Multi-Hop Transmission Method of 5G Millimeter Wave Relay Based on Network Coding

Yilin Wang ,Weisheng He ,Xuwei Fan ,Lianfen Huang ,Jie Yang ,Yuliang Tang,*

1 School of Information,Xiamen University,Xiamen 361005,China

2 Department of Software Engineering,Xiamen Institute of Software Technology,Xiamen 361005,China

Abstract: With the rapid development of the Internet technology,millimeter wave (mmWave) will be used as a supplement to 5G low frequency bands to meet the extremely high system capacity requirements of 5G in hot spots.Although 5G mmWave communication can adapt to the needs of 5G network and carry a large amount of transmitted data,transmission stability has become one of the key technical issues of 5G network mmWave communication due to problems such as strong attenuation and poor penetration of mmWave.In order to improve the efficiency of the mmWave multi-hop transmission,we propose a 5G mmWave multi-hop transmission method based on network coding,which can adapt to the current wireless network environment,improve spectrum efficiency and increase network throughput.Based on MATLAB simulation experiments,it is verified that the proposed method can greatly improve the transmission efficiency and reduce the signal loss under the premise of ensuring the accurate signal transmission.

Keywords: 5G network;millimeter wave relay;multihop communication;network coding

I.INTRODUCTION

The commercialization of 5G has gradually integrated more efficient,intelligent and user-friendly network services into our lives,but there are still many key technologies that require academic and industrial research.MmWave has more abundant available spectrum resources and higher transmission rate.It is an important support for 5G network in terms of transmission rate and capacity in the future.Therefore,the extension of 5G communication to mmWave has become an inevitable trend[1—3].

In order to make use of the advantages of mmWave in transmission rate and capacity in 5G network,mmWave relay communication is applied to solve the problems of strong attenuation and weak stability of mmWave transmission [4,5].The essence of mmWave relay equipment is to improve the existing wireless relay equipment and add corresponding function modules to carry out the communication transmission between mmWaves.Its most basic purpose is still to relay the wireless signal from the previous network device node to the next one,thereby gradually expanding the coverage of the wireless signal and adding new wireless coverage area.Faced with the problems of large space loss,short effective transmission distance,and susceptibility to mmWave interference,mmWave relay can continue transmission,expand coverage,and improve communication stability[6—9].

With the widespread application of mmWave relays and wireless communication,bandwidth resources will become increasingly scarce in the future,and multi-hop transmission will become more and more common.In the past,the data transmission efficiency of multi-hop transmission was relatively low.Therefore,it’s necessary to improve the stability of mmWave relay transmission and focus on the optimization of transmission efficiency in the future.In the current stage of 5G commercialization,for making use of the advantages of 5G mmWave in transmission efficiency and capacity,the relay technology should not only support mmWave communication in transmission,but also further combine with other technologies to optimize 5G mmWave communication from multiple perspectives[10,11].Network coding technology is an emerging coding technology,which mainly acts on the communication relay node in the network.The information received on each channel of the node can be linearly or nonlinearly coded,and then forwarded to the downstream.The communication node enables the relay node to increase the utility of encoding processing,so as to improve the throughput and robustness of the network.Applying network coding to mmWave relays can improve network security and communication efficiency at the same time[12—16].

Aiming at the low efficiency of multi-hop transmission,a mmWave multi-hop transmission method based on network coding is studied.We combine relay technology with network coding to effectively improve network throughput and save bandwidth,thus improving the stability and transmission efficiency of 5G mmWave relay communication and ultimately optimize the efficient communication of 5G network.

The rest of this paper is organized as follows.Section II introduces related research in this field.The main content of the Section III is the principle of multi-hop transmission technology based on network coding.Section IV is the program realization process.In Section V,experiments are carried out to verify the optimization effect of the method on the throughput and transmission efficiency of the mmWave multi-hop transmission.Finally,conclusions are given in Section VI.

II.RELATED WORK

For the purpose of improving the stability and efficiency of massive multiple input multiple output(MIMO) mmWave communication technology,we can introduce wireless relays to make up for the shortcomings of mmWave communications.

The research of mmWave relay communication process optimization is diversified because there are many different ways to optimize the communication transmission process.

In [17],a novel scheme has been proposed to improve coverage area,spectrum and energy efficiency for mmWave communications by Jaipreet Kaur and Maninder Lal Singh.Xingqin Lin and Jeffrey G.Andrews presents a novel stochastic geometry approach to the connectivity of mmWave networks with multihop relaying.They find that multi-hop relaying can greatly improve the connectivity versus single hop mmWave transmission [18].Ahmed Abdelreheem,Osama A.Omer etc.proposed a novel deep learningbased relay selection scheme in mmWave device-todevice (D2D) communication underlying the 5G cellular networks.This proposed deep learning model is developed to overcome the challenges of selecting the optimal relay based on low complexity and high efficiency[19].

As an important collaboration technology commonly used in wireless relay,network coding technology can effectively improve the throughput,stability and transmission efficiency of the relay transmission process.At present,network coding technology is not only applied to wireless networks to improve network throughput,but also has been mature in peer-to-peer(P2P) systems,distributed file storage systems and network security integration and collaboration[20].

In [21],network coding and wireless relay cooperation are applied to link retransmission operations during uplink and downlink transmission.In the experimental simulation,the collaboration is conducted in a poor channel environment with low or medium signal-to-noise ratio.The method can significantly improve the overall throughput of the system,and when the full-duplex relay is located in the optimal location of the cell,the throughput reaches the best optimization effect.In[22],random linear network coding with wireless relay in D2D network is adopted.Compared with direct D2D communication,it can optimize the quality and stability of the transmission process significantly.In [23],a method of cooperative application of physical layer network coding and polar code on wireless relay is proposed.This application method significantly improves the channel fading effect in the transmission process,and the bit error rate of the transmission is also largely reduced.In[24],a collaborative method of applying network coding and relay equipment to a distributed storage system is proposed.It can not only make use of the technical advantages of network coding to effectively reduce the probability of data failure and balance the load balance between devices,but also rely on the cooperation of relay equipment and network coding to reduce transmission delay of data access,and finally to achieve the purpose of improving the transmission efficiency of data access and optimizing distributed storage applications.In[25],a network coding collaboration method based on multiple relays is proposed,which selects the best coding relay and forwarding relay according to dynamic changes of the user’s cell environment,and then plan the best network coding communication transmission method.Compared with the method of directly and randomly selecting the nearest relay base station,this method significantly reduces the influence of channel fading,increases data transmission rate,and optimizes the transmission efficiency in dynamic scenes.

It can be seen that network coding and wireless relay,when combined in a certain way and applied to the appropriate scenes,can bring optimization effects in terms of transmission rate,delay,stability,etc.,thereby improving the transmission effectiveness of the corresponding scene.Therefore,for massive MIMO mmWave communications,the application of mmWave relay to make up for the deficiencies of mmWave transmission can also be combined with network coding technology to improve the stability of massive MIMO mmWave communications[26,27].At the same time,the throughput and data transmission efficiency of the network are improved to accommodate the larger scale data transmission that the 5G network needs to carry in the future.

III.TRANSMISSION PRINCIPLE OF MULTI-HOP TRANSMISSION BASED ON NETWORK CODING

In order to solve the problem that the communication link is easily blocked by obstacles due to poor penetration ability and signal diffraction in mmWave communication,the introduction of mmWave relay technology can not only provide an alternative transmission path,but also provide enhanced signal methods such as improving the robustness of communications through collaboration.It plays an important role in improving the reliability and communication quality of mmWave communication system.

Although multi-hop relay can solve the problem of the mmWave transmission,it will consume more communication resources,increase work complexity and power overhead.How to effectively improve the communication efficiency of mmWave multi-hop transmission is of great significance to the future 5G mmWave communication.

Common application scenarios for mmWave multihop are:

(1)Chain undertaking transmission: The direct link between the source node and the destination node is in a deep fading scene such as a building or rain attenuation,where the signal attenuation is too large,and the mmWave signal blocking phenomenon is serious.At this time,if one or more relay nodes are introduced between the source node and the destination node,the original direct transmission can be extended to two or more hops,which can resist the influence of channel fading on signal transmission,eliminate the blind spots and blind area,and expand the coverage of the signal.

(2) Array cooperative transmission: Multiple relay nodes are used to collaborate to form a virtual distributed antenna array,which can enable the system to obtain diversity gain and improve the reliability of communication.

In addition,the relay transmission can be divided into single-user relay transmission and multi-user relay transmission according to the number of users simultaneously served by a relay.

We mainly carry out research in the chain undertaking transmission scenario,which achieves communication transmission through a single relay or multiple relay series to form a multi-hop link.Figure 1 presents a simplified scenario of mmWave communication.In the figure,the direct link between node A and node B is interrupted and information is exchanged through nearby relay node R1 and R2.

Figure 1.Scenario of millimeter wave relay communication.

Whether it is necessary to introduce relay assistance in the transmission process is mainly measured by the interruption probability and path loss.

Outage probability is defined as the probability that the channel capacityCis less than a given system rateR:

The outage probability is actually another expression of the link capacity.When the link capacity cannot meet the required user rate,the outage event will occur.This event is probabilistic and depends on the average signal-to-noise ratio (SNR) of the link and its channel fading distribution model.According to formula (1),the greater the probability that the capacity is less than the given rate is,the greater the probability of system outage is(that is,the smaller the system capacity is,the worse the system performance is,and the greater the probability of system interruption is).

In order to make up for the higher probability of outage and transmission loss in the transmission process,there are two main options: (a) Increasing the transmission power of the base station transmitter and the sensitivity of the device receiver.(b)Introduce relayassisted transmission between the transmitter and the receiver.Due to the structure of 5G super-dense heterogeneous network,increasing the power of the transmitter may increase the communication interference in the micro area.So the introduction of mmWave relay to assist transmission in 5G network is a more applicable way.

Massive MIMO technology has been added to 5G mmWave relay,and precoding is one of the core technologies of massive MIMO.According to the precoding coded instructions,the synchronous transmission of multiple antennas can be realized,thus improving the system capacity,energy efficiency and link reliability without increasing the signal transmission power.In addition,parallel and separate transmission of multiple antennas can be realized to bring spatial diversity gain and improve transmission throughput by taking advantage of the narrow beam and strong directivity of mmWave.

As an important cooperation technology of relay,network coding technology can be used to improve the efficiency of multi-hop transmission.Figure 2 shows the comparison of the working modes of network coding in two-hop relay interactive transmission.

Assuming that Figure 2(a),(b),(c)is set to consume only a single channel resource transmission between two points:

Figure 2.Comparison of working modes in two-hop transmission.(a)Traditional scheduling scheme (b)Straightforward network coding(c)Physical-layer network coding.

Figure 3.Schematic diagram of millimeter wave multi-hop transmission method based on network coding.

(a) represents the traditional scheduling scheme,which takes 4 time slots to complete the two-hop relay interactive transmission;

(b)represents the working mode of simple network coding,which can save a transmission time slot;

(c) represents the working method of using PNC(Physical-layer Network Coding).The principle of PNC is that the relay node can receive transmitted packets from both terminal nodes,encode the two messages into one packet,and then broadcast the encoded packet to the terminal node.Through these two steps,two time slots are saved to exchange packets between them,thus improving the overall network throughput and spectral efficiency of the network.This further improves efficiency by allowing interactions to be completed in two time slots only.

It can be seen that the PNC can increase the transmission efficiency by two times compared with the traditional original relay working mode.

The PNC mainly applies the same architecture and method to the reception and modulation of electromagnetic wave signals in the physical layer of wireless communication.Take a two-way transmission network for example.For simplicity,suppose there is a relay node R between A and B.3GPP TS23.501 separates the control plane(CU)and the user plane(DU)in the 5G system architecture,and expands the function from the core network to the wireless network.So node A and node B are in a synchronous system,sending and receiving data strictly according to the time slot through CU negotiation.Assuming that the simplest binary phase-shift keying modulation method is used in the physical layer modulation method,nodes A and B interact through the relay node R1,supposing that all signals are synchronized in time,frequency and phase,and the power amplification factor is the same.

In the first time slot,node A and node B send information to relay node R at the same time,and the joint baseband signal received by relay node R in one symbol period is:

In formula (2),ωis the carrier frequency,S1(t),S2(t) are the baseband signals sent by node A and node B,anda1,a2are the BPSK modulated signals of node A and node B respectively (i.e.,a ∈{1,?1}).In this way,the relay node R will receive the baseband signalI=a1+a2.Through this signal,the relay node R cannot decode the signala1,a2transmitted by the node A and the node B,but the relay node R does not need to obtain specific information and only needs to forward the information required for the decoding of the node A and the node B to achieve end-to-end information delivery.

In order to realize this process,the relay node R needs to perform a special modulation and demodulation mapping operation to make the information equivalent to the sum information of the bit stream sent by node A and node B on the finite field GF(2).This is the PNC mapping.

After PNC mapping,the relay node R performs BPSK modulation on the mapping resultS3to obtaina3,whereS3is equivalent to the result of mod 2 addition ofS1andS2:

Map(I)represents the demodulation mapping operation ofIby the relay node R,and the modulation and demodulation mapping operations of the node A,the node B,and the relay node R in this process are shown in Table 1.

Table 1.PNC modulation and demodulation mapping of two-hop relay interactive transmission network.

Table 2.Model parameters.

In the second time slot,the relay node R sendsa3:

After node A and node B receiveS3(t),they can obtainS3through BPSK demodulation,and then decodeS1andS2respectively by the XOR network coding calculation principle,thereby completing the information exchange between node A and node B in two time slots.For a frame exchange,ignoring the slightBER difference,PNC requires two time slots,traditional scheduling scheme requires four,while straightforward network coding requires three.Therefore,PNC can improve the system throughput of the threenode wireless network by 100% and 50% compared with traditional transmission scheduling and straightforward network coding,respectively[28].

Therefore,the application of network coding technology to multi-hop transmission links can also improve the transmission efficiency of the network.Figure 3 shows the information interaction process between the network gateway and the mmWave edge base station.The principles of the information interaction will be explained in detail in the next section.We propose a mmWave multi-hop transmission method based on network coding to optimize the efficiency of the 5G network mmWave multi-hop transmission.At the same time,this method is also applicable to other multi-hop networks.

IV.IMPLEMENTATION OF MMWAVE MULTI-HOP TRANSMISSION METHOD BASED ON NETWORK CODING

The method of mmWave multi-hop transmission based on network coding in this paper can use network coding to encode some data packets,so that these packets can be transmitted in “free riding” mode without additional channel resource overhead,so as to improve the efficiency of multi-hop transmission.Figure 4 is a working diagram of the transmission method.In this case,there are three relay nodes in the middle.Let us take this case for example to illustrate the working principle of the method.

Figure 4.Working diagram of multi-hop transmission method based on network coding.

Firstly,the setting conditions of the working state of each equipment in this method are explained:

1.Full duplex mode: Network relay and edge devices or network gateway can send and receive at the same time;

2.Data storage function: In addition to network gateway and edge devices at the beginning and end of the transmission link,all the network relay nodes in the middle also need to have the storage capacity.For the received and forwarded data,at least two time slots should be stored and then released;

3.Limit of transmission channel consumption: In the interaction scenario,each device is set to consume only two channel resources per time slot.One channel resource is used for sending,and the other for receiving,and one channel resource can only transmit one packet.

4.Network coding technology: This method uses physical layer network coding.The intermediate node only encodes and transmits according to the currently received content combined with historical storage content.The intermediate node does not need to decode and know the content of the packet.To be more specific,XOR network coding calculation is used,including the following calculation principles:

This principle shows that two packets A and B can get a coded packet (A ⊕B) after XOR network coding calculation.

This principle represents the decoding process of XOR network coding,which is the XOR calculation between the encoded data packet (A ⊕B) and one of the original data packets can get another original data packet.

This principle represents the compound operation process between data packets of XOR network coding.The principle of calculation is that even-numbered items cancel each other out,and odd-numbered items are retained.

The above is the calculation principle of network coding involved in this method.

In Figure 4,the dark blue rectangle represents the network gateway or edge devices,both of which need to transmit interactive packets,and the light blue rectangle represents the relay device node passing through the multi-hop chain link.Each arrow indicates the transmission action between devices.The content of the transmitted packet is marked above the arrow.The black arrow indicates effective data transmission.The gray arrow indicates unreceived or useless data transmission.The working time slot is marked on the left,except for the first row,the remaining two rows represent a time slot.This schematic diagram shows the dynamic changes of the data transmitted by each device after the method starts to work.

Take these three relay nodes in the middle for example.The network edge device needs to actively upload packets{1,2,3,...}to the gateway.In this process,the data packets{x1,x2,x3,...}can be transmitted to edge devices by“free riding”without additional channel resource overhead.

The process from time slot 1 to time slot 4 is a normal working process of the chain transmission of the topology network.

In slot 5,the network gateway sends the data packet(1⊕x1)which is processed by network coding.The data packet 1 is sent to the network gateway in slot 4,andx1 is the data packet stored by the network gateway and needs to be transmitted.After XOR network coding,the(1⊕x1)data packet is sent in this slot.

In time slot 6,the network gateway continues to send the network encoded data packet(2⊕x2)to the network relay node 3 just like in time slot 5.Node 3 broadcasts the data packet (3⊕x1) to the front and back devices(that is,the relay node 2 and the network gateway device).The data packet(3⊕x1)is obtained by XOR network coding of the combination of data packet 3,the data packet (1⊕x1) and data packet 1,which was stored before in time slot 5:

Because the physical layer network coding is used,the data packets can be stored after network coding operation.Only one data packet can be received in a time slot.

The working principle of time slot 7 is the same as that of time slot 6.Time slot 7 is the working principle of Figure 3,where the first relay node corresponds to the mmWave edge base station.

In time slot 8,the network relay node 1 transmits data packets (7⊕x1) to edge devices.The edge device can obtain the data packetx1 by decoding the data packet using its own data packet 7:

Another example is given to illustrate the origin of the data packet (5⊕x3) broadcasted and transmitted by the network relay node 3 at this time.The data packet is obtained by XOR network coding of:

1.Data packet (5⊕x1) transmitted by the relay node 2 in slot 7;

2.Data packet (3⊕x1) stored by relay node 3,which will be send in slot 6;

3.Data packet(3⊕x3)from the network gateway.

Following this principle,the edge device can upload data packets to the network gateway.In the meantime,the physical layer network coding technology can be used to gradually transmit the data packets of the network gateway to the edge device without additional channel consumption,so as to efficiently realize the data interaction of the intermediate multi-hop link.The above example shows that the method is suitable for the scenario where network topology can be abstracted as a chain relay transmission model.If the device meets the setting requirements mentioned above,the efficient transmission of the method can be realized.

At the same time,thanks to the use of PNC,the data packets stored in the relay device do not need to be decoded in the process of intermediate interaction.Therefore,it can also improve the security and privacy of the transmission process: firstly,it reduces the probability of the original packet content being decoded in the interaction path;secondly,even if packets are stolen,the network coding processing will make the content more difficult to decode and know.

The above is mmWave multi-hop transmission method based on network coding.In order to study the optimization effect of the method,we will use the MATLAB platform to carry out the simulation experiment in the next section.

V.SIMULATION EXPERIMENT AND DISCUSSION

In the previous section,the mmWave multi-hop transmission method based on network coding is proposed,and an example is given to illustrate its working principle in detail.In this section,the simulation experiment of this method will be carried out on MATLAB platform.

The propagation of mmWave can be divided into two cases,line-of-sight (LOS) and non-line-of-sight(NLOS).The channels corresponding to the two cases have different path losses.MmWaves are very sensitive to obstacles.It is generally considered that mmWave LOS propagation is similar to free space propagation.The path loss index can be set to 2,and the path loss index of the NLOS link is considered to be greater than that of the LOS link.

The calculation formula of the propagation loss of the urban macro(UMa)large-scale spatial loss modelin the LOS/NLOS scenario is as follows.

whered3D(m)is the distance between transmitter and receiver,fc(GHz) is the center frequency.hBS(m)andhUT(m) are antenna height for base station and user terminal respectively.

Compared to the free space(FS)path loss equation:

which ideally describes the attenuation a signal experience.

3GPP has released a path loss model for frequencies ranging from 0.5 to 100 GHz.Here we mainly simulate in the UMa scenario.Table 2 shows our parameter settings.

Now we study the comparison of changes in the number of transmitted data packets before and after the use of this transmission method in the mmWave multi-hop chain link model.

Figure 5 shows the change of the number of packets before and after we use this method when the number of intermediate nodes is from 1 to 4.The blue line indicates the change in the number of packets of the original transmission data.The red line indicates the change in the number of packets of the new transmission data,and the yellow line indicates the change in the number of packets of the total transmitted data.

In data transmission,throughput is the amount of data transmitted successfully from one terminal node to another in a given time period.Therefore,in the case of the same transmission time slot,the throughput improvement can be reflected by the increase in the amount of data transmitted.It can be found from Figure 5 that the completion of the first transmission of new data usually needs to go through N time slots(N=“number of intermediate nodes+1”)after the completion of the first transmission of the original transmission data.In addition,the less the number of intermediate nodes is,the lower the ”start time cost of time slot”of incremental data transmission is,and the faster the amount of data transmitted increases.

In order to facilitate further integration and comparison of the change of the amount of transmission data with the transmission time slot and the number of intermediate nodes,a three-dimensional diagram as shown in Figure 6 is adopted.

As shown in Figure 6,as the number of intermediate nodes increases,more time slots need to be spent to improve the amount of data transmitted by this method.Therefore,after passing through the same time slot,the more the number of intermediate nodes is,the less the amount of network data transmitted by this method is(the intersection of the red rectangle and the coordinate when the transmission time slot is 50 in Figure 6,the amount of data transmitted when the number of relays is 4 is less than that when the number of relays is 1).However,when the transmission process continues,which means that the transmission time slot approaches infinity,the improvement effect of this method on the amount of transmitted data is expected to double.As a result,throughput can be nearly twice that of the traditional method,which can improve the transmission efficiency and save transmission time and channel bandwidth.

As shown in Figure 7,we further analyzed the percentage of data volume based on the data in Figure 6.Two conclusions can be drawn:

Figure 5.Comparison of the number of data packet transmissions based on PNC in mmWave multi-hop transmission methods.

Figure 6.Three-dimensional comparison of data packet transmitted in mmWave multi-hop transmission method based on PNC.

Figure 7.Three-dimensional graph of the percentage increase in the amount of transmitted data.

(1)The completion of the first transmission of new data usually needs to go through N time slots (N=“number of intermediate nodes+1”)after the completion of the first transmission of the original transmission data.

(2) The more the number of intermediate nodes is,the less obvious the improved effect of this method on data transmission is.However,when the time slot approaches infinity,the transmission network with different numbers of intermediate nodes will transmit twice the amount of data at different rates through this method.

As shown in Figure 8,we simulated the path loss in different modes.In different frequencies and modes,the attenuation of NLOS scenario is the most obvious and the loss in LOS path is very similar to that in FS path.The path loss of using multi-hop transmission method based on PNC is smaller than that of using traditional transmission method.This indicates that the method proposed in this paper can not only increase throughput and improve data transmission efficiency,but also ensure the stability of communication transmission.With the rapid development of mobile communication networks and intelligent terminals,our method can be applied to current and future wired and wireless communication environments.

Figure 8.Path loss comparison of multi-hop transmission in UMa scenario.

VI.CONCLUSION

Aiming at the intelligent optimization of the 5G mmWave relay,a 5G mmWave multi-hop transmission method based on network coding was proposed.Our method combines network coding technology with multi-hop transmission,which can not only adapt to the current wireless network environment,but also improve spectrum efficiency and network throughput while ensuring the stability of communication.MATLAB simulation results show that this method can gradually approach the transmission of twice the amount of data on the premise of ensuring accurate transmission as the transmission time slot progresses.It is also verified that this method can effectively improve the data transmission efficiency of multi-hop transmission and save channel bandwidth.

MmWave communication is an important support for future 5G network in terms of transmission rate and transmission capacity.The combination of network coding technology and multi-hop transmission has a significant effect on network communications.We introduced network coding technology and mmWave relay to work together,so as to improve the stability of mmWave communication while increasing network throughput and decreasing path loss.If we can optimize and improve the beam management technology,and increase its stability and reduce sensitivity,it will be more conducive to the wide application of mmWave communication,to bring optimization effects to more 5G network communication scenarios.Moreover,our solution is simple and easy to implement,it is worth continuing to explore in the future to expand the collaborative method of network coding and relay base stations to more network scenarios.Also,more 5G network communication scenarios can be optimized in combination with other methods.

ACKNOWLEDGEMENT

This work was partially supported by National Natural Science Foundation of China (No.61871339,61731012,61971365),Key Laboratory of Digital Fujian on IoT Communication,Architecture and Security Technology (No.2010499),National Ministry of Science and Technology,863 Project (No.2015AA01A705) and Industry-University Cooperation Project of Department of Science and Technology of Fujian Province(No.2020H6001).