A data compression and transmission method for wireless wind speed sensor

Through the compression perception method combined with wavelet dictionary and MQTT protocol, the problem of high energy consumption and poor robustness in wireless wind speed sensor data transmission is solved, and the wind speed data transmission with low energy consumption, low bandwidth, and high robustness is achieved. Effectively reconstruct wind speed data in the absence.

CN118474794BActive Publication Date: 2024-11-19ZHEJIANG UNIV
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202410663791.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-27
Publication Date
2024-11-19
Estimated Expiration
2044-05-27

AI Technical Summary

Technical Problem

Wireless wind speed sensors have problems such as high energy consumption, large bandwidth usage and poor robustness during data acquisition and transmission, especially the poor reconstruction of non-stationary wind speed data, and the existing methods have failed to effectively handle data transmission. Missing problem.

Method used

The compression perception method based on wavelet dictionary is used to convert wind speed data into compressed data sequences, and transmit it through the MQTT lightweight protocol. It uses two threads to achieve continuous acquisition and transmission, correct data loss during transmission, and passes through LASSO algorithm or orthogonal matching. Tracking algorithm reconstructs wind speed data.

Benefits of technology

It realizes wireless transmission of wind speed data with low energy consumption, low bandwidth and high robustness, improves data transmission accuracy and applicability, and can effectively reconstruct the original wind speed data in the absence of data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118474794B_ABST
    Figure CN118474794B_ABST
Patent Text Reader

Abstract

A data compression and transmission method for a wireless wind speed sensor comprises: A. The sensor end collects wind speed data at a fixed frequency, records the timestamp of the first collection and the wind speed data sequence collected, and converts the collected wind speed data sequence into a compressed data sequence. B. The timestamp of the first collection and the compressed data sequence are transmitted from the sensor end to the data terminal using the Message Queuing Telemetry Transport (MQTT) lightweight communication protocol. C. The wind speed data sequence is reconstructed in the data terminal using a compressed sensing algorithm based on a wavelet dictionary. D. The sensor is installed in an actual environment for testing, and a complete section of data is pre-collected to test the transmission effect and determine the transmission parameters. E. The adjusted transmission parameters are sent to the wind speed sensor, and the wind speed data is collected and transmitted for a long time using dual threads. The present invention realizes low-energy, low-bandwidth, and highly robust wireless transmission of wind speed data, which ensures the integrity of the wind speed data while effectively extending the service life of the wireless wind speed transmission equipment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of environmental wind speed monitoring and data transmission, and in particular to a compressed sensing data transmission method. Background Art

[0002] In recent years, engineering structures have developed towards being light, tall, and long-span. Such engineering structures are often highly flexible, have low damping, and low natural frequency, and are therefore very sensitive to wind loads, which have become a key control load in structural design. Wind monitoring has become an effective means of understanding the wind load characteristics of service structures, and wind speed is an important parameter for structural wind monitoring.

[0003] The emerging wireless sensor network technology has the advantages of low cost, easy installation and maintenance, large deployment scale, and low interference, and is gradually being used for wind monitoring of large civil structures. Wireless health monitoring systems are limited by sensor battery life and transmission bandwidth, making it difficult to conduct long-term data collection. Compressed sensing is a new sampling theory that can reconstruct the original wind speed data with less sampled data, which is expected to solve the energy consumption problem of wireless sensor networks and improve the working life of wireless health monitoring systems.

[0004] Choosing a suitable sparse dictionary is critical for reconstructing wind speed data, which is usually not naturally sparse. When commonly used dictionaries fit non-stationary data, the Gibbs effect will occur, resulting in a decrease in the sparsity of the data, which is an important factor affecting the effect of compressed sensing data transmission. Common dictionaries show great advantages in reconstructing acceleration data, but are not effective in reconstructing wind speed data with strong non-stationarity. In addition, existing transmission methods rarely consider the problem of data missing that may exist during data transmission, and data transmission is not very robust. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention proposes a data compression and transmission method for a wireless wind speed sensor to achieve wireless transmission of wind speed data with low energy consumption, low bandwidth and high robustness.

[0006] A data compression and transmission method for a wireless wind speed sensor comprises the following steps:

[0007] A. The sensor collects wind speed data at a fixed frequency, records the timestamp of the first collection and the collected wind speed data sequence, and converts the collected wind speed data sequence into a compressed data sequence.

[0008] B. Use the Message Queuing Telemetry Transport (MQTT) lightweight communication protocol to transmit the first collected timestamp and compressed data sequence from the sensor end to the data terminal.

[0009] C. Reconstruct the wind speed data sequence in the data terminal using the compressive sensing algorithm based on the wavelet dictionary.

[0010] D. Install the sensor in the actual environment for testing, pre-collect a complete set of data to test the transmission effect, and determine the transmission parameters.

[0011] E. Send the adjusted transmission parameters to the wind speed sensor, and collect and transmit the wind speed data in a long-term dual-threaded manner.

[0012] Further, step A specifically includes:

[0013] A1. Set the sampling frequency (usually 1 Hz) and collect the wind speed data;

[0014] A2. Record the first data collection timestamp t and the wind speed data value x collected at the i-th time i , and generate an n-dimensional wind speed data sequence x;

[0015] A3. At the sensor end, pre-generate an m×n (m < n) order random Bernoulli observation matrix Φ, and use the observation matrix to convert the wind speed data sequence x into an m-dimensional compressed data sequence y. This process can be expressed as:

[0016] y = Φx. (1)

[0017] Further, step B specifically includes:

[0018] B1. The sensor end uses the lightweight Message Queuing Telemetry Transport (MQTT) protocol to transmit the compressed data sequence back to the data terminal. The Quality of Service (QoS) of the MQTT protocol information service is set to 0, that is, the sensor device transmits the message once. This transmission mode has low energy consumption, but there may be data loss problems during the transmission process;

[0019] B2. When the sensor end sends the compressed data sequence, the content of a single transmission includes a single-precision floating-point number and a short integer. The single-precision floating-point number represents the i-th value y in the compressed data sequence i , and the short integer represents its index value i;

[0020] B3. The data terminal receives the data, and records the m' single-precision floating-point numbers and short integers received, and temporarily stores them as the received compressed data sequence y' and the index value sequence {i}. If there is data loss during the transmission process, then m' < m.

[0021] Further, step C specifically includes:

[0022] C1. Discrete wavelet function ψj,k (t) can be expressed as

[0023]

[0024] Where ψ(t) represents the initial wavelet function, a is the scaling factor, and b is the translation factor. j , b=2 j k, j, k∈Z, we get the binary wavelet function as follows

[0025] ψ j,k (t) = 2 -j / 2 ψ(2 j t-k2); (3)

[0026] C2. A specific type of wavelet includes an initial father function and an initial mother function, which are respectively denoted as φ(t), after scaling and translation, we get φ j,k (t). The parent function only takes the case of j = L-1, arranges them in order, and records them as P; the mother function takes the case of j = 0, 1, ..., L-1, arranges them in order, and combines the mother functions with the same j value together, and records them as Q j

[0027]

[0028] P, Q j Combine them and use the wavelet function to construct a wavelet dictionary Ψ with a layer number of L, which can be expressed as:

[0029] Ψ=[P Q0 Q1 … Q L-1 ]; (5)

[0030] C3. The coefficients of wind speed data series usually expanded in the wavelet dictionary have sparse characteristics, namely:

[0031] x=Ψw (6)

[0032] Where w is the sparse coefficient. Combined with formula (1), the relationship between the compressed data sequence and the sparse coefficient can be expressed as:

[0033] y=Φx=ΦΨw=Aw (7)

[0034] Where A = ΦΨ is the measurement matrix. If the i-th compressed data sequence is missing during transmission, remove the i-th row in the measurement matrix A and correct it to the measurement matrix A'. At this time, the relationship between the received compressed data sequence and the sparse coefficients can be obtained:

[0035] y'=A'w; (8)

[0036] C4. When w is sparse, the following optimization problem can be solved by sparse algorithms such as LASSO algorithm or orthogonal matching pursuit algorithm:

[0037]

[0038] Among them, ||·||1 represents the l1 norm, that is, the sum of the absolute values ​​of each element in the vector. Solve to get the estimated value of the sparse coefficient w After that, the reconstructed wind speed data sequence can be obtained

[0039]

[0040] C5. According to the collected acquisition frequency and the time of the first data acquisition, the acquisition time corresponding to each data value in the wind speed data sequence can be obtained.

[0041] Further, step D specifically includes:

[0042] D1. Test the sensor in the actual environment, collect a complete piece of data in advance and transmit it back to the data terminal;

[0043] D2. Evaluate the wind speed data collection effect under different wind speed data length n and compressed data length m combinations, using mean square error As an indicator to measure the effectiveness of wind speed data transmission;

[0044] D3. The longer the wind speed data sequence length, the longer the delay time between collection and reception. Stable data reconstruction requires the wind speed data sequence length to reach a certain threshold. The shorter the compressed data sequence, the lower the transmission energy consumption and the greater the reconstructed data error. Therefore, it is necessary to select the wind speed data length n and compressed data length m according to the transmission effect and actual collection requirements.

[0045] Further, step E specifically includes:

[0046] E1. Send the adjusted transmission parameters to the wind speed sensor, collect the wind speed data sequence and process it into a compressed data sequence;

[0047] E2. In order to perform continuous and uninterrupted data collection, and considering that sending data requires a certain amount of time, the sensor performs dual-threaded tasks. The first task is to send the compressed data sequence of the previous time period, and the second task is to collect the wind speed data sequence of the current time to achieve continuous collection of wind speed data.

[0048] The present invention utilizes compressed sensing technology and only needs to transmit a small amount of data through the MQTT lightweight protocol to restore the wind speed data sequence. The wavelet dictionary is constructed to improve the sparse characteristics of the wind speed sparse coefficient, reduce the error between the reconstructed wind speed data and the original wind speed data, and effectively improve the data transmission accuracy. The present invention realizes low-energy, low-bandwidth, and high-robustness wireless transmission of wind speed data.

[0049] The advantages of the present invention are: first, the compressed sensing technology is used and a wavelet dictionary is constructed to improve the sparse expression of wind speed data, and the original wind speed data can be restored by transmitting a small number of data points, reducing the energy consumed and bandwidth occupied by the sensor when transmitting data. Second, the lightweight MQTT protocol is used to transmit data to further reduce the energy consumption of data transmission, and considering the possible data missing problem, the observation matrix is ​​corrected according to the received data index number, and the ability to reconstruct the original wind speed data is still possessed, which effectively improves the robustness of the transmission method. Third, considering the site randomness of wind speed, a sensor parameter adjustment optimization scheme is given to effectively improve the applicability of the transmission method. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a flow chart of the method of the present invention;

[0051] Figure 2 are the initial father function and the initial mother function of the sym7 wavelet used in the present invention;

[0052] Figure 3 It is a scene photo of the test scene of the present invention;

[0053] Figure 4 It is the transmission effect of the present invention under different parameter values;

[0054] Figure 5 It is an indicator of data transmission accuracy under different parameter values ​​of the present invention. DETAILED DESCRIPTION

[0055] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings.

[0056] This embodiment relates to a data compression and transmission method for a wireless wind speed sensor. Figure 1 As shown, the following steps are included:

[0057] A. The sensor collects wind speed data at a fixed frequency, records the timestamp of the first collection and the collected wind speed data sequence, and converts the collected wind speed data sequence into a compressed data sequence. Specifically, it includes:

[0058] A1. The sensor collects wind speed data at a sampling frequency of 1 Hz.

[0059] A2. Record the first data collection timestamp t and the i-th collected wind speed data value x i , generate m-dimensional wind speed data sequence x;

[0060] A3. At the sensor end, a random Bernoulli measurement matrix Φ of order m×n (m<n) is generated in advance, and the collected wind speed data sequence x is converted into an m-dimensional compressed data sequence y using the measurement matrix.

[0061] B. Use the Message Queuing Telemetry Transport (MQTT) lightweight communication protocol to transmit the compressed data sequence from the sensor end to the data terminal.

[0062] B1. The sensor uses the Message Queuing Telemetry Transport (MQTT) lightweight protocol to transmit the compressed data sequence back to the data terminal, and the MQTT protocol information quality of service (QoS) is set to 0.

[0063] B2. When the sensor sends a compressed data sequence, a single transmission includes a single-precision floating-point number and a short integer. The single-precision floating-point number represents the i-th value y in the compressed data sequence. i , the short integer represents its index value i;

[0064] B3. The data terminal receives data and records the received m' single-precision floating point numbers and short integers, temporarily storing them as the received compressed data sequence y' and index value sequence {i}. If there is data missing during the transmission process, m'<m.

[0065] C. Reconstruct the wind speed data sequence using the compressed sensing algorithm based on wavelet dictionary in the data terminal.

[0066] C1. Binary wavelet function ψ j,k (t) can be expressed as

[0067] ψ j,k (t) = 2 -j / 2 ψ(2 j t-k2); (1)

[0068] C2. This example uses the sym7 wavelet test, which includes the initial parent function and the initial generating function φ(t), such as Figure 2 As shown, after scaling and translation, we can get φ j,k(t). The parent function only takes the case of j = 9, arranges them in order, and records them as P; the mother function takes the case of j = 0, 1, ..., 9, arranges them in order, and combines the mother functions with the same j value together, and records them as Q j ; P, Q j The wavelet dictionary Ψ with the number of layers L is constructed by combining the wavelet functions.

[0069] C3. Calculate the observation matrix A, and obtain the modified observation matrix A' according to the received data index value;

[0070] C4. Solve the following optimization problem using the LASSO algorithm:

[0071]

[0072] The above formula can be transformed into an optimization problem that iteratively solves the following formula:

[0073]

[0074] Where α represents the regularization parameter that controls the sparsity. The estimated value of the sparse coefficient w is obtained After that, the reconstructed wind speed data sequence can be obtained

[0075]

[0076] C5. According to the collected acquisition frequency and the time of the first data acquisition, the acquisition time corresponding to each data value in the wind speed data sequence can be obtained.

[0077] D. Install the sensor in the actual environment for testing, and pre-collect a complete set of data to test the transmission effect and determine the transmission parameters.

[0078] D1. Test the sensor in a real environment. Figure 3 As shown, a complete data of 300 characters in length is collected in advance and transmitted back to the data terminal;

[0079] D2. Evaluate the wind speed data collection effect under different wind speed data length n and compressed data length m combinations, using mean square error As an indicator to measure the effect of wind speed data transmission, the data collection effect is as follows: Figure 4 As shown;

[0080] D3. Test the mean square error of data transmission under different parameter combinations. The results are as follows Figure 5 As shown; if the mean square error is required not to exceed 6%, the wind speed data length n is set to 240, and the compressed data length m is set to 168, which is expected to save 30% of transmission energy consumption.

[0081] E. Send the adjusted transmission parameters to the wind speed sensor, and use dual threads to collect and transmit wind speed data for a long time.

[0082] E1. Send the adjusted transmission parameters to the wind speed sensor, collect the wind speed data sequence and process it into a compressed data sequence;

[0083] E2. In order to perform continuous and uninterrupted data collection, and considering that sending data requires a certain amount of time, the sensor performs dual-threaded tasks. The first task is to send the compressed data sequence of the previous time period, and the second task is to collect the wind speed data sequence of the current time to achieve continuous collection of wind speed data.

[0084] The contents described in the embodiments of this specification are merely an enumeration of the implementation forms of the inventive concept. The protection scope of the present invention should not be regarded as limited to the specific forms described in the embodiments. The protection scope of the present invention also extends to equivalent technical means that can be conceived by those skilled in the art based on the inventive concept.

Claims

1. A data compression and transmission method for a wireless wind speed sensor, comprising the following steps: A. The sensor end collects wind speed data at a fixed frequency, records the timestamp of the first collection and the wind speed data sequence obtained by the collection, and converts the collected wind speed data sequence into a compressed data sequence; B. Use the Message Queuing Telemetry Transport (MQTT) lightweight communication protocol to transmit the timestamp of the first collection and the compressed data sequence from the sensor end to the data terminal; C. Reconstruct the wind speed data sequence using a compressive sensing algorithm based on a wavelet dictionary in the data terminal; D. Install the sensor in the actual environment for testing, pre-collect a complete set of data to test the transmission effect, and determine the transmission parameters; E. Send the adjusted transmission parameters to the wind speed sensor, and continuously collect and transmit wind speed data in a dual-threaded manner; Step A specifically includes: A1. Set the sampling frequency and collect wind speed data; A2. Record the first data collection timestamp t and the i-th collected wind speed data value x i , generate n-dimensional wind speed data sequence x; A3. At the sensor end, pre-generate an m×n order random Bernoulli observation matrix Φ, where m < n, and use the observation matrix to convert the wind speed data sequence x into an m-dimensional compressed data sequence y. This process is expressed as: y = Φx (1) Step B specifically includes: B1. The sensor end uses the Message Queuing Telemetry Transport (MQTT) lightweight communication protocol to transmit the compressed data sequence back to the data terminal, and the information service quality (QoS) of the MQTT lightweight communication protocol is set to 0; B2. When the sensor sends a compressed data sequence, a single transmission includes a single-precision floating-point number and a short integer. The single-precision floating-point number represents the i-th value y in the compressed data sequence. i , the short integer represents its index value i; B3. The data terminal receives the data, records the m' single-precision floating-point numbers and short integers received, and temporarily stores them as the received compressed data sequence y' and the index value sequence {i}. If there is data loss during the transmission process, then m' < m.

2. A data compression and transmission method for a wireless wind speed sensor as claimed in claim 1, characterized in that: Step E specifically includes: E1. Send the adjusted transmission parameters to the wind speed sensor, collect the wind speed data sequence and process it into a compressed data sequence; E2. In order to be able to perform continuous and uninterrupted data collection, and considering that it takes a certain amount of time to send data, the sensor executes a dual-threaded task. The first task is to send the compressed data sequence of the previous time period, and the second task is to collect the wind speed data sequence of the current time, so as to achieve continuous collection of wind speed data.

Citation Information

Patent Citations

  • Cloud computing platform-based wind generator state monitoring method and apparatus adopting machine learning

    CN107608937A

  • Data compression method, device and system and storage medium

    CN115941806A

  • Structural health monitoring missing data reconstruction method based on WGANGP-Unet

    CN116502060A

  • Method for Pan-Sharpening Panchromatic and Multispectral Images Using Wavelet Dictionaries

    US20130129201A1