CN113286309A - Heterogeneous communication method and system based on CSI - Google Patents
Heterogeneous communication method and system based on CSI Download PDFInfo
- Publication number
- CN113286309A CN113286309A CN202110554717.0A CN202110554717A CN113286309A CN 113286309 A CN113286309 A CN 113286309A CN 202110554717 A CN202110554717 A CN 202110554717A CN 113286309 A CN113286309 A CN 113286309A
- Authority
- CN
- China
- Prior art keywords
- csi
- zigbee
- wifi
- signal
- 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.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/18—Network planning tools
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/22—Traffic simulation tools or models
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/06—Optimizing the usage of the radio link, e.g. header compression, information sizing, discarding information
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Signal Processing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Networks & Wireless Communication (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention relates to a heterogeneous communication method and a system based on CSI, wherein the method at least comprises the following steps: under the condition that the ZigBee signal is overlapped with the WiFi signal in a frequency domain overlapping mode and embedded with the CSI sequence, the WiFi signal is received, and information characteristics of the CSI sequence containing the ZigBee code are classified based on a classification model. On the basis that overlapping frequency spectrum distribution exists between the WiFi signal and the ZigBee signal, the direct contact can be established by analyzing the interference WiFi CSI sequence caused by the encoded ZigBee signal through the overlapping frequency spectrum, and the direct communication is realized between incompatible wireless devices.
Description
Technical Field
The present invention relates to the field of wireless communication technologies, and in particular, to a heterogeneous communication method and system based on CSI.
Background
Cross-technology communication (CTC) is intended to enable direct communication between incompatible wireless devices without the need for additional hardware. A key challenge is to embed symbol information in the physical spectrum so that it can be accurately decoded by another heterogeneous communication protocol.
According to the prediction of IOT Analytics, the number of globally active internet of things devices will reach 220 hundred million by 2025. The rich diversity of internet of things devices presents challenges to the interconnection of heterogeneous devices. This will result in congestion of the radio spectrum. Devices using different wireless technologies (e.g., WiFi, ZigBee, and bluetooth) must share unlicensed spectrum (e.g., the ISM band) when public spaces coexist.
Traditionally, bridging wireless technology enables indirect connections between heterogeneous devices through multiple wireless gateways. Gateways, also known as protocol converters, are internet connectors used for the transmission and exchange of information between networks. By providing multiple wireless interfaces, the local gateway can convert data from different devices, conforming to different communication standards. However, this approach can introduce additional hardware, maintenance costs, deployment complexity, and traffic overhead into and out of the gateway.
The prior art proposes a cross-protocol communication technique that allows direct communication between incompatible devices without the need for additional hardware. Existing cross-protocol communication technologies can be divided into packet-level cross-protocol communication and physical-level cross-protocol communication. The core idea of the packet-level cross-protocol communication technology is to embed transmitted symbol information into packet-level characteristics (including packet length, time information, sequence characteristics and the like), so as to realize direct communication and cross-technology communication among WiFi, ZigBee and Bluetooth. However, the core idea of the packet-level cross-protocol communication technology is to embed transmitted symbol information into packet-level features, thereby realizing direct communication and cross-technology communication among WiFi, ZigBee and bluetooth. Packet-level cross-technology communication cannot achieve higher throughput rates due to limitations on the number and energy of packets.
For example, chinese patent CN107682830B discloses a method and apparatus for ZigBee to WiFi communication. The communication method comprises the following steps: the method comprises the steps that a second WiFi terminal obtains link information of a WiFi link from a first WiFi terminal to the second WiFi terminal, wherein the link information at least comprises CSI sequences of a first frequency band and sub-channels in each decoding window; the second WiFi terminal analyzes each section of CSI sequence to obtain each piece of data to be sent, wherein the sub-channel is in the coverage range of a second frequency band, the second frequency band is the frequency band of the ZigBee terminal for sending the ZigBee data packet, the second frequency band is in the coverage range of the first frequency band, the ZigBee terminal is used for analyzing the data to be sent to judge whether the ZigBee data packet is sent or not in a preset sending window corresponding to each piece of data to be sent, if the data is sent, the ZigBee data packet is sent in the second frequency band in the time period corresponding to the sending window, the decoding window corresponds to the sending window, and the length of the decoding window and the length of the sending window are equal.
For example, chinese patent CN107750061B discloses an adaptive communication method from ZigBee to WiFi. The method comprises the steps that a second WiFi terminal sends first data and second data to a first WiFi terminal and a ZigBee terminal at a first frequency band, so that the first WiFi terminal establishes a WiFi link from the first WiFi terminal to the second WiFi terminal at the first frequency band according to first sending power, the WiFi packet length and the WiFi packet interval, and the ZigBee terminal judges whether to send a ZigBee data packet at a sending window corresponding to each bit of data to be sent or not according to analysis of the data to be sent; and the second WiFi terminal acquires CSI information of the link information of the WiFi link and analyzes the CSI information to obtain the data to be sent.
However, the bridging wireless technology in the prior art enables indirect connection between heterogeneous devices through a plurality of wireless gateways. This approach can introduce additional hardware, maintenance costs, deployment complexity, and traffic overhead into and out of the gateway.
The invention provides a CSI heterogeneous communication method and system, which can realize heterogeneous communication from ZigBee to WiFi without additional hardware support.
Furthermore, on the one hand, due to the differences in understanding to the person skilled in the art; on the other hand, since the inventor has studied a lot of documents and patents when making the present invention, but the space is not limited to the details and contents listed in the above, however, the present invention is by no means free of the features of the prior art, but the present invention has been provided with all the features of the prior art, and the applicant reserves the right to increase the related prior art in the background.
Disclosure of Invention
In view of the deficiencies of the prior art, the present invention provides a CSI-based heterogeneous communication method, which at least comprises: under the condition that the ZigBee signal is overlapped with the WiFi signal in a frequency domain overlapping mode and embedded with the CSI sequence, the WiFi signal is received, and information characteristics of the CSI sequence containing the ZigBee code are classified based on a classification model. According to the method, the CSI sequence containing the ZigBee signal information is selected by utilizing the signal change influence of the ZigBee signal and the WiFi signal on the WiFi signal in the overlapped frequency domain, and the ZigBee signal information is obtained by classifying the CST sequence. According to the invention, the ZigBee signal sent by the ZigBee sending terminal is embedded into the WiFi signal for transmission, and can be received by the WiFi receiving terminal and the ZigBee signal information can be obtained, so that a good effect of enabling the ZigBee sending terminal and the WiFi receiving terminal to communicate without arranging extra hardware is achieved.
Preferably, the method further comprises: and extracting amplitude characteristics of the CSI sequence, and selecting the CSI sequence containing the ZigBee code based on the amplitude characteristics of CSI sequence fluctuation and a limited amplitude range. Since the amplitude of the WiFi signal is increased significantly by the ZigBee signal, the CSI sequence containing the ZigBee information can be screened accurately by selecting the amplitude range.
Preferably, the method further comprises: and extracting information characteristics of the CSI sequence containing the ZigBee code based on at least one deep learning algorithm. The accuracy of information feature extraction can be accurately improved through a deep learning algorithm, so that ZigBee information can be more accurately acquired.
Preferably, the method further comprises: and classifying the information characteristics of the CSI sequence containing the ZigBee code based on at least one neural network classification model. The neural network model is used for rapidly and accurately classifying a plurality of information characteristics, so that the identification accuracy is up to 95%, and the accuracy is higher than that of equipment for performing heterogeneous communication through hardware switching in the prior art.
Preferably, the information characteristic includes at least one of variance of CSI value, time domain peak-to-peak, kurtosis, bias, standard deviation, mean, mode, median, and combinations thereof. The information characteristics are the characteristics developed after the WiFi signals are influenced by the embedding of the ZigBee signals. Therefore, the CSI sequence of the WiFi containing the ZigBee signal can be selected and obtained through a single information characteristic. Moreover, the more information features are, the more accurate the CSI sequence of the WiFi containing the ZigBee signal can be screened out.
Preferably, the method further comprises: after receiving the WiFi signal, at least one receive antenna and its CSI sequence information are selected based on the signal strength of the receive antenna. For the plurality of receiving antennas, under the influence of the transmission environment and the influence of the signal path, some receiving antennas receive more WiFi signals with large amplitude change, and some receiving antennas receive less WiFi signals with large amplitude change, so that more CSI sequences of the WiFi signals containing the ZigBee information can be obtained through selection of the receiving antennas, and the accuracy of obtaining the ZigBee information is improved.
The invention also provides a CSI-based heterogeneous communication system, which at least comprises a receiving module and an analyzing device, wherein the receiving module can receive WiFi signals, the analyzing device at least comprises a feature extraction module and a classification module, the receiving module receives the WiFi signals, and the classification module classifies the information features of the CSI sequence containing ZigBee codes extracted by the feature extraction module based on a classification model under the condition that the ZigBee signals are overlapped with the WiFi signals in a frequency domain overlapping mode and embedded into the CSI sequence. According to the communication system, the ZigBee signal and the WiFi signal can be received together in a mode that the ZigBee signal and the WiFi signal share a frequency domain, the CSI sequence containing the ZigBee information can be screened out according to the information characteristics, the ZigBee information in the CSI can be extracted accurately, and the accuracy is as high as 95%. According to the heterogeneous communication system, the receiving end only needs the WiFi receiving end and the analysis device, and extra information conversion hardware is not needed, so that more hardware maintenance is not needed. The hardware setting of the heterogeneous communication system is simpler, and the resolution accuracy is higher.
Preferably, the parsing device further comprises a preprocessing module, wherein the preprocessing module extracts amplitude characteristics of the CSI sequence, and selects the CSI sequence containing the ZigBee code based on the amplitude characteristics of CSI sequence fluctuation and a limited amplitude range. Due to the arrangement of the preprocessing module, WiFi signals which do not contain ZigBee signals can be extracted and discharged, and the data processing amount of the feature extraction module and the classification module is reduced.
Preferably, the feature extraction module extracts information features of the CSI sequence containing the ZigBee code based on at least one deep learning algorithm. The deep learning algorithm in the feature extraction module can accurately extract information features through the learning training of sample data.
The invention also provides a network framework based on the CSI heterogeneous communication system, which at least comprises a feature extraction layer and a classification layer, wherein under the condition that the ZigBee signal is overlapped with the WiFi signal in a frequency domain overlapping mode and is embedded into the CSI sequence, the classification layer classifies the information features of the CSI sequence containing the ZigBee code, which are extracted by the feature extraction layer, based on a classification model. The network framework of the invention is an LSTM network framework, and the robust information decoding is carried out by utilizing the sequence correlation of the time sequence CSI data.
Preferably, the network framework based on the CSI heterogeneous communication system further comprises a preprocessing layer, wherein the preprocessing layer extracts amplitude characteristics of the CSI sequence, and selects the CSI sequence containing the ZigBee code based on the amplitude characteristics of CSI sequence fluctuation and a limited amplitude range.
Drawings
FIG. 1 is a schematic diagram of a heterogeneous communication system architecture of the present invention;
FIG. 2 is a schematic diagram of the frequency spectrum distribution of WiFi and ZigBee;
FIG. 3 is a graphical illustration of the effect of ZigBee transmission on WiFi CSI;
FIG. 4 is a diagram of the classifier window size versus accuracy for a support vector machine;
FIG. 5 is a diagram of an LSTM network architecture;
FIG. 6 is a diagram showing the setup of the experimental apparatus of the present invention.
List of reference numerals
10: a sample data set; 20: testing the data set; 30: a receiving module; 40: a preprocessing module; 50: a feature extraction module; 60: and (5) a classification module.
Detailed Description
The following detailed description is made with reference to the accompanying drawings.
The invention provides a heterogeneous communication method and system based on CSI, which can also be a communication device or a CTC system based on machine learning and deep learning.
In the invention, the heterogeneous communication based on the CSI is a cross-protocol communication system which embeds ZigBee information into a CSI sequence of WiFi in a frequency domain overlapping mode, so that the ZigBee information can be received by receiving the CSI sequence of the WiFi.
In the prior art, the transmission of signals from ZigBee to WiFi can be realized only by a variety of hardware, and such disadvantages are that the number of hardware is large, the deployment is complex, and additional maintenance cost and traffic overhead of entering and exiting the gateway exist. How to realize direct communication between incompatible wireless devices without additional hardware is a technical problem which is not solved in the prior art.
The basic idea of the invention is that since both protocols have overlapping spectrum distributions, direct contact can be established by analyzing the coded ZigBee signal through the CSI sequence interfering WiFi caused by overlapping spectrum. The present invention proposes a new long-short time memory (LSTM) network framework that studies sequence correlation in time series CSI data to achieve robust information decoding.
WiFi is a technology that allows electronic devices to connect to WLANs based on the ieee802.11 standard. ZigBee is a low-power consumption local area network protocol technology which accords with the ieee802.15.4 standard. ZigBee and WiFi typically use 2.4G bandwidth. ZigBee is suitable for industrial application with small data volume, low power consumption and good stability. However, WiFi has a large data volume, a short distance, and is relatively unstable, and is mainly suitable for home and business applications.
The spectrum distribution of 2.4GHz ZigBee and WiFi is shown in figure 2. The spectrum distribution of ZigBee and WiFi is overlapped, so that theoretically heterogeneous communication from ZigBee to WiFi signal layer can be realized.
CSI is a very important practical data describing channels in wireless communications. In wireless communications, CSI represents the propagation characteristics of a communication link and describes the combined response of various effects in the channel, such as scattering, fading, and power attenuation. The amplitude and phase of the CSI are affected by multipath effects, including amplitude attenuation and phase shift. It describes how a signal propagates from a transmitter to a receiver, containing the amplitude and phase of each subcarrier
Wherein, H (f)i) Is a center frequency of fiAnd H is the phase. The effect of ZigBee transmission on WiFi is observed by changes in CSI.
The left diagram in fig. 3 represents the CSI sequence without ZigBee transmission. The right diagram in fig. 3 represents the CSI sequence with ZigBee transmission. As shown in fig. 3, there is ZigBee information transmission in the WiFi signal, and the CSI sequence amplitude range is 16-20. ZigBee information transmission does not exist in the WiFi signal, and the amplitude range of the CSI sequence is 9-12. The fluctuation range of the CSI sequence containing the ZigBee information is larger than that of the CSI sequence without ZigBee information transmission. Therefore, the amplitude change of the CSI can be used to determine whether there is a ZigBee information transmission.
Therefore, the CSI sequence with the amplitude range of 16-20 is selected as the CSI sequence to be processed for further extracting information characteristics.
Fig. 4 shows a fitted curve between window length and SVM classifier accuracy. The fitting function is f (x) p1x4+p2x3+p3x2+p4x+p5
Where f (x) is the classification accuracy, x is the window length, P1 ═ 0.0001, P2 ═ -0.0019, P3 ═ 0.0249$, P4 ═ -0.1541$, P5 $ 1.2557.
With the increase of the window length, the classification accuracy of the SVM classifier is improved, but the classification accuracy is hardly improved after a certain window length. Therefore, a window length of 16 is selected, and an accuracy rate of more than 94% can be achieved in the SVM classifier.
Example 1
As shown in fig. 1, the CSI-based heterogeneous communication system of the present invention includes a transmitting apparatus and a parsing apparatus.
The transmitting device comprises a WiFi transmitting terminal and a ZigBee transmitting terminal.
The WiFi sending terminal is, for example, an intel5300 network card and a data sampling module. The data sampling module is, for example, a CSITool data sampling module. The sampling rate of the CSITool data sampling module is set to 2 KHz. The WiFi packet interval is 0.5ms and the packet length is 145 bytes.
The ZigBee transmitting terminal is, for example, a TelosB node CC2420 module. The WiFi signal is transmitted on channel 6 and the ZigBee signal configuration is transmitted on channel 18 so that the WiFi signal overlaps with the ZigBee signal in the frequency domain. The ZigBee packet interval is 0.192ms, and the packet length is 28 bytes.
As shown in fig. 6, a first distance between the WiFi transmitting terminal and the ZigBee transmitting terminal is smaller than a second distance between the WiFi transmitting terminal and the WiFi receiving terminal. The advantage that so set up lies in, zigBee transmitting terminal is close to wiFi transmitting terminal, and the probability that zigBee signal and wiFi signal contact is must be great, and the zigBee signal must be able to contact and coincide with the route of part wiFi signal to transmit in the same frequency domain.
Preferably, the ZigBee transmission terminal is disposed between the WiFi transmission terminal and the WiFi reception terminal. According to the arrangement, the ZigBee transmitting terminal is located in a transmission path of the WiFi signal, so that the ZigBee signal can be contacted with the WiFi signal at any time in the same frequency domain and can be transmitted together, and the success rate of embedding the ZigBee signal into a CSI sequence of the WiFi signal can be improved.
Preferably, the first distance is one half of the second distance, which is more beneficial to acquiring more CSI sequences containing ZigBee signals, so that the accuracy of classification of the CSI creep information features is improved.
The parsing means comprises at least a receiving module 30, a pre-processing module 40, a feature extraction module 50 and a classification module 60.
The preprocessing module 40 is not necessary and may not be provided. The preprocessing module 40 is arranged to enable the quality of the filtered signals to be higher and the obtained information to be more accurate.
The receiving module 30 is used for receiving signals. The receiving module 30 includes at least one WiFi terminal and its receiving antenna.
The preprocessing module 40 is configured to preprocess the received signal sent by the receiving module 30.
The preprocessing step includes at least the selection of the receive antennas and the selection of the subcarriers.
Preferably, the preprocessing module 40 includes an antenna selection module 41 and a subcarrier selection module 42. The antenna selection module 41 is configured to select a receiving antenna. The basis for the antenna selection module 41 to select the receiving antenna includes: and selecting one antenna with the most obvious CSI sequence fluctuation through the variance of the CSI. The larger the variance, the more obvious the CSI sequence fluctuation is. According to the method and the device, the sample data sets are classified according to the information characteristics of the CSI sequences on the three receiving antennas, and the receiving antenna with the highest classification accuracy is selected.
The selection of the antennas is beneficial to selecting one receiving antenna which is most affected by ZigBee so as to improve the classification accuracy of the CSI sequence. The antenna selection module 41 sends the signals received by the selected receive antenna to the subcarrier selection module 42. Since the larger the variance of the CSI, the more obvious the amplitude fluctuation of the CSI sequence, the easier it is to extract the characteristics of the information. The subcarrier selection block 42 selects the carrier with the largest variance as the subcarrier from which information is extracted based on the variance value of the CSI values.
In the present invention, each receive antenna has 30 subcarriers. The receiving antenna is selected first, and then the sub-carrier is selected, so that a CSI sequence with large amplitude fluctuation and obvious amplitude fluctuation can be obtained.
Each subcarrier of the WiFi signal will be affected differently by the ZigBee signal. And the subcarrier with the largest influence by the ZigBee signal is selected, so that better classification accuracy can be obtained.
The feature extraction module 50 is configured to perform information feature extraction on the subcarrier information. The information characteristic includes at least a variance of the CSI value. The information features may further include one or more of time-domain peak-to-peak, kurtosis, bias, standard deviation, mean, mode, and median.
The information characteristics represent different changes in the WiFi-CSI when ZigBee transmissions occur.
The classification module 60 is configured to classify information features such as time domain peak-to-peak, kurtosis, bias, standard deviation, mean, mode, median, and the like.
The classification module 60 includes a learning classifier 61 and an LSTM neural network module 62.
The learning classifier 61 automatically analyzes and acquires rules from the data based on a machine learning algorithm, and predicts unknown data using the rules. Examples of machine learning algorithms include decision trees, discriminant analysis, logistic regression, SVM, KNN, and component classifiers.
The purpose of class learning is to learn a classification function or classification model from a given set of manually labeled class training samples. When new data is entered, classification module 60 may map the new data item to a class within a given class according to a functional prediction.
An LSTM neural network is disposed within the LSTM neural network module 62. As shown in fig. 5, the LSTM neural network is a Recurrent Neural Network (RNN) that is capable of learning long-term correlations between time steps of sequence data. In the training of the original RNN, with the increase of training time and the increase of the number of network layers, the problem of gradient explosion or gradient disappearance is likely to occur, so that the remote data cannot be processed and the information of the remote data cannot be acquired. Applications of the LSTM neural network include text generation, sequence classification, speech recognition, image description generation, and videotagging.
Preferably, the classification module 60 performs learning classification training on the preprocessed sample data set to construct an LSTM classification model.
For example, in a data sample, the CSI sequence is labeled as "0" or "1". Only the CSI sequence for WiFi transmissions (which do not overlap with ZigBee) is marked as "0". The CSI sequence from overlapping transmissions of WiFi and ZigBee is marked as "1".
After constructing the LSTM classification model, CSI sequence classification is performed on the received data.
The invention can be widely applied.
For example, existing sensors are commonly referred to as ZigBee sensors, which collect information such as temperature, humidity, air quality, etc. The smart phone can only collect WiFi signals generally. The ZigBee sensor cannot directly send the ZigBee signal to the mobile phone. According to the method and the communication system, the ZigBee signal and the WiFi signal sent by the ZigBee sensor are sent in the same frequency domain, so that the ZigBee signal and the WiFi signal are received by the WiFi receiving module of the mobile phone in the same frequency domain, and the ZigBee data packet can be analyzed. The invention enables the ZigBee sensor and the WiFi receiving terminal to realize direct communication so as to transmit information.
Example 2
This embodiment is a further improvement of embodiment 1, and repeated contents are not described again.
FIG. 6 is one of the settings application scenarios of the present invention. The WiFi terminal is an Intel link 5300WiFi network card. The Intel link 5300WiFi network card is equipped with a 2.16GHz Intel Celeron n2830 processor, a 2GB memory and a 12.04 version Ubuntu operating system.
The ZigBee terminal uses the IOT-NODE24T NODE. The IOT-NODE24T NODE is an ultra-low power consumption wireless sensor network NODE, conforms to IEEE802.15.4 protocol specification and uses a PCB antenna. The node adopts an MSP430F1611 processor of TI company, has the working frequency of 8MHz, and is integrated with a wireless transceiver chip CC2420 conforming to the IEEE802.15.4 protocol specification. The IOT-NODE24T NODE is suitable for research and development of ZigBee protocols.
The invention uses TinyOS to control the working power, communication channel and data packet length of the ZigBee node. TinyOS is an open source operating system developed by University of California at Berkeley (University of California, Berkeley). TinyOS is designed specifically for embedded wireless sensor networks. The component-based operating system architecture enables programs to be updated quickly while reducing the code length of the memory constraints of the sensor network nodes.
The antenna of the receiving module is as shown in fig. 6. The distance between the WiFi sending terminal and the WiFi receiving terminal is 100 cm. The distance between the WiFi sending terminal and the ZigBee sending terminal is 50 cm.
The WiFi transmitting terminal is set to transmit on channel 6. The ZigBee terminal is arranged to transmit on the channel 18 so that the two signals overlap in the frequency domain.
Preferably, the WiFi sending terminal starts to transmit signals after the ZigBee sending terminal starts, so that data packets of the ZigBee signals can be prevented from being omitted, and the ZigBee signals and the WiFi signals can be transmitted in the same frequency domain.
The relative arrangement positions of the ZigBee transmitting terminal, the WiFi transmitting terminal and the WiFi receiving terminal are not limited, and the ZigBee transmitting terminal, the WiFi transmitting terminal and the WiFi receiving terminal can be any arrangement positions.
Preferably, as shown in fig. 6, the ZigBee transmitting terminal is arranged on the same straight line with the WiFi transmitting terminal and at least one receiving antenna of the WiFi receiving terminal, and the ZigBee transmitting terminal is arranged in the middle of the three. By the arrangement, data with large WiFi CSI sequence change can be collected, and machine learning classification and neural network classification can be performed.
The sampling rate of CSITool is set to 2 KHz. The WiFi packet interval is 0.5ms and the packet length is 145 bytes. The ZigBee packet interval is 0.192ms, and the packet length is 28 bytes. So set up, can guarantee that a zigBee data packet overlaps with at least one wiFi data packet.
The feature extraction module adopts 23 machine learning classification methods to extract 8 features from the data. Two features are peak-to-peak variance; the three features are variance, peak-to-peak and kurtosis; the four characteristics are variance, peak-to-peak, kurtosis and skewness; the five characteristics are variance, peak-to-peak, kurtosis, skewness and standard deviation; the six characteristics are variance, peak-to-peak, kurtosis, skewness, standard deviation and mean value; seven characteristics are variance, peak-to-peak, kurtosis, skewness, standard deviation and mean; eight features refer to variance, peak-to-peak, kurtosis, skewness, standard deviation, mean, mode, and medium.
In implementation, as the information features increase, the recognition accuracy of most of the learning classifiers 61 is improved correspondingly.
Because some classifiers have different sensitivities to features of complex trees, intermediate trees, sample trees, and so on. Preferably, the recognition features of the present invention incorporate standard deviation features, and the accuracy of the learning classifier 61 is greatly improved. In the case where the learning classifier is a linear discrimination classifier or a quadratic discrimination classifier, increasing or decreasing the information features does not affect the accuracy of the classifier.
The Standard Deviation (Standard development) is the arithmetic square root of the arithmetic mean (i.e., the variance) of the squared Deviation, and is expressed as sigma. The standard deviation is also called standard deviation, or experimental standard deviation, and is most commonly used in probability statistics as a measure of the degree of statistical distribution. The standard deviation is calculated as follows:
Preferably, when the learning classifier 61 is an SVM classifier, the accuracy of information feature classification is high.
Preferably, when the LSTM neural network is trained on a sample dataset, the accuracy of the sample dataset is 96.3%. The accuracy of the test data set of the present invention was 94.2%. The classification precision of the LSTM neural network is high.
It should be noted that the above-mentioned embodiments are exemplary, and that those skilled in the art, having benefit of the present disclosure, may devise various arrangements that are within the scope of the present disclosure and that fall within the scope of the invention. It should be understood by those skilled in the art that the present specification and figures are illustrative only and are not limiting upon the claims. The scope of the invention is defined by the claims and their equivalents.
The present specification encompasses multiple inventive concepts and the applicant reserves the right to submit divisional applications according to each inventive concept. The present description contains several inventive concepts, such as "preferably", "according to a preferred embodiment" or "optionally", each indicating that the respective paragraph discloses a separate concept, the applicant reserves the right to submit divisional applications according to each inventive concept.
Claims (10)
1. A CSI-based heterogeneous communication method, the method comprising at least:
under the condition that the ZigBee signal is overlapped with the WiFi signal in a frequency domain overlapping mode and embedded with the CSI sequence, the WiFi signal is received, and information characteristics of the CSI sequence containing the ZigBee code are classified based on a classification model.
2. The CSI-based heterogeneous communication method of claim 1, wherein the method further comprises:
and extracting amplitude characteristics of the CSI sequence, and selecting the CSI sequence containing the ZigBee code based on the amplitude characteristics of CSI sequence fluctuation and a limited amplitude range.
3. The CSI-based heterogeneous communication method according to claim 2, wherein the method further comprises:
and extracting information characteristics of the CSI sequence containing the ZigBee code based on at least one deep learning algorithm.
4. The CSI-based heterogeneous communication method of claim 1, wherein the method further comprises:
and classifying the information characteristics of the CSI sequence containing the ZigBee code based on at least one neural network classification model.
5. The CSI-based heterogeneous communication method according to any one of claims 1 to 4,
the information characteristics at least comprise one of variance, time domain peak-to-peak, kurtosis, bias, standard deviation, mean, mode, median of CSI values and combination thereof.
6. The method of CSI based heterogeneous communication according to claim 5, wherein said method further comprises:
after receiving the WiFi signal, at least one receive antenna and its CSI sequence information are selected based on the signal strength of the receive antenna.
7. A CSI-based heterogeneous communication system, the system being capable of receiving at least a reception module (30) and an analysis device of WiFi signals,
said parsing means comprising at least a feature extraction module (50) and a classification module (60),
the receiving module (30) receives a WiFi signal,
in case the ZigBee signal overlaps with the WiFi signal in a frequency domain overlapping manner and embeds a CSI sequence, the classification module (60) classifies information features extracted by the feature extraction module (50) containing the ZigBee-encoded CSI sequence based on a classification model.
8. The CSI-based heterogeneous communication system according to claim 7, wherein the parsing means further comprises a pre-processing module (40),
the preprocessing module (40) extracts amplitude characteristics of the CSI sequence, and selects the CSI sequence containing the ZigBee code based on the amplitude characteristics of CSI sequence fluctuation and a limited amplitude range.
9. The CSI-based heterogeneous communication system according to claim 7, wherein the feature extraction module extracts information features containing ZigBee-coded CSI sequences based on at least one deep learning algorithm.
10. A network framework based on a CSI heterogeneous communication system is characterized by at least comprising a feature extraction layer and a classification layer,
under the condition that the ZigBee signal is overlapped with the WiFi signal in a frequency domain overlapping mode and embedded with the CSI sequence, the classification layer classifies the information characteristics extracted by the characteristic extraction layer and containing the ZigBee coded CSI sequence based on a classification model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110554717.0A CN113286309B (en) | 2021-05-18 | 2021-05-18 | Heterogeneous communication method and system based on CSI |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110554717.0A CN113286309B (en) | 2021-05-18 | 2021-05-18 | Heterogeneous communication method and system based on CSI |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113286309A true CN113286309A (en) | 2021-08-20 |
| CN113286309B CN113286309B (en) | 2023-02-07 |
Family
ID=77280446
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110554717.0A Active CN113286309B (en) | 2021-05-18 | 2021-05-18 | Heterogeneous communication method and system based on CSI |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113286309B (en) |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107682830A (en) * | 2017-09-27 | 2018-02-09 | 清华大学 | A kind of ZigBee to WiFi communication means and device |
| CN107750061A (en) * | 2017-09-27 | 2018-03-02 | 清华大学 | A kind of ZigBee to WiFi adaptive communication method and device |
| CN108763216A (en) * | 2018-06-01 | 2018-11-06 | 河南理工大学 | A kind of text emotion analysis method based on Chinese data collection |
| US20190097865A1 (en) * | 2012-12-05 | 2019-03-28 | Origin Wireless, Inc. | Apparatus, systems and methods for event recognition based on a wireless signal |
| CN109548046A (en) * | 2018-11-16 | 2019-03-29 | 清华大学 | A kind of channel self-adapting is across protocol communication system and method |
| US20190158340A1 (en) * | 2012-12-05 | 2019-05-23 | Origin Wireless, Inc. | Apparatus, systems and methods for fall-down detection based on a wireless signal |
| CN110414468A (en) * | 2019-08-05 | 2019-11-05 | 合肥工业大学 | Based on the auth method of hand signal under WiFi environment |
| CN110659677A (en) * | 2019-09-10 | 2020-01-07 | 电子科技大学 | Human body falling detection method based on movable sensor combination equipment |
| CN110826327A (en) * | 2019-11-05 | 2020-02-21 | 泰康保险集团股份有限公司 | Emotion analysis method and device, computer readable medium and electronic equipment |
| CN111053549A (en) * | 2019-12-23 | 2020-04-24 | 威海北洋电气集团股份有限公司 | Intelligent biological signal abnormality detection method and system |
| US20200177418A1 (en) * | 2017-06-19 | 2020-06-04 | Nokia Technologies Oy | Data transmission network configuration |
| CN111327379A (en) * | 2020-02-03 | 2020-06-23 | 清华大学 | Channel allocation method and device based on cross-protocol signals |
| CN111325112A (en) * | 2020-01-31 | 2020-06-23 | 贵州大学 | Cutter wear state monitoring method based on depth gate control circulation unit neural network |
| CN111835492A (en) * | 2020-06-09 | 2020-10-27 | 北京邮电大学 | Asymmetric channel-oriented symmetric cross-protocol communication method and device |
| CN112637093A (en) * | 2020-12-09 | 2021-04-09 | 齐鲁工业大学 | Signal detection method based on model-driven deep learning |
-
2021
- 2021-05-18 CN CN202110554717.0A patent/CN113286309B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190097865A1 (en) * | 2012-12-05 | 2019-03-28 | Origin Wireless, Inc. | Apparatus, systems and methods for event recognition based on a wireless signal |
| US20190158340A1 (en) * | 2012-12-05 | 2019-05-23 | Origin Wireless, Inc. | Apparatus, systems and methods for fall-down detection based on a wireless signal |
| US20200177418A1 (en) * | 2017-06-19 | 2020-06-04 | Nokia Technologies Oy | Data transmission network configuration |
| CN107750061A (en) * | 2017-09-27 | 2018-03-02 | 清华大学 | A kind of ZigBee to WiFi adaptive communication method and device |
| CN107682830A (en) * | 2017-09-27 | 2018-02-09 | 清华大学 | A kind of ZigBee to WiFi communication means and device |
| CN108763216A (en) * | 2018-06-01 | 2018-11-06 | 河南理工大学 | A kind of text emotion analysis method based on Chinese data collection |
| CN109548046A (en) * | 2018-11-16 | 2019-03-29 | 清华大学 | A kind of channel self-adapting is across protocol communication system and method |
| CN110414468A (en) * | 2019-08-05 | 2019-11-05 | 合肥工业大学 | Based on the auth method of hand signal under WiFi environment |
| CN110659677A (en) * | 2019-09-10 | 2020-01-07 | 电子科技大学 | Human body falling detection method based on movable sensor combination equipment |
| CN110826327A (en) * | 2019-11-05 | 2020-02-21 | 泰康保险集团股份有限公司 | Emotion analysis method and device, computer readable medium and electronic equipment |
| CN111053549A (en) * | 2019-12-23 | 2020-04-24 | 威海北洋电气集团股份有限公司 | Intelligent biological signal abnormality detection method and system |
| CN111325112A (en) * | 2020-01-31 | 2020-06-23 | 贵州大学 | Cutter wear state monitoring method based on depth gate control circulation unit neural network |
| CN111327379A (en) * | 2020-02-03 | 2020-06-23 | 清华大学 | Channel allocation method and device based on cross-protocol signals |
| CN111835492A (en) * | 2020-06-09 | 2020-10-27 | 北京邮电大学 | Asymmetric channel-oriented symmetric cross-protocol communication method and device |
| CN112637093A (en) * | 2020-12-09 | 2021-04-09 | 齐鲁工业大学 | Signal detection method based on model-driven deep learning |
Non-Patent Citations (3)
| Title |
|---|
| XIUZHEN GUO等: "ZigFi: Harnessing Channel State Information for Cross-Technology Communication", 《IEEE/ACM TRANSACTIONS ON NETWORKING》 * |
| 周健: "无线感知网络中基于CSI的室内入侵检测与行为识别研究", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
| 陈慕涵等: "基于深度学习的大规模MIMO信道状态信息反馈", 《物联网学报》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113286309B (en) | 2023-02-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Jiang et al. | Transparent cross-technology communication over data traffic | |
| EP2229792B1 (en) | Using information obtained from one wireless medium for communication in another wireless medium | |
| Maglogiannis et al. | Enhancing the coexistence of LTE and Wi-Fi in unlicensed spectrum through convolutional neural networks | |
| Ahmed et al. | CR-IoTNet: Machine learning based joint spectrum sensing and allocation for cognitive radio enabled IoT cellular networks | |
| CN107750061B (en) | Self-adaptive communication method and device from ZigBee to WiFi | |
| CN106533623B (en) | Across protocol communication method and device | |
| Subray et al. | Towards enhancing spectrum sensing: Signal classification using autoencoders | |
| Chan et al. | DeepSense: Enabling carrier sense in low-power wide area networks using deep learning | |
| WO2022047410A2 (en) | Unsupervised learning for simultaneous localization and mapping in deep neural networks using channel state information | |
| Yi et al. | Interference source identification for ieee 802.15. 4 wireless sensor networks using deep learning | |
| CN109450573B (en) | frequency spectrum sensing method based on deep neural network | |
| US11863495B2 (en) | Signaling for a channel state information reference signal (CSI-RS) | |
| CN113922899A (en) | Spectrum sensing access method based on neural network and Welch transformation | |
| Camelo et al. | A general approach for traffic classification in wireless networks using deep learning | |
| Pattanayak et al. | Artificial intelligence based model for channel status prediction: A new spectrum sensing technique for cognitive radio | |
| Mirza et al. | Fingerprinting 802.11 rate adaption algorithms | |
| CN113286309B (en) | Heterogeneous communication method and system based on CSI | |
| CN101444129A (en) | Method and apparatus for communication system | |
| CN117713967A (en) | Intelligent frequency spectrum sensing anti-interference communication system for ad hoc network unmanned aerial vehicle | |
| Huang et al. | Semi-Supervised Deep Learning Based Wireless Interference Identification for IIoT Networks | |
| Kim et al. | Implementation of bistatic backscatter wireless communication system using ambient Wi-Fi signals | |
| CN115002703A (en) | Passive indoor people number detection method based on Wi-Fi channel state information | |
| Ferreira et al. | Improving mmwave backhaul reliability: a machine-learning based approach | |
| Kapetanovic et al. | Classifying wlan packets from the rf envelope: Towards more efficient wireless network performance | |
| Almohamad et al. | A deep learning model for LoRa signals classification using cyclostationay features |
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 |