CN109600180B - Wireless communication intelligent flow sensing system based on frequency spectrum information - Google Patents
Wireless communication intelligent flow sensing system based on frequency spectrum information Download PDFInfo
- Publication number
- CN109600180B CN109600180B CN201811540307.5A CN201811540307A CN109600180B CN 109600180 B CN109600180 B CN 109600180B CN 201811540307 A CN201811540307 A CN 201811540307A CN 109600180 B CN109600180 B CN 109600180B
- Authority
- CN
- China
- Prior art keywords
- frequency spectrum
- data
- flow
- information
- spectrum information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/382—Monitoring; Testing of propagation channels for resource allocation, admission control or handover
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/391—Modelling the propagation channel
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/06—Testing, supervising or monitoring using simulated traffic
Abstract
The invention belongs to the technical field of wireless communication, and relates to a wireless communication intelligent flow sensing system based on frequency spectrum information. The traffic awareness system of the present invention is used for a heterogeneous wireless system, and includes: an offline learning subsystem: collecting system flow data and frequency spectrum information corresponding to time, calculating the characteristics of the frequency spectrum information, and then automatically obtaining a corresponding perception model of the flow data and the frequency spectrum data under the system by using machine learning; the on-line perception subsystem: and the terminal antenna collects the frequency spectrum information of the coexisting system in real time and senses the real-time flow state by using the frequency spectrum information and a sensing model obtained by the offline learning subsystem. Compared with the existing high-level information-based flow sensing system, the method provided by the invention does not need direct information interaction between heterogeneous systems, and only needs to utilize physical layer frequency spectrum information. Compared with the existing clustering receiver, the clustering receiver has the advantages of small deployment difficulty, low complexity and strong adaptability.
Description
Technical Field
The invention belongs to the technical field of wireless communication, and relates to a wireless communication intelligent flow sensing system based on frequency spectrum information.
Background
Heterogeneous wireless systems (such as coexistence systems of Wi-Fi and LAA-LTE) are important components of future wireless communication networks. Traffic awareness is an important component of heterogeneous wireless systems, and has the role of enabling one system to acquire traffic information of other systems that coexist to optimize transmission strategies. Its efficiency directly affects the availability and deployment complexity of the system. The existing traffic sensing method needs information of a Media Access Control (MAC) layer or a system upper layer (such as a network layer and a data link layer), that is, requires direct information interaction of a high-level system between heterogeneous wireless systems. This approach limits the deployment and application of heterogeneous wireless systems.
Disclosure of Invention
The invention aims to provide a heterogeneous wireless system flow sensing system based on physical layer spectrum information. The system is designed to eliminate the need for a heterogeneous system to have direct information transfer, thereby reducing deployment complexity and increasing its availability.
The technical scheme adopted by the invention is as follows:
a wireless communication intelligent flow sensing system based on frequency spectrum information, which is used for a heterogeneous wireless system, and is characterized by comprising:
an offline learning subsystem: collecting system flow data and frequency spectrum information corresponding to time, calculating the characteristics of the frequency spectrum information, and then automatically obtaining a corresponding perception model of the flow data and the frequency spectrum data under the system by using machine learning;
the on-line perception subsystem: and the terminal antenna collects the frequency spectrum information of the coexisting system in real time and senses the real-time flow state by using the frequency spectrum information and a sensing model obtained by the offline learning subsystem.
In order to overcome the defects caused by the fact that the existing heterogeneous wireless system flow sensing scheme needs to directly transmit MAC layer information and the information of the layers above the MAC layer information, the scheme of the invention directly collects the physical layer frequency spectrum information of the coexisting system, and the subsystem learns the corresponding relation between the frequency spectrum information and the flow under the system by using a regression model under the on-line state, so that the subsystem on the line can directly sense the flow information by using the collected physical layer information in real time.
Furthermore, the invention provides a heterogeneous wireless system flow sensing system based on first-order statistical characteristics of frequency spectrum information, which is characterized in that,
an offline learning subsystem: simultaneously collecting system flow data and frequency spectrum information and calculating first-order statistical characteristics of the frequency spectrum information, then analyzing various regression models to learn the corresponding relation between the flow data and the first-order statistical characteristics of the frequency spectrum data under the system and automatically obtaining an optimal corresponding regression model;
the on-line perception subsystem: and calculating the first-order statistical characteristics of the measured spectral data in real time, and then sensing the flow information in real time by using the model obtained by learning of the offline learning subsystem.
The spectral information statistical features used in the present invention are not limited to first order statistical features. The invention further provides a heterogeneous wireless system flow sensing system jointly using the first-order characteristic and the high-order characteristic of the frequency spectrum information, which is characterized in that,
the offline learning subsystem and the online sensing subsystem not only calculate the first-order statistical features of the collected spectrum data, but also calculate the high-order statistical features. And then, the plurality of statistical characteristics are jointly used for carrying out flow sensing by using a plurality of spectrum characteristics.
Compared with the existing high-level information-based flow sensing system, the method provided by the invention does not need direct information interaction between heterogeneous systems, and only needs to utilize physical layer frequency spectrum information. Compared with the existing clustering receiver, the clustering receiver has the advantages of small deployment difficulty, low complexity and strong adaptability.
Drawings
Fig. 1 shows the overall design of the flow sensing system proposed by the present invention;
FIG. 2 shows a design of an offline learning subsystem proposed by the present invention;
FIG. 3 shows a design of an on-line perception subsystem proposed by the present invention;
FIG. 4 illustrates an offline subsystem data processing flow proposed by the present invention;
FIG. 5 illustrates an optimal regression model auto-learning scheme proposed by the present invention;
FIG. 6 shows a spectrum-flow fit plot for a Wi-Fi in a use case of the present invention;
figure 7 shows a comparison of mean square errors for different fits at a Wi-Fi in a use case of the present invention.
Detailed Description
The invention is described in detail below with reference to the drawings and simulation examples so that those skilled in the art can better understand the invention.
The invention provides a system for sensing flow in a heterogeneous wireless system by using physical layer frequency spectrum information, which is divided into an offline learning subsystem and an online sensing subsystem. The operation steps of the system are as follows:
1. data acquisition: acquiring frequency spectrum data and flow data corresponding to time, wherein each data corresponds to a period of fixed length, and acquiring N data to form a training set;
2. model training: carrying out model training by using the acquired training set to obtain an optimal model and corresponding model parameters;
3. and (3) online perception: and collecting frequency spectrum data in real time and inputting the frequency spectrum data into the trained model, and outputting the flow sensing information in real time by the model.
Fig. 1 shows the overall design of the system proposed by the invention. The input of the offline learning subsystem is a frequency spectrum data training set and a flow data training set. The subsystem outputs a corresponding perception model between the spectrum information and the flow information corresponding to the heterogeneous wireless system. The input of the on-line sensing subsystem is frequency spectrum information and a sensing model which are collected in real time, and sensed flow information is output in real time.
Fig. 2 shows a design of the offline learning subsystem proposed by the present invention. The subsystem directly obtains a flow training set from a perception target system and acquires corresponding frequency spectrum data through a frequency spectrum testing instrument. The data sets are stored in a database. According to the training set data in the database, the subsystem performs model learning and outputs a corresponding perception model.
Fig. 3 shows a design of an on-line sensing subsystem proposed by the present invention. The frequency spectrum testing instrument collects frequency spectrum information of a target system in real time and inputs the frequency spectrum information into a real-time streaming frequency spectrum database. The sensing model acquires real-time frequency spectrum data from the streaming database as input and outputs flow information obtained by real-time sensing in real time.
Fig. 4 shows a data processing flow of the offline learning subsystem proposed by the present invention. The time length of each processing unit in the time domain is T. For nth T long time unit, flow dataThe collection and calculation mode of (A) is to carry out the total data throughput in the time period
Dividing by T, namely:
meanwhile, in the nth T long time unit, the physical layer spectrum information is continuously and uniformly sampled by J points to obtain the spectrum information vector in the time unit, namely the spectrum information vector in the time unit
After continuously collecting data of N T long-time units, obtaining a traffic data training set and a spectrum data training set, namely:
inputting the training set into a training module, wherein the training module comprises W pre-defined alternative regression models, and the set of the models is as follows:
after training, the optimal model is output for use by the online sensing subsystem.
Fig. 5 shows a scheme of auto-learning an optimal regression model proposed by the present invention. And (3) dividing the training set data into training data and verification data in the database, wherein the frequency spectrum data is subjected to feature extraction. The training data are respectively input into different regression models to obtain corresponding optimal model parameters. The optimal model of different regression models will be tested and the mean square error calculated with the verification data obtained by segmentation as input. The model with the minimum mean square error is output as a perception model under the system.
The first spectrum data feature extraction method provided by the invention is to calculate first-order statistical features, namely:
the invention further provides a feature extraction mode of the mixed high-order statistical features, which comprises the following specific modes:
1. extracting first-order statistical characteristics;
2. sequentially extracting second-order and fourth-order statistical characteristics from the second order;
3. and splicing the extracted statistical features into vectors to form mixed high-order features which are used as the input of a regression algorithm.
The calculation method of the high-order features comprises the following steps:
1. variance:
2. kurtosis:
the invention further takes a Wi-Fi target system as an example to explain the specific use method of the wireless communication system flow sensing system based on the frequency spectrum information.
In the system, the unit sampling time length T is 30 seconds, the flow unit is megabyte per second, the frequency spectrum information is the total energy corresponding to the bandwidth, and the unit is dBm. The corresponding relationship between the spectrum information and the traffic information collected at different user access times at a certain time is shown in fig. 6.
Fig. 6 shows a fitting curve obtained by fitting different polynomials in an illustrative example of the present invention, where "poly-n" represents fitting using an nth order polynomial as a model, and "Sigmoid" represents fitting using a Sigmoid function as a model.
Fig. 7 shows fitting errors obtained by using different polynomial fitting manners in an illustrative example of the present invention, where a term with abscissa of 6 indicates that the fitting model is a sigmoid model. It can be seen that a model of an excessively high order is not suitable, and the sigmoid model achieves the best verification effect under the system and is the optimal perception model of the target system.
Claims (4)
1. A wireless communication intelligent flow sensing system based on frequency spectrum information, the flow sensing system is used for a heterogeneous wireless system, the flow sensing is used for enabling one system in the heterogeneous wireless system to obtain the flow information of other coexisting systems, and the other systems with the obtained flow information are defined as monitored systems, and the system is characterized by comprising:
an offline learning subsystem: collecting monitored system flow data and frequency spectrum information corresponding to time, calculating the characteristics of the frequency spectrum information, and then automatically obtaining a corresponding sensing model of the flow data and the frequency spectrum data under the system by using machine learning;
the on-line perception subsystem: and the terminal antenna collects the frequency spectrum information of the heterogeneous wireless system in real time and senses the real-time flow state by using the frequency spectrum information and a sensing model obtained by the offline learning subsystem.
2. The system according to claim 1, wherein the specific method for acquiring the system traffic data and the spectrum information corresponding to the time is as follows:
setting the time length of each processing unit in the time domain as T, and setting the flow data for the nth T long time unit
The collection and calculation mode of (A) is to carry out the total data throughput in the time period
Dividing by T, namely:
meanwhile, in the nth T long time unit, J points of sampling the physical layer spectrum information continuously and uniformly are obtained, and a spectrum information vector in the time unit is obtained, that is:
after continuously collecting data of N T long-time units, obtaining a flow data training set R and a spectrum data training set F, namely:
3. the system according to claim 2, wherein the spectral information-based wireless communication intelligent flow sensing system is characterized in that feature extraction is performed on spectral data collected by the spectral data training set F and the on-line sensing subsystem, and the feature of the calculated spectral information is to calculate a first-order statistical feature and a high-order statistical feature of the spectral information:
1) calculating the first order statistical features:
wherein E [. cndot. ] represents the expectation, i.e. the first order statistical feature;
2) calculating first order statistical features and higher order statistical features, including:
a. calculating the first order statistical features:
wherein E [. cndot. ] represents the expectation, i.e. the first order statistical feature;
b. calculating the second moment characteristic variance and the fourth moment characteristic kurtosis in the following specific mode:
variance:
kurtosis:
c. splicing the extracted first-order, second-order and fourth-order statistical features into a three-dimensional vector to form mixed high-order features; the spectrum feature extraction technology is respectively applied to a spectrum data training set F and spectrum data acquired by an online sensing subsystem in real time.
4. The system according to claim 3, wherein the specific method for automatically obtaining the corresponding sensing model of the flow data and the spectrum data in the system by using machine learning comprises:
in the on-line learning subsystem, a flow data training set R and a characteristic data set obtained by extracting characteristics from a spectrum data training set F are used
And the optimal models of the different regression models are obtained by adopting verification data obtained by segmentation as input to test and calculate the mean square error, and the model with the minimum mean square error takes the output as a corresponding perception model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811540307.5A CN109600180B (en) | 2018-12-17 | 2018-12-17 | Wireless communication intelligent flow sensing system based on frequency spectrum information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811540307.5A CN109600180B (en) | 2018-12-17 | 2018-12-17 | Wireless communication intelligent flow sensing system based on frequency spectrum information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109600180A CN109600180A (en) | 2019-04-09 |
| CN109600180B true CN109600180B (en) | 2020-01-03 |
Family
ID=65962612
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811540307.5A Active CN109600180B (en) | 2018-12-17 | 2018-12-17 | Wireless communication intelligent flow sensing system based on frequency spectrum information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109600180B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010137777A1 (en) * | 2009-05-28 | 2010-12-02 | Lg Electronics Inc. | Apparatus and method for determination of frame structure for reducing interference in frequency reuse system using cognitive radio |
| CN102130798A (en) * | 2011-03-25 | 2011-07-20 | 中国电子科技集团公司第三十研究所 | Method and device for detecting multidimensional flow anomalies of distributed network |
| CN103141142A (en) * | 2010-09-30 | 2013-06-05 | 频谱桥公司 | System and method for collaborative spectrum analysis |
| WO2016043739A1 (en) * | 2014-09-17 | 2016-03-24 | Resurgo, Llc | Heterogeneous sensors for network defense |
| CN106535104A (en) * | 2016-12-16 | 2017-03-22 | 中南大学 | Adaptive Bluetooth transmission method based on flow perception |
| CN107360032A (en) * | 2017-07-20 | 2017-11-17 | 中国南方电网有限责任公司 | A kind of network stream recognition method and electronic equipment |
| CN108173610A (en) * | 2018-02-11 | 2018-06-15 | 南京邮电大学 | The collaborative frequency spectrum sensing method of heterogeneous wireless network based on second-order statistic |
-
2018
- 2018-12-17 CN CN201811540307.5A patent/CN109600180B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010137777A1 (en) * | 2009-05-28 | 2010-12-02 | Lg Electronics Inc. | Apparatus and method for determination of frame structure for reducing interference in frequency reuse system using cognitive radio |
| CN103141142A (en) * | 2010-09-30 | 2013-06-05 | 频谱桥公司 | System and method for collaborative spectrum analysis |
| CN102130798A (en) * | 2011-03-25 | 2011-07-20 | 中国电子科技集团公司第三十研究所 | Method and device for detecting multidimensional flow anomalies of distributed network |
| WO2016043739A1 (en) * | 2014-09-17 | 2016-03-24 | Resurgo, Llc | Heterogeneous sensors for network defense |
| CN106535104A (en) * | 2016-12-16 | 2017-03-22 | 中南大学 | Adaptive Bluetooth transmission method based on flow perception |
| CN107360032A (en) * | 2017-07-20 | 2017-11-17 | 中国南方电网有限责任公司 | A kind of network stream recognition method and electronic equipment |
| CN108173610A (en) * | 2018-02-11 | 2018-06-15 | 南京邮电大学 | The collaborative frequency spectrum sensing method of heterogeneous wireless network based on second-order statistic |
Non-Patent Citations (1)
| Title |
|---|
| LTE-LAA网络频谱利用方法研究;付卓然;《中国优秀硕士学位论文全文数据库》;20181031(第10期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109600180A (en) | 2019-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109347668B (en) | Training method and device for service quality assessment model | |
| CN109543829A (en) | Method and system for hybrid deployment of deep learning neural network on terminal and cloud | |
| CN109905696A (en) | A kind of recognition methods of the Video service Quality of experience based on encryption data on flows | |
| JP7351082B2 (en) | Interference source prediction model generation method, interference source recognition method, device, and terminal device | |
| CN110675395A (en) | Intelligent on-line monitoring method for power transmission line | |
| CN111756600B (en) | Multi-communication system and method for realizing multiple switch test machines | |
| CN106612511B (en) | Wireless network throughput evaluation method and device based on support vector machine | |
| CN111174370A (en) | Fault detection method and device, storage medium and electronic device | |
| CN109640335B (en) | Wireless sensor fault diagnosis method based on convolutional neural network | |
| CN103957551B (en) | Communications status method for real-time monitoring in automatic weather station | |
| CN111862951A (en) | Voice endpoint detection method and device, storage medium and electronic equipment | |
| CN112307969A (en) | Pulse signal classification identification method and device and computer equipment | |
| CN110838075A (en) | Training and predicting method and device for prediction model of transient stability of power grid system | |
| CN112986492A (en) | Method and device for establishing gas concentration prediction model | |
| CN105427047A (en) | Multi-dimensional operation and maintenance data fusion method and system for power transmission and transformation equipment | |
| CN111224726A (en) | Video live broadcast system based on long-term and short-term memory network and implementation method thereof | |
| CN113612650A (en) | Monitoring method for edge computing equipment | |
| CN109450573A (en) | A kind of frequency spectrum sensing method based on deep neural network | |
| CN114879081A (en) | Lightning damage area analysis method based on synchronous dynamic monitoring data of lightning arrester | |
| CN117388893B (en) | Multi-device positioning system based on GPS | |
| CN114413409A (en) | Detection method and device for air conditioner fault probability and intelligent air conditioner | |
| CN109600180B (en) | Wireless communication intelligent flow sensing system based on frequency spectrum information | |
| CN116684878B (en) | 5G information transmission data safety monitoring system | |
| CN108613820A (en) | A kind of online allophone monitoring algorithm for GIS bulk mechanicals defect diagonsis and positioning | |
| CN114584230B (en) | Predictive channel modeling method based on countermeasure network and long-term and short-term memory network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |