

8. The communication tower data acquisition and monitoring system based on multi-sensor integration according to claim 1, characterized in that, The method for obtaining the noise interference level of the monitoring interval is as follows: The sum of the normalized differences between the energy values ​​of all odd harmonic frequencies and the previous adjacent even harmonic frequency in the infrared light signal data of the detection interval is recorded as the infrared light sum of the detection interval. The temperature sum and humidity sum of the detection interval are also calculated. The spectral distortion of the detection interval is calculated based on the sum of infrared light, temperature, and humidity within the detection interval. The positive correlation between the peak periodicity trend and the spectral distortion of the monitoring interval is denoted as the noise interference degree of the monitoring interval.

9. The communication tower data acquisition and monitoring system based on multi-sensor integration according to claim 8, characterized in that, The method for determining the spectral distortion of the detection interval is as follows: The sum of the infrared light accumulation, temperature accumulation, and humidity accumulation in the detection interval is denoted as the third accumulation of the detection interval. The positive correlation result of the third accumulation of the detection interval is denoted as the spectral distortion of the detection interval.

10. The communication tower data acquisition and monitoring system based on multi-sensor integration according to claim 1, characterized in that, The specific steps for obtaining the monitoring results of communication tower data based on the tower environment anomaly and noise interference levels in all monitoring intervals are as follows: Based on the tower environment anomaly and noise interference of all monitoring intervals, the comprehensive score of the monitoring interval is obtained; the normalized value of the comprehensive score of the monitoring interval is recorded as the fault anomaly factor of the monitoring interval. The fault level of the communication tower in the detection range is assigned based on the fault anomaly factors in the monitoring range; the fault level is used as a label, and the fault level, temperature data, humidity data and infrared light signal data of the detection range are input into the anomaly classification model to obtain the anomaly classification results of the communication tower in the detection range.

Images