WO2022036801A1

WO2022036801A1 - Method and system for achieving coexistence of heterogeneous networks

Info

Publication number: WO2022036801A1
Application number: PCT/CN2020/116547
Authority: WO
Inventors: 姚俊梅; 郑玮东; 伍楷舜
Original assignee: 深圳大学
Priority date: 2020-08-18
Filing date: 2020-09-21
Publication date: 2022-02-24
Also published as: CN111935721B; CN111935721A

Abstract

Disclosed in the present invention is a method and system for achieving coexistence of heterogeneous networks. The method comprises: on the basis of an acquired environmental sample, a WiFi transmitting terminal determining a channel state by using a set first idle channel detection threshold; when it is determined that a channel is in an idle state, inputting, within a set distributed inter-frame spacing time, the environmental sample into a pre-trained neural network model to identify whether a heterogeneous signal exists; and when the existence of a heterogeneous signal is identified, the WiFi transmitting terminal reducing the first idle channel detection threshold to a second idle channel detection threshold, terminating the distributed inter-frame spacing time, entering a backoff process, and detecting, in a backoff time slot, the channel state by using the second idle channel detection threshold, thereby determining a transmission opportunity of a WiFi data packet according to a detection result. In the present invention, a heterogeneous signal can be accurately identified, thus reducing the interference between heterogeneous networks.

Description

A method and system for realizing coexistence of heterogeneous networks

technical field

The present invention relates to the technical field of heterogeneous network communication, and more particularly, to a method and system for realizing coexistence of heterogeneous networks.

Background technique

The vigorous development of the Internet of Things technology has brought about the explosive growth of wireless communication devices. Many devices use various communication protocols with different performances due to different needs. Wireless communication between devices using different protocols in the same frequency band will lead to cross-protocol interference. . For example, in the most common ISM2.4GHZ frequency band, WiFi devices cannot perceive ZigBee signals in the same frequency band with low transmit power due to factors such as large occupied bandwidth, high transmit power, and long transmission range. ZigBee receiving nodes collide, which greatly interferes with the information transmission of ZigBee nodes.

In recent years, various wireless networks are expanding their application fields, resulting in the coexistence of multiple wireless protocols in the same space. As a result, cross-protocol interference has become more and more serious. The problem of coexistence of heterogeneous protocols has become a bottleneck restricting the development of wireless networks. At present, alleviating cross-protocol interference and solving the problem of heterogeneous signal coexistence mainly start from two aspects: heterogeneous signal identification and coexistence mechanism design.

Without effective detection and identification of heterogeneous sources of interference, countermeasures cannot be taken to address coexistence issues. Traditional methods identify heterogeneous signals based on spectral characteristics at high sampling rates through custom-built spectrum analyzers. In recent years, researchers have realized the identification of interference sources of common communication chips through the unique time-frequency characteristics of heterogeneous interference. For example, using the periodicity of WiFiBeacon to identify WiFi interference on the ZigBee chip, or using decision tree learning to identify multiple heterogeneous signals through the RSSI value in the WiFi chip.

After learning the source of interference, designing a reasonable mechanism to allow heterogeneous protocols to use the channel harmoniously becomes a top priority. Common approaches include detecting various congestion levels through cognitive radio or interference detection mechanisms to select an appropriate channel. In recent years, researchers have proposed to improve the visibility of ZigBee to avoid the WiFi transmitter, for example, placing additional ZigBee nodes near the WiFi transmitter to send redundant data; another example, adding ZigBee Preamble (preamble) to reserve channels for remote ends Another example is to use a network server equipped with cross-protocol communication technology to achieve channel coordination and planning; or, to use the sub-carrier zeroing technology to reserve some sub-carriers that overlap with the ZigBee channel through the WiFi transmitter. Frequency Domain Isolation of Heterogeneous Signals.

The existing heterogeneous signal identification schemes are all based on artificially extracted time-frequency features for heterogeneous signal identification. The biggest drawback of this method is that the required sampling time is long, that is, better identification results can be obtained only when the spectral resolution is high. However, in the existing coexistence mechanism design, the solution of switching channels becomes inadvisable due to the increased number of devices, which leads to more and more busy channels. Adding nodes to improve ZigBee visibility will bring extra cost and little effect. Cross-protocol communication technology transmission The information is inefficient and the physical layer modification process is cumbersome and complicated. The subcarrier zeroing technology needs to be modified on both the WiFi transmitter and the receiver. Modifying the WiFi receiving device is trivial and unrealistic.

SUMMARY OF THE INVENTION

The purpose of the present invention is to overcome the above-mentioned defects of the prior art, and to provide a method and system for realizing the coexistence of heterogeneous networks, so as to solve the heterogeneous network in which the WiFi equipment causes serious interference to the ZigBee nodes of the overlapping channels caused by the unequal transmission power. coexistence problem.

According to a first aspect of the present invention, a method for realizing coexistence of heterogeneous networks is provided. The method includes the following steps:

Step S1: The WiFi sending end uses the set first idle channel detection threshold to determine the channel state based on the collected environment samples, and in the case of determining that the channel is in the idle state;

Step S2: in the case of judging that the channel is in an idle state, within the set distributed inter-frame gap time, input the environment sample into a pre-trained neural network model to identify whether there is a heterogeneous signal;

Step S3: In the case of identifying the existence of heterogeneous signals, the WiFi transmitter reduces the first idle channel detection threshold to the second idle channel detection threshold, ends the distributed inter-frame gap time, enters the backoff process, and is in the backoff process. The time slot uses the second idle channel detection threshold to detect the channel state, and then determines the transmission timing of the WiFi data packet according to the detection result.

According to a second aspect of the present invention, a system for realizing coexistence of heterogeneous networks is provided. The system includes:

Channel state detection unit: the WiFi sending end uses the set first idle channel detection threshold to determine the channel state based on the collected environment samples;

Heterogeneous signal identification unit: in the case of judging that the channel is in an idle state, within the set distributed inter-frame gap time, the environment sample is input into the pre-trained neural network model to identify whether there is a heterogeneous signal;

Coexistence protocol unit: In the case of identifying the presence of heterogeneous signals, the WiFi transmitter reduces the first idle channel detection threshold to the second idle channel detection threshold, ends the distributed inter-frame gap time, enters the backoff process, and The backoff time slot uses the second idle channel detection threshold to detect the channel state, and then determines the transmission timing of the WiFi data packet according to the detection result.

Compared with the prior art, the advantage of the present invention is that the provided heterogeneous signal identification design can automatically extract the ZigBee signal with a low signal-to-noise ratio through a deep learning model, so as to achieve excellent performance even under the microsecond-level delay. Recognition rate. The coexistence mechanism design only needs to modify the medium access control layer at the WiFi sending end, and the rest are compatible with the existing network, which is more convenient for deployment and large-scale promotion. The heterogeneous signal identification module is designed in combination with the coexistence mechanism, which can realize the existence of low-power ZigBee signals at the WiFi transmitter with a very short delay, so as to prevent the WiFi transmitter from simultaneously transmitting WiFi data packets during ZigBee data transmission. The invention starts from the interference source, solves the coexistence problem of heterogeneous networks with unequal power, and realizes a more fair and reasonable channel competition situation.

Other features and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention with reference to the accompanying drawings.

Description of drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention.

1 is a flowchart of a method for realizing coexistence of heterogeneous networks according to an embodiment of the present invention;

2 is an overall process diagram of a method for realizing coexistence of heterogeneous networks according to an embodiment of the present invention;

3 is a flowchart of identifying a heterogeneous signal according to an embodiment of the present invention;

4 is a topology diagram of an experimental device for collecting sample data according to an embodiment of the present invention;

5 is a schematic diagram of a process for identifying heterogeneous signals according to an embodiment of the present invention;

6 is a structural diagram of a convolutional neural network according to an embodiment of the present invention;

7 is a schematic diagram of a coexistence mechanism in a heterogeneous network according to an embodiment of the present invention;

8 is a schematic diagram of an experimental result of identifying a heterogeneous signal according to an embodiment of the present invention;

9 is a topology diagram of a protocol performance verification simulation according to an embodiment of the present invention;

FIG. 10 is a schematic diagram of protocol performance of a WiFi transmitter using different parameters according to an embodiment of the present invention;

FIG. 11 is a schematic diagram of protocol performance for different topological distances according to an embodiment of the present invention.

detailed description

Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, the numerical expressions and numerical values set forth in these embodiments do not limit the scope of the invention unless specifically stated otherwise.

The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such techniques, methods, and apparatus should be considered part of the specification.

In all examples shown and discussed herein, any specific values should be construed as illustrative only and not limiting. Accordingly, other instances of the exemplary embodiment may have different values.

It should be noted that like numerals and letters refer to like items in the following figures, so once an item is defined in one figure, it does not require further discussion in subsequent figures.

Referring to Fig. 1 , the method for realizing coexistence of heterogeneous networks provided by the present invention generally includes: step S110, the WiFi sending end uses the set first idle channel detection threshold based on the collected environment samples to determine the channel state; step S120, in In the case of judging that the channel is in an idle state, within the set distributed inter-frame gap time, input the environment sample into the pre-trained neural network model to identify whether there is a heterogeneous signal; step S130, after identifying the existence of a heterogeneous signal In the case of , the WiFi transmitter reduces the first idle channel detection threshold to the second idle channel detection threshold, ends the distributed inter-frame gap time, enters the back-off process, and uses the second idle channel detection threshold in the back-off time slot The channel state is detected, and then the transmission timing of the WiFi data packet is determined according to the detection result. The present invention mainly includes the design of heterologous signal recognition and the design of heterologous signal coexistence mechanism.

In the following, the heterogeneous signal identification and coexistence mechanism (or called coexistence protocol) will be introduced in detail to facilitate understanding. Taking a heterogeneous network composed of WiFi and ZigBee as an example, it will be described in conjunction with FIG. 2 .

1. About the identification of heterogeneous signals

For the identification of heterogeneous signals, the present invention adopts deep learning technology to realize the identification of ZigBee signals when the signal-to-noise ratio is low. In general, the process of heterogeneous signal identification includes: collecting a large amount of data to construct the original data set; after labeling the original data set according to the frequency domain features of higher spectral resolution, and then dividing into frames to obtain fewer sample points and Corresponding labels, obtain the training set; send the training set to CNN (Convolutional Neural Network, convolutional neural network) for training. In addition, the robustness of the model can be further verified by the test set, so as to realize the effective identification of ZigBee signals with low signal-to-noise ratio in extremely short delay.

Specifically, with reference to Figure 3, Figure 4 and Figure 5, constructing a training set includes the following steps:

Step S101, collecting signals.

To ensure the purity of the original signal, signal acquisition was performed in a shielded room. The distance and transmission parameters of the experimental equipment are set as shown in Figure 4. Two general software radio equipment USRPN210 is used as the WiFi protocol transceiver. The WiFi transmitter is in WiFi channel 6, with a standard transmit power of 16dBm and a rate of 54Mbps. Continuously transmit up to 1500 byte packets. The TelosB device is used as the ZigBee protocol transmission node, and the ZigBee sending node sends 20-byte data packets every 1ms with the ZigBee channel 16 overlapping with the WiFi channel 6, the standard transmission power of the ZigBee device of 0dBm, and the data transmission rate of 250kbps.

In the acquisition experiment, first only run the WiFi receiving device and ZigBee transceiver device, collect at 25Mbps sampling rate for 10s, and construct a pure ZigBee signal dataset 1; then only run the WiFi transceiver device, collect at 25Mbps for 10s, construct a pure WiFi signal dataset 2; Then, run the WiFi transceiver and ZigBee transceiver at the same time, and collect 10s data at a sampling rate of 25Mbps to construct dataset 3. Since there is a gap between data packets sending and receiving, the ZigBee signal in data set 1 accounts for about 60%, and the rest is the noise floor. Data set 2 is used to improve the robustness of the model, and there is no need to care about the proportion of WiFi signals. The signal accounts for about 7%, and the rest is the noise floor and WiFi signal. Data set 1 and data set 2 constitute a training set, and data set 3 is a test set.

Step S102, divide long frames and acquire labels.

The original sample data streams of all data sets are divided into long frames at every 1024 sample points, and Hanning window is applied frame by frame to reduce the interference caused by signal fluctuations, and then each long frame is subjected to fast Fourier transform to obtain spectral images ,

Further, two spectral features, the signal bandwidth and the center frequency, are extracted from the spectrogram to identify the existence of the ZigBee signal. In one embodiment, a peak finding method is used to find spectral features. Specifically, the peak search method compares the change of the slope of the line connecting two adjacent points. If the adjacent slope rises by more than 30%, it is regarded as a peak. A ZigBee signal spectrogram with a spectral resolution of 1024 points has several peaks. , the center of the peak is the center frequency, and the width of the peak boundary is the bandwidth. The presence or absence of a ZigBee signal in the frame can be easily identified using the bandwidth and center frequency at high spectral resolution. For example, if the signal bandwidth is 2MHz and its center frequency falls near 2.435GHz, it can be determined that there is a ZigBee signal in this data, and the label is set to ZigBee (indicating that there is a heterogeneous signal ZigBee), otherwise, whether it is noise or WiFi signal , are marked with the Noise label.

Step S103, obtaining short frames and corresponding labels

In step S102, the long frame divided by 1024 sample points has higher spectral resolution. In order to meet the requirement of shorter time delay, the method of truncating the sample points is adopted to obtain a short frame of every 64 points. . At this time, the spectral characteristics of the spectrogram of the short frame are no longer obvious, so the means of extracting the spectral bandwidth and the center frequency are no longer feasible.

In one embodiment, the label of the short frame is obtained by means of the label of the long frame. For example, when a long frame of 1024 sample points is subdivided into a short frame of 64 sample points, the corresponding label of the short frame can be obtained only by expanding the label by 16 times. Since a ZigBee data packet continuously occupies about 36 frames divided by 1024 points, the signal boundary frame is at the junction of ZigBee signal and noise, which may contain both ZigBee signal and noise signal. Direct expansion will cause wrong labels. In order to avoid confusion and ensure label accuracy, preferably, the signal boundary long frame and the corresponding label are removed, and then the short frame operation is performed.

Step S104, data vectorization

In a preferred embodiment, for the short frame obtained in step S103, a vectorization operation is performed to convert the short frame of IQ data into a four-dimensional vector form N _example × Dim _IQ ×Dim _value ×Dim _channel which is convenient for CNN to read and use. Nexample represents the number of short frames, Dim _IQ = 2 stores the data of the two IQ channels respectively, Dim _value = 64 means that each short frame contains only 64 sample points, and Dim _channel = 1 is the value representing black and white images in image processing . For example, after using the numpy library to vectorize the short frame data stream, the cPickle library is used to serialize and store it into a file for the CNN model to use.

For the specific network structure of the CNN model, see Figure 6. The short frame vector and corresponding label (ie, the vectorized sample data set) obtained through the above data processing process will be sent to the CNN model for training. Using CNN can find the hidden features of weak ZigBee signals and the recognition speed is fast. Referring to Figure 6, the first layer of the network is a zero-padding layer (Zero Padding) to save boundary information; then, 128 convolutions with a convolution kernel size of (1,3) and a stride of (1,1) are input The convolution layer composed of filters, and the ReLU (linear rectification function) function is used as the activation function to introduce nonlinearity. This convolution layer is the key to extracting the hidden features of the signal; then, Dropout is performed to randomly discard half of the neurons to avoid overfitting. , which is beneficial to improve the robustness of the model; perform the same zero-padding, convolution, and dropout operations to obtain deeper and more representative hidden features; then, send the multi-dimensional data to the Flatten (flattening) layer One-dimensionalization is convenient for subsequent fully-connected layers; then two fully-connected layers are set as output layers, in order to obtain the weighted sum of the high-dimensional features obtained by the convolutional layer to obtain the scores of each category, and then map them to probabilities through the Softmax function. Finally, a CNN model that can effectively identify heterologous signals such as ZigBee with very few sample points in a short time is obtained.

2. Design of Coexistence Mechanism in Heterogeneous Networks

Before WiFi sends data packets, a small number of environmental signal samples will be sampled for ED (EnergyDetection, energy detection) to determine whether the channel is idle. In the embodiment of the present invention, the environmental signal sampling points are sent into the deep learning model to determine the existence of the ZigBee signal. If it does not exist, it will be business as usual. If it exists, the CNN classifier will set a ZigBeeOn flag and pass it to the MAC. (Medium Access Control, medium access control) layer, the MAC layer will lower the CCA (Clear Channel Assessment, idle channel detection) threshold, so that it will be more sensitive to ZigBee packets with low signal energy, so the energy detection will determine that the channel is busy, Trigger the backoff mechanism, delay the current transmission, and reserve the channel for ZigBee signal transmission, so as to avoid the phenomenon of the ZigBee data packet being destroyed caused by the simultaneous transmission of two protocols, and realize the fair competition of the channel between the two signals with unequal power. The lowering degree of the threshold may be determined according to the actual application scenario of the network, or determined according to simulation, which is not limited in the present invention.

After the CNN model identifies that the weak ZigBee signal is being transmitted in the channel, the next step is to design a reasonable MAC layer mechanism to delay the transmission of the WiFi signal to prevent its access channel from interfering with the normal reception of ZigBee data packets. Since WiFi devices cannot wait all the time, they must find an opportunity to re-access the channel to achieve a fair and reasonable channel allocation scheme for heterogeneous protocols.

The coexistence mechanism designed by the present invention is shown in Figure 7. By modifying the parameters of the inherent DCF (Distributed Coordination Function) mechanism of WiFi, that is, reducing the CCA threshold, the WiFi transmitter is more sensitive to the weak ZigBee signal, so as to determine the channel Busy and triggers its inherent back-off mechanism, delaying the transmission of WiFi packets and giving up the channel for ZigBee data transmission.

Specifically, when the WiFi sender has a data packet to transmit, it first samples a small amount of environmental samples for energy detection to determine the channel state. At this time, since the ZigBee transmission power is lower than the energy detection threshold, the WiFi will determine that the channel is idle and wait for the protocol 802.11 DIFS (Distributed Inter-frame Spacing) of 9 microseconds specified by g/n. In the embodiment of the present invention, a small number of environmental samples used for energy detection are sent into the pre-trained CNN model during the DIFS time to determine the existence of the ZigBee signal, and if there is, a ZigBeeOn flag is set. The WiFiMAC layer will reduce the CCA threshold after receiving the flag bit. At this time, the sender ends the DIFS and enters the backoff process, and randomly selects a value from the contention window as the value of the random backoff counter. The backoff time consists of several backoff time slots. CCA will be performed in each backoff time slot. Since the CCA threshold has been lowered, the presence of the ZigBee signal will be sensed, and the channel will be determined to be busy, thus suspending the random backoff counter and delaying the transmission of WiFi data packets. In this way, there is no scenario in which the WiFi signal and the ZigBee signal exist in the channel at the same time, thus solving the problem that the WiFi data packet destroys the ZigBee data packet at the ZigBee receiving end.

Correspondingly, the present invention also provides a system for realizing coexistence of heterogeneous networks, which is used to realize one aspect or multiple aspects of the above method. For example, the system includes: a channel state detection unit, where the WiFi sending end uses a set first idle channel detection threshold to determine the channel state based on the collected environmental samples; a heterogeneous signal identification unit: when it is determined that the channel is in an idle state, the In the set distributed inter-frame gap time, input the environment sample into the pre-trained neural network model to identify whether there is a heterogeneous signal; coexistence protocol unit: in the case of identifying the existence of a heterogeneous signal, the WiFi sender The idle channel detection threshold is reduced to the second idle channel detection threshold, the distributed inter-frame gap time is ended, the backoff process is entered, and the second idle channel detection threshold is used to detect the channel state in the backoff time slot, and then the WiFi is determined according to the detection result. The transmission timing of the packet. The present invention can be applied to actually deployed heterogeneous networks, and can also be used for laboratory simulation, network performance testing equipment, and the like.

In order to further verify the effect of the present invention, the effectiveness of the heterogeneous signal identification and coexistence mechanisms are respectively verified.

Specifically, for heterogeneous signal recognition, the wireless protocol data is obtained from the experimental platform built by USRP N210 and TelosB, and the convolutional neural network is obtained by offline training on the training set via the graphics card loaded with Intel Graphics 620 and the central processing unit of Intel Corei7-8550U. model, and then send the test set into the model to obtain the recognition accuracy shown in Figure 8(a) and the confusion matrix shown in Figure 8(b) (sample points are 64). c) The running times corresponding to different sample points are shown.

It can be seen from Figure 8(a) that all the test set data obtained a recognition accuracy of more than 99.9%, which verified the accuracy of different sample points generated using the same data set. Since the smaller the number of sample points, the same data set is divided by The more short frame vectors obtained by the short frame operation, the higher the corresponding accuracy. In addition, when dividing short frames, only one short frame is selected from multiple short frames, and the number of short frames obtained by corresponding different sample points is the same, as shown in Figure 8(a), the higher the number of sample points, the higher the frequency spectrum The larger the resolution, the higher the corresponding recognition accuracy. Figure 8(b) is the corresponding confusion matrix when the number of sample points is 64 points. The predicted value of ZigBee label is consistent with the actual value, which means that all ZigBee signals are recognized, and 0.0187% of noise signals will be mistaken for ZigBee The signal, after analysis, is caused by part of the WiFi signal (the label of the WiFi signal is also Noise, because it only cares whether there is a ZigBee signal), but the proportion is too small, and the impact on the performance is negligible. Figure 8(c) is the average time for the model to run each short frame. The longer the number of sample points, the longer the corresponding running time. When the number of sample points is 64 points, the average recognition time is 66 microseconds. If a special neural network chip is used, such as TPU, the speed will be increased by more than 10 times, which can be shortened to 6us. Experiments show that the present invention uses the convolutional neural network model to identify heterogeneous signals, which fully complies with the design objectives and has excellent performance.

Further, the performance of the coexistence protocol is verified, and the verification is carried out by means of simulation. The simulation uses the ns-3 library, and the simulation topology is shown in Figure 9. The general settings are consistent with the signal acquisition experiment, but due to the ns-3 library, if the distance between the ZigBee transceiver nodes is less than 2 meters, the phenomenon of continuous channel preemption may occur. Therefore, the default distance is set to 5 meters. In the simulation experiment, the throughput of WiFi is used as the performance indicator, and since ZigBee sends less data and the rate is low, it is not suitable to measure the throughput. Therefore, it is preferable to use PRR (PacketReceiveRate, packet reception rate) as the measurement indicator.

The specific coexistence protocol simulation process is as follows: firstly, modify the settings of the WiFi transmitter to verify the performance of the coexistence protocol of the present invention under various WiFi standard transmission rates. It can be seen from Fig. 10(a) that, compared with the standard protocol, the protocol designed by the present invention brings more than twice the improvement of the PRR of ZigBee, which means that the protocol design brings great benefits to the ZigBee protocol. When the sending rate is low, the WiFi throughput is slightly lost compared to the standard protocol. When the sending rate is high, the loss of WiFi throughput is minimal.

Figure 10(b) shows the performance of different WiFi data packet lengths. At this time, the WiFi data transmission rate has been adjusted back to the default 54Mbps. It can be seen from the figure that the larger the packet length, the larger the throughput, and this protocol has little effect on the WiFi throughput, and the use of the coexistence mechanism protocol of the present invention will increase the PRR by more than two times, with excellent performance.

Next, change the experimental topology structure, by modifying the distance d _z between ZigBee devices, the distance d _wz between the ZigBee sending node and the WiFi sending device, and observe the changes of the corresponding performance indicators, get the different topological distances as shown in Figure 11 Protocol performance.

When d _z is within 2m, the signal-to-noise ratio of the receiving end is large enough, and the PRR is close to 100%; when it increases to 3m, the PRR can be increased from 90% to 100% by using the protocol proposed by the present invention; while increasing d _z to more than 4m , the PRR begins to drop sharply, and the PRR will be increased by nearly 2 times by using the present invention. Since the distance between the ZigBee sender and the WiFi sender remains unchanged, the throughput is constant, and the protocol of the present invention has little impact on the throughput.

Although when the d _wz is within 2m, if the protocol of the present invention is adopted, ZigBee will continue to occupy the channel, resulting in the WiFi throughput dropping to 0. Fortunately, in reality, the WiFi transmitter is usually located at a high place, and there is little chance of close contact with the ZigBee transmitter. When d _wz is more than 3m, the protocol proposed in the present invention will increase the PRR by about two times, and the WiFi throughput will be almost unchanged.

To sum up, the present invention proposes to apply deep learning to ZigBee signal recognition for the first time. With the help of the convolutional neural network model, the hidden features of a small number of ZigBee signal samples with low signal-to-noise ratio can be automatically extracted to realize the recognition of weak ZigBee signals at the microsecond level. The present invention applies the convolutional neural network model to wireless protocol identification, and provides a brand-new idea of heterogeneous signal identification. Further, the present invention designs a MAC layer coexistence protocol in combination with the deep learning identification results, and solves the problem of cross-protocol interference caused by power asymmetry by simply modifying the CCA threshold of the DCF mechanism adopted by WiFi, starting from the interference side, and realizing WiFi and ZigBee. The protocol is more fair and reasonable channel contention strategy.

Compared with the existing heterogeneous signal identification scheme, the deep learning-based heterogeneous signal identification scheme proposed by the present invention achieves a recognition accuracy rate as high as 99.9%, does not require high spectral resolution, and achieves extremely short delay at the microsecond level. The existing coexistence mechanism design scheme either requires additional nodes to be laid, or needs to modify the transmitter and the receiver at the same time, while the coexistence protocol design proposed by the present invention, combined with the existing DCF mechanism, only needs to modify the WiFi transmitter, and The current commercial equipment has high compatibility and is more convenient for large-scale promotion and deployment.

The present invention may be a system, method and/or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of the present invention.

A computer-readable storage medium may be a tangible device that can hold and store instructions for use by the instruction execution device. The computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (non-exhaustive list) of computer readable storage media include: portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM) or flash memory), static random access memory (SRAM), portable compact disk read only memory (CD-ROM), digital versatile disk (DVD), memory sticks, floppy disks, mechanically coded devices, such as printers with instructions stored thereon Hole cards or raised structures in grooves, and any suitable combination of the above. Computer-readable storage media, as used herein, are not to be construed as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (eg, light pulses through fiber optic cables), or through electrical wires transmitted electrical signals.

The computer readable program instructions described herein may be downloaded to various computing/processing devices from a computer readable storage medium, or to an external computer or external storage device over a network such as the Internet, a local area network, a wide area network and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers. A network adapter card or network interface in each computing/processing device receives computer-readable program instructions from a network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing/processing device .

The computer program instructions for carrying out the operations of the present invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state setting data, or instructions in one or more programming languages. Source or object code, written in any combination, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as the "C" language or similar programming languages. The computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on the remote computer or server implement. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (eg, using an Internet service provider through the Internet connect). In some embodiments, custom electronic circuits, such as programmable logic circuits, field programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), can be personalized by utilizing state information of computer readable program instructions. Computer readable program instructions are executed to implement various aspects of the present invention.

Aspects of the present invention are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer readable program instructions.

These computer readable program instructions may be provided to a processor of a general purpose computer, special purpose computer or other programmable data processing apparatus to produce a machine that causes the instructions when executed by the processor of the computer or other programmable data processing apparatus , resulting in means for implementing the functions/acts specified in one or more blocks of the flowchart and/or block diagrams. These computer readable program instructions can also be stored in a computer readable storage medium, these instructions cause a computer, programmable data processing apparatus and/or other equipment to operate in a specific manner, so that the computer readable medium on which the instructions are stored includes An article of manufacture comprising instructions for implementing various aspects of the functions/acts specified in one or more blocks of the flowchart and/or block diagrams.

Computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other equipment to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other equipment to produce a computer-implemented process , thereby causing instructions executing on a computer, other programmable data processing apparatus, or other device to implement the functions/acts specified in one or more blocks of the flowcharts and/or block diagrams.

The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more functions for implementing the specified logical function(s) executable instructions. In some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It is also noted that each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented in dedicated hardware-based systems that perform the specified functions or actions , or can be implemented in a combination of dedicated hardware and computer instructions. It is well known to those skilled in the art that implementation in hardware, implementation in software, and implementation in a combination of software and hardware are all equivalent.

Various embodiments of the present invention have been described above, and the foregoing descriptions are exemplary, not exhaustive, and not limiting of the disclosed embodiments. Numerous modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein was chosen to best explain the principles of the embodiments, the practical application or technical improvement in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein. The scope of the invention is defined by the appended claims.

Claims

A method for realizing coexistence of heterogeneous networks, comprising the following steps:

Step S1: The WiFi sending end uses the set first idle channel detection threshold to determine the channel state based on the collected environment samples;

Step S2: in the case of judging that the channel is in an idle state, within the set distributed inter-frame gap time, input the environment sample into a pre-trained neural network model to identify whether there is a heterogeneous signal;

Step S3: In the case of identifying the existence of heterogeneous signals, the WiFi transmitter reduces the first idle channel detection threshold to the second idle channel detection threshold, ends the distributed inter-frame gap time, enters the backoff process, and is in the backoff process. The time slot uses the second idle channel detection threshold to detect the channel state, and then determines the transmission timing of the WiFi data packet according to the detection result.
The method according to claim 1, wherein step S3 comprises the following sub-steps:

Set the heterogeneous signal existence flag and communicate it to the medium access control layer of WiFi;

The medium access control layer of WiFi reduces the first idle channel detection threshold to the second idle channel detection threshold, and the WiFi sender ends the distributed inter-frame gap time, enters the backoff process, and randomly selects a value from the contention window as the For the value of the random backoff counter, idle channel detection is performed for each backoff time slot included in the backoff time. When it is detected that the channel is busy, the random backoff counter is suspended to delay the transmission of WiFi data packets.
The method of claim 1, wherein training the neural network model comprises the steps of:

A sample data set is constructed, and the sample data set represents the corresponding relationship between the spectral characteristics of the collected signal and the existence label of the heterogeneous signal, and the collected signal includes pure ZigBee signal data, pure WiFi signal, and includes ZigBee signal, background noise and WiFi signal data;

performing a vectorization operation on the sample data set;

Use the vectorized sample data set to train the neural network model to obtain the neural network model that satisfies the set optimization goal.
The method of claim 3, wherein said constructing a sample data set comprises:

Divide the collected original sample data stream into frames at every 1024 sample points to obtain long frames, and perform fast Fourier transform on each long frame to obtain a spectrogram;

Extract the signal bandwidth and center frequency from the spectrogram as spectral features;

The long frame of 1024 sample points is subdivided into short frames of 64 sample points, and the label of the long frame is expanded by 16 times as the label of the short frame.
The method according to claim 4, wherein before performing the operation of dividing the short frame, it further comprises removing the signal boundary long frame and the corresponding label.
The method of claim 4, wherein performing a vectorization operation on the sample data set comprises:

The sample data set is represented as a four-dimensional vector form N example ×Dim IQ ×Dim value ×Dim channel , where Nexample represents the number of short frames, Dim IQ =2 represents the data of two IQ channels, and Dim value =64 represents each short frame. The frame contains only 64 sample points, and Dim channel = 1 represents the value representing the black and white image in image processing.
The method according to claim 1, wherein the neural network model sequentially comprises: a zero-padding layer for storing boundary information; a convolution layer for extracting hidden features of the signal, and using a linear rectification function for nonlinearization; A Dropout layer for dropping neurons; a flattening layer for making multidimensional data one-dimensional; and two fully connected layers used as output layers.
A system for realizing coexistence of heterogeneous networks, including:

Channel state detection unit: the WiFi sending end uses the set first idle channel detection threshold to determine the channel state based on the collected environment samples;

Heterogeneous signal identification unit: in the case of judging that the channel is in an idle state, within the set distributed inter-frame gap time, the environment sample is input into the pre-trained neural network model to identify whether there is a heterogeneous signal;

Coexistence protocol unit: In the case of identifying the presence of heterogeneous signals, the WiFi transmitter reduces the first idle channel detection threshold to the second idle channel detection threshold, ends the distributed inter-frame gap time, enters the backoff process, and The backoff time slot uses the second idle channel detection threshold to detect the channel state, and then determines the transmission timing of the WiFi data packet according to the detection result.
A computer-readable storage medium having stored thereon a computer program, wherein the program, when executed by a processor, implements the steps of the method of claim 1 .