Disastrous rainfall monitoring method based on Internet of Things signal transmission effect

Abstract

The invention relates to a disastrous rainfall monitoring method based on an Internet of Things signal transmission effect, and the method comprises the steps: constructing a signal feature matrix through employing the signal parameters of an Internet of Things terminal obtained by a communication base station; dividing the reference rainfall data acquired by the auxiliary rainfall monitoring equipment into a rainfall grade matrix; constructing a training set and a test set based on the signal feature matrix and the rainfall level matrix; training a machine learning model by using the training set to obtain a rainfall intensity grade estimation model; inputting the signal features in the test set into the rainfall intensity grade estimation model to obtain a rainfall grade estimation result; and carrying out spatial interpolation processing on the rainfall grade estimation result to obtain regional rainfall intensity grade distribution data. According to the invention, the attenuation effect of the communication link signal of the honeycomb Internet of Things caused by rainfall is fully utilized, large-range and high-temporal-spatial-resolution rainfall monitoring can be realized, and the method can be applied to early warning scenes of disasters such as urban inland inundation and debris flow in mountainous areas.

Description

Technical Field

[0001] This invention relates to the field of precipitation detection technology, and in particular to a method for monitoring disastrous rainfall based on the signal transmission effect of the Internet of Things. Background Technology

[0002] Traditional rainfall monitoring primarily relies on technologies such as ground-based rain gauges, weather radar, and satellite remote sensing. While ground-based rain gauges provide accurate measurements, their sparse spatial distribution limits their ability to detect small-scale heavy rainfall. Weather radar can provide information on rainfall distribution over a wider area, but factors such as beam obstruction and bright band effects lead to significant estimation errors in near-surface precipitation, and the equipment is also expensive. Although satellite remote sensing technology offers broad coverage, its temporal resolution is low; geostationary satellites update data approximately every 30 minutes, while polar-orbiting satellites require revisiting every few hours, and its ability to detect weak precipitation is limited.

[0003] In recent years, rainfall monitoring methods based on wireless signal attenuation have gradually attracted attention. Among them, commercial microwave link technology uses rain-induced attenuation inversion path to average rainfall intensity through a fixed line-of-sight microwave link. However, this method requires dedicated communication equipment and has poor deployment flexibility. Another attempt is to estimate rainfall using low-frequency non-line-of-sight signals between 4G / 5G mobile terminals and base stations. However, due to the non-fixed location of mobile terminals and random antenna orientation, the accuracy of rainfall measurement still needs to be improved.

[0004] With the widespread adoption of cellular IoT technology, the massive deployment of smart meters, environmental monitoring terminals, and other IoT devices in fixed locations provides numerous low-frequency non-line-of-sight links with fixed signal transmission paths, offering new possibilities for high-precision, high-spatiotemporal resolution rainfall monitoring. However, existing technologies still face several key challenges: First, signal attenuation between IoT terminals and base stations is affected not only by rainfall but also by environmental factors such as building obstruction and vegetation growth; effectively separating precipitation and non-precipitation attenuation becomes a technical hurdle. Second, individual terminals can only provide point-like rainfall information, lacking effective algorithms to fuse discrete observation data into continuous regional rainfall distribution. Furthermore, existing rainfall attenuation models are primarily designed for high-frequency bands (above 10 GHz), while commonly used IoT frequency bands (such as 2.4 GHz) have lower sensitivity to rainfall, necessitating the development of new inversion models. Therefore, how to utilize existing cellular IoT infrastructure to achieve high spatiotemporal resolution and high accuracy in regional rainfall intensity estimation without adding dedicated monitoring equipment, especially improving the detection sensitivity of low-frequency signals for weak rainfall, solving multipath interference in complex urban environments, and ensuring the spatial reconstruction accuracy of rainfall fields in sparse terminal areas, has become a key issue that urgently needs to be addressed in the current technological field. Summary of the Invention

[0005] The purpose of this invention is to propose a method for monitoring disastrous rainfall based on the signal transmission effect of the Internet of Things (IoT) to address the problems existing in the prior art. This method involves real-time acquisition of communication signal parameters between IoT terminals and base stations within a target area to extract rainfall characteristics; establishing a mapping relationship between signal attenuation characteristics and rainfall intensity levels; and building a rainfall intensity level estimation model using machine learning algorithms. The real-time signal characteristics are then input into the model to estimate the current rainfall intensity level, and combined with multi-terminal spatial distribution data to generate a regional rainfall level distribution map. This method fully utilizes the signal attenuation effect of rainfall on cellular IoT communication links, enabling large-scale, high spatiotemporal resolution rainfall monitoring, and can be applied to disaster early warning scenarios such as urban flooding and mountain mudslides.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] Disastrous rainfall monitoring methods based on IoT signal transmission effects include:

[0008] A signal feature matrix is ​​constructed using the signal parameters of IoT terminals obtained from communication base stations;

[0009] A rainfall level matrix is ​​formed based on the baseline rainfall data acquired by auxiliary rainfall monitoring equipment;

[0010] Based on the signal feature matrix and rainfall level matrix, a training set and a test set are constructed;

[0011] The training set is used to train the machine learning model to obtain a rainfall intensity level estimation model;

[0012] The signal features in the test set are input into the rainfall intensity level estimation model to obtain the rainfall level estimation results;

[0013] Spatial interpolation is performed on the rainfall level estimation results to obtain regional rainfall intensity level distribution data.

[0014] Optionally, constructing the signal feature matrix includes:

[0015] Based on base station measurement reports, obtain signal parameters of several IoT terminals near the communication base station during a preset time period;

[0016] For the signal parameters of each IoT terminal, calculate the statistical characteristics and construct an initial signal feature matrix;

[0017] The initial signal feature matrix is ​​processed using a normalization method to obtain the final signal feature matrix.

[0018] Optionally, the rainfall level matrix includes:

[0019] The rainfall data is spatially nearest neighbor matched with IoT terminals to obtain a rainfall data matrix;

[0020] The rainfall data matrix is ​​divided according to a preset rainfall level standard to obtain the rainfall level matrix.

[0021] Optionally, the initial signal feature matrix is:

[0022]

[0023] Where C represents the initial signal feature matrix, Std represents the statistical features, m represents the number of IoT terminals, and t1, t2, t3, t4, t5, t6, t7, t8, t9, t1 ... n Indicates the start and end times of a preset time period.

[0024] Optionally, the rainfall level matrix is:

[0025]

[0026] Where L represents the rainfall level matrix, L m This represents the rainfall intensity level at the m-th terminal.

[0027] Optionally, spatial interpolation processing of the rainfall level estimation results includes:

[0028] Based on the spatial distribution characteristics of multiple terminals, the inverse distance weighted interpolation method is used to perform spatial interpolation calculations on the rainfall level estimation results of discrete IoT nodes, generating continuous regional rainfall intensity level distribution data.

[0029] Optionally, the signal parameters include a signal strength indication.

[0030] Optionally, the statistical characteristic includes: sliding standard deviation.

[0031] The beneficial effects of this invention are as follows:

[0032] Compared with existing technologies, the disaster rainfall monitoring method based on IoT signal transmission effect proposed in this invention fully explores rainfall-related characteristics based on the statistical results of IoT-communication base station measurement reports, and realizes real-time and effective acquisition of rainfall information in the region. In addition, the method is based on the widely existing cellular IoT communication links, eliminating the need to deploy dedicated meteorological equipment. It utilizes existing IoT infrastructure to achieve high spatiotemporal resolution rainfall monitoring, significantly reducing deployment and maintenance costs. It is particularly suitable for scenarios such as urban flooding and mountain debris flow early warning. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments 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.

[0034] Figure 1 This is a schematic diagram showing the distribution of communication base stations, IoT terminals, and auxiliary rainfall monitoring equipment in the disaster rainfall monitoring method based on IoT signal transmission effect according to an embodiment of the present invention.

[0035] Figure 2 This is a flowchart illustrating the process of a disaster rainfall monitoring method based on the Internet of Things signal transmission effect, according to an embodiment of the present invention.

[0036] Figure 3 This is a schematic diagram of the BP neural network model structure established in an embodiment of the present invention;

[0037] Figure 4 This is the confusion matrix between the rainfall intensity level estimated based on the cellular IoT communication link and the raindrop spectrometer measurement value in an embodiment of the present invention. Detailed Implementation

[0038] 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.

[0039] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0040] This embodiment proposes a method for monitoring disastrous rainfall based on the signal transmission effect of the Internet of Things, including:

[0041] A signal feature matrix is ​​constructed using the signal parameters of IoT terminals obtained from communication base stations;

[0042] A rainfall level matrix is ​​formed based on the baseline rainfall data acquired by auxiliary rainfall monitoring equipment;

[0043] Training and testing sets are constructed based on the signal feature matrix and the rainfall level matrix;

[0044] The machine learning model is trained using the training set to obtain a rainfall intensity level estimation model;

[0045] The signal features in the test set are input into the rainfall intensity level estimation model to obtain the rainfall level estimation results;

[0046] Spatial interpolation is performed on the rainfall level estimation results to obtain regional rainfall intensity level distribution data.

[0047] Furthermore, constructing the signal feature matrix includes:

[0048] Based on base station measurement reports, obtain signal parameters of several IoT terminals near the communication base station during a preset time period;

[0049] For the signal parameters of each IoT terminal, calculate the statistical characteristics and construct an initial signal feature matrix;

[0050] The initial signal feature matrix is ​​processed using a normalization method to obtain the final signal feature matrix.

[0051] Furthermore, the rainfall level matrix includes:

[0052] The rainfall data is spatially nearest neighbor matched with IoT terminals to obtain a rainfall data matrix;

[0053] The rainfall data matrix is ​​divided according to the preset rainfall level standard to obtain the rainfall level matrix.

[0054] Furthermore, the initial signal feature matrix is:

[0055]

[0056] Where C represents the initial signal feature matrix, Std represents the statistical features, m represents the number of statistical features, and t1, t2, t3, t4, t5, t6, t7, t8, t9, t1 ... n Indicates the start and end times of a preset time period.

[0057] Furthermore, the rainfall level matrix is ​​as follows:

[0058]

[0059] Where L represents the rainfall level matrix.

[0060] Furthermore, spatial interpolation processing of the rainfall level estimation results includes:

[0061] Based on the spatial distribution characteristics of multiple terminals, the inverse distance weighted interpolation method is used to perform spatial interpolation calculations on the rainfall level estimation results of discrete IoT nodes, generating continuous regional rainfall intensity level distribution data.

[0062] Furthermore, signal parameters include, but are not limited to, signal strength indication.

[0063] Furthermore, statistical characteristics include, but are not limited to, the moving standard deviation.

[0064] Furthermore, the normalization methods for the feature matrix include, but are not limited to, Z-score normalization.

[0065] Furthermore, machine learning models include, but are not limited to, backpropagation (BP) neural networks.

[0066] Furthermore, spatial interpolation techniques include, but are not limited to, inverse distance-weighted interpolation.

[0067] Figure 1 This is a schematic diagram showing the distribution of communication base stations, IoT terminals, and auxiliary rainfall monitoring equipment in the disaster rainfall monitoring method based on the Internet of Things signal transmission effect described in this invention; Figure 2 This is a flowchart illustrating the process of the disaster rainfall monitoring method based on the Internet of Things signal transmission effect described in this invention. Figure 3 This is a schematic diagram of the BP neural network model structure established in the example described in this invention; Figure 4 It is the confusion matrix between the rainfall intensity level estimated based on the cellular IoT communication link and the raindrop spectrometer measurement value in the example described in this invention.

[0068] This embodiment obtains signal parameters from IoT terminals through communication base stations, extracts rainfall-related features, establishes a rainfall intensity level estimation model using machine learning algorithms, and calculates a regional rainfall level distribution map by combining terminal spatial distribution information. The main steps include:

[0069] 1. Obtain signal parameters from IoT terminals via communication base stations and extract rainfall-related features.

[0070] (1) Collect base station measurement report data to obtain the data of 25 nearby IoT terminals from t1 to t2. 1440 Received Signal Strength Indication (RSSI) at any given time:

[0071]

[0072] (2) For each terminal's signal parameter sequence, calculate the sliding standard deviation (Std) feature with a window length of 60 and construct the feature matrix C:

[0073]

[0074] The feature matrix C is then processed using the Z-score normalization method to obtain C′.

[0075] (2) Obtain baseline rainfall data through auxiliary rainfall monitoring equipment near the communication base station, and perform spatial nearest neighbor matching between the rainfall data and the Internet of Things terminal to obtain the rainfall data matrix R;

[0076]

[0077] The rainfall data matrix is ​​then divided according to a preset rainfall level standard to obtain the rainfall level matrix L;

[0078]

[0079] (3) The standardized feature matrix C′ and the rainfall level matrix L are divided into training and testing sets in an 8:2 ratio. The training set is used to train a BP neural network algorithm to learn the mapping relationship between signal features and rainfall intensity levels, thus establishing a rainfall intensity level estimation model. The model consists of an input layer, a hidden layer, and an output layer, with the structure as follows: Figure 3 As shown, the input layer receives the normalized signal features of the input, the hidden layer learns the high-dimensional hidden features of the input data, and the output layer establishes the nonlinear relationship between the high-dimensional hidden features and the rainfall level.

[0080] (4) Input the normalized signal features in the test set into the rainfall inversion model M, and output the rainfall intensity level estimate L′ for each IoT terminal location;

[0081]

[0082] (5) Based on the spatial distribution characteristics of multiple terminals, the inverse distance weighted interpolation method is used to perform spatial interpolation calculation on the estimated rainfall level of discrete IoT nodes, and the location to be estimated (lon) unknown ,lat unknown The rainfall level L″ can be expressed as:

[0083]

[0084] Where, λ k The weighting coefficients for the location to be estimated are used to estimate the rainfall level at the k-th IoT terminal. k Let be the Euclidean distance between the k-th IoT terminal location and the location to be estimated, r (10km) be the radius of influence, and ρ (2) be the power parameter. Therefore, a continuous regional rainfall intensity level distribution map can be obtained.

[0085] In practical applications, based on the signal characteristics of IoT terminals extracted from base station measurement reports, the rainfall intensity level of a single terminal location and the spatial distribution of rainfall levels in the entire area can be obtained in real time.

[0086] Compared with existing technologies, the disaster rainfall monitoring method based on IoT signal transmission effect proposed in this embodiment fully explores rainfall-related characteristics based on the statistical results of IoT-communication base station measurement reports, and realizes real-time and effective acquisition of rainfall information in the region. In addition, the method is based on the widely existing cellular IoT communication links, eliminating the need to deploy dedicated meteorological equipment. It utilizes existing IoT infrastructure to achieve high spatiotemporal resolution rainfall monitoring, significantly reducing deployment and maintenance costs. It is particularly suitable for scenarios such as urban flooding and mountain debris flow early warning.

[0087] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for monitoring disastrous rainfall based on the signal transmission effect of the Internet of Things, characterized in that, include: A signal feature matrix is ​​constructed using the signal parameters of IoT terminals obtained from communication base stations; A rainfall level matrix is ​​formed based on the baseline rainfall data acquired by auxiliary rainfall monitoring equipment; Based on the signal feature matrix and rainfall level matrix, a training set and a test set are constructed; The training set is used to train the machine learning model to obtain a rainfall intensity level estimation model; The signal features in the test set are input into the rainfall intensity level estimation model to obtain the rainfall level estimation results; Spatial interpolation is performed on the rainfall level estimation results to obtain regional rainfall intensity level distribution data.

2. The method for monitoring disastrous rainfall based on the signal transmission effect of the Internet of Things according to claim 1, characterized in that, Constructing the signal feature matrix includes: Based on base station measurement reports, obtain signal parameters of several IoT terminals near the communication base station during a preset time period; For the signal parameters of each IoT terminal, calculate the statistical characteristics and construct an initial signal feature matrix; The initial signal feature matrix is ​​processed using a normalization method to obtain the final signal feature matrix.

3. The method for monitoring disastrous rainfall based on the signal transmission effect of the Internet of Things according to claim 1, characterized in that, The rainfall level matrix includes: The rainfall data is spatially nearest neighbor matched with IoT terminals to obtain a rainfall data matrix; The rainfall data matrix is ​​divided according to a preset rainfall level standard to obtain the rainfall level matrix.

4. The method for monitoring disastrous rainfall based on the signal transmission effect of the Internet of Things according to claim 2, characterized in that, The initial signal feature matrix is: Where C represents the initial signal feature matrix, Std represents the statistical features, m represents the number of IoT terminals, and t1, t2, t3, t4, t5, t6, t7, t8, t9, t1 ... n Indicates the start and end times of a preset time period.

5. The method for monitoring disastrous rainfall based on the signal transmission effect of the Internet of Things according to claim 1, characterized in that, The rainfall level matrix is ​​as follows: Where L represents the rainfall level matrix, L m This represents the rainfall intensity level at the m-th IoT terminal.

6. The method for monitoring disastrous rainfall based on the signal transmission effect of the Internet of Things according to claim 1, characterized in that, Spatial interpolation processing of the rainfall level estimation results includes: Based on the spatial distribution characteristics of multiple terminals, the inverse distance weighted interpolation method is used to perform spatial interpolation calculations on the rainfall level estimation results of discrete IoT nodes, generating continuous regional rainfall intensity level distribution data.

7. The method for monitoring disastrous rainfall based on the Internet of Things signal transmission effect according to claim 2, characterized in that, The signal parameters include a signal strength indicator.

8. The method for monitoring disastrous rainfall based on the signal transmission effect of the Internet of Things according to claim 2, characterized in that, The statistical characteristics include: sliding standard deviation.

Images