阅读说明：本技术 面向非对称毫米波大规模mimo的多用户上下行波束对准方法 (Multi-user uplink and downlink beam alignment method for asymmetric millimeter wave large-scale MIMO ) 是由 刘永城 李旻 赵民建 李立言 张嘉瑀 于 2021-09-17 设计创作，主要内容包括：本申请提出了一种面向非对称毫米波大规模MIMO的多用户上下行波束对准方法、装置及介质,该方法包括：构造用于探测多个方向的全数字多指向波束；进行基站端到用户端的多轮下行波束训练；控制用户端根据接收信号进行波束判决,确定出最大接收功率的目标下行发送-接收波束对；对接收信号进行数据处理,生成预测上行发送窄波束的训练数据,并对预设的神经网络进行训练；基于训练后的网络参数和接收信号进行在线实时信号检测,预测目标上行发送窄波束；根据反馈的目标下行发送窄波束的标号,将目标下行发送窄波束展宽为基站端目标上行接收波束。该方法以较少的训练开销和较高的可靠性,实现了多用户上下行波束的快速对准,简化了训练过程。(The application provides a multi-user uplink and downlink beam alignment method, a device and a medium for asymmetric millimeter wave large-scale MIMO, wherein the method comprises the following steps: constructing a full digital multi-directional beam for detecting multiple directions; performing multiple rounds of downlink beam training from a base station end to a user end; the control user end carries out beam judgment according to the received signal and determines a target downlink transmitting-receiving beam pair with the maximum receiving power; processing the received signal to generate training data for predicting uplink transmission narrow beams, and training a preset neural network; performing on-line real-time signal detection based on the trained network parameters and the received signals, and predicting a target uplink transmission narrow beam; and widening the target downlink transmission narrow beam into a target uplink receiving beam at the base station end according to the fed back label of the target downlink transmission narrow beam. The method realizes the rapid alignment of uplink and downlink beams of multiple users with less training overhead and higher reliability, and simplifies the training process.)

1. A multi-user uplink and downlink beam alignment method for asymmetric millimeter wave large-scale MIMO is characterized by comprising the following steps:

constructing a multi-directional beam codebook, and constructing a full-digital multi-directional beam for detecting multiple directions through optimized code words in the codebook;

controlling the base station end to send the multi-directional beams, controlling the user end to receive the multi-directional beams by using the receiving beams, and calculating the matched filtering output of the receiving signals of the user end so as to perform multiple rounds of downlink beam training from the base station end to the user end;

controlling the user terminal to carry out beam judgment according to the received signals, and determining a target downlink transmitting-receiving beam pair with maximum receiving power from the received signals received in the multiple rounds of downlink beam training;

carrying out data processing on received signals received by the user side under different signal-to-noise ratios, generating training data for predicting narrow beams sent by the user side to the base station side in an uplink mode, and carrying out offline training on a preset convolutional neural network through the training data to obtain trained network parameters;

performing online real-time signal detection based on the trained network parameters and the received signals of the user side to predict target uplink transmission narrow beams from the user side to the base station side;

and controlling the user side to feed back the label of the target downlink transmission narrow beam of the base station end to the base station end, and widening the target downlink transmission narrow beam into a target uplink receiving beam of the base station end.

2. The alignment method of claim 1, wherein the constructing a multi-directional beam codebook comprises:

constructing a minimum mean square error function for optimizing the multi-directional beam;

and solving the minimum mean square error function through a Riemannian manifold optimization algorithm to obtain optimized code words emitted to different directions, and combining the code words to generate the codebook.

3. The alignment method of claim 2, wherein the minimum mean square error function is expressed by the following formula:

wherein, s.t.w^Hw＝1，

Where, w is a single directional beam,in order to be a vector of the response of the array,for the departure angle, g is the ideal gain vector, L_TIs the size of the codebook, g, sent by the base station during downlink transmission^*Is the conjugate transpose of the ideal gain vector;

constructing an all-digital multi-directional beam for detecting multiple directions by the following formula:

wherein M [. C]Is the direction of the multi-directional beam, M is the number of directions of the multi-directional beam,indicating to either directionThe transmitted optimized codeword.

4. The alignment method according to claim 1, wherein in each round of downlink beam training from the base station to the user terminal, the base station transmits a plurality of training symbols and determines a combination of directions for collecting the observation signals in any training symbol, and the round of downlink beam training from the base station to the user terminal comprises:

in each round of downlink beam training, the base station end sends a preset number of pilot frequency sequences;

determining a combined signal corresponding to any training symbol in the current round of downlink beam training received by the user terminal;

and calculating the matched filtering output of the combined signal corresponding to any training symbol.

5. The alignment method according to claim 1, wherein the performing data processing on the received signals with different signal-to-noise ratios received by the user end to generate training data for predicting narrow beam uplink transmission from the user end to the base station end comprises:

combining the matched filtered outputs of the received signals received by the user terminal under different signal-to-noise ratios by the following formula:

wherein the content of the first and second substances,r is the training round, n is any user, k^*Is a receiving beam of a user terminal, L_RReceiving codebook size for a user terminal, L ═ L_T/M，L_TTransmitting the codebook size for the base station;

sequentially normalizing the received signals of each round of downlink beam training and then combining the normalized received signals;

and taking the real part and the imaginary part of the normalized signal and a receiving beam of the user end as the training data.

6. The alignment method according to claim 1, wherein the broadening the target downlink transmission narrow beam into a base station target uplink reception beam comprises:

based on the asymmetric characteristic of the uplink and downlink beams, the central angle of the target uplink receiving beam of the base station end is kept consistent with the target downlink sending narrow beam, and the width of the target uplink receiving beam of the base station end is widened to be the preset beam width.

7. A multi-user uplink and downlink beam alignment device facing asymmetric millimeter wave massive MIMO is characterized by comprising:

the system comprises a construction module, a detection module and a control module, wherein the construction module is used for constructing a multi-directional beam codebook and constructing a full-digital multi-directional beam for detecting multiple directions through optimized code words in the codebook;

the training module is used for controlling the base station end to send the multi-directional beams, controlling the user end to receive the multi-directional beams by using the receiving beams, and calculating the matched filtering output of the receiving signals of the user end so as to carry out multi-round downlink beam training from the base station end to the user end;

a decision module, configured to control the user side to perform beam decision according to the received signal, and determine a target downlink transmit-receive beam pair with the maximum received power from the received signals received in the multiple rounds of downlink beam training;

the data processing module is used for carrying out data processing on the received signals received by the user side under different signal-to-noise ratios, generating training data for predicting narrow beam uplink transmission from the user side to the base station side, and carrying out offline training on a preset convolutional neural network through the training data to obtain trained network parameters;

the prediction module is used for carrying out online real-time signal detection on the basis of the trained network parameters and the received signals of the user side so as to predict the target uplink transmission narrow beam from the user side to the base station side;

and the broadening module is used for controlling the user side to feed back the label of the target downlink transmission narrow beam of the base station side to the base station side, and broadening the target downlink transmission narrow beam into a target uplink receiving beam of the base station side.

8. The alignment device according to claim 8, wherein the building block is specifically configured to:

constructing a minimum mean square error function for optimizing the multi-directional beam;

and solving the minimum mean square error function through a Riemannian manifold optimization algorithm to obtain optimized code words emitted to different directions, and combining the code words to generate the codebook.

9. The alignment device of claim 8, wherein the build module is further configured to:

the minimum mean square error function is expressed by the following formula:

wherein, s.t.w^Hw＝1，

Where, w is a single directional beam,in order to be a vector of the response of the array,for the departure angle, g is the ideal gain vector, L_TIs the size of the codebook, g, sent by the base station during downlink transmission^*Is the conjugate transpose of the ideal gain vector;

constructing an all-digital multi-directional beam for detecting multiple directions by the following formula:

wherein M [. C]Is the direction of the multi-directional beam, M is the number of directions of the multi-directional beam,represents the optimized codeword transmitted to either direction.

10. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the method for multiuser uplink and downlink beam alignment for asymmetric millimeter wave massive MIMO according to any of claims 1-6.

Technical Field

The present application relates to the field of wireless communications technologies, and in particular, to a method, an apparatus, and a storage medium for aligning uplink and downlink beams of multiple users for asymmetric millimeter wave massive MIMO.

Background

With the continuous development of social economy, applications such as VR, AR, 3D media, and ultra-high definition video transmission are in a wide range, and the service volume of wireless communication data is also in a significantly rising situation. However, as an indispensable carrier for constructing a new generation of information infrastructure, radio spectrum resources are increasingly scarce, and the problem of insufficient resource structure is increasingly highlighted.

The millimeter wave has a great development potential due to the advantages of sufficient spectrum resources, high spectrum efficiency and the like, but the characteristics of serious channel transmission path loss and the like bring a great number of technical problems to the actual implementation and deployment of millimeter wave communication. To overcome the severe path loss, it is necessary to use massive antenna arrays and beamforming on the transceiver to accommodate directional transmission in millimeter wave systems. In order to obtain a large beamforming gain, it is necessary to adaptively control and align transmission and reception beams of a Base Station (BS) and a User Equipment (UE). Given perfect channel knowledge, there are many optimized hybrid analog and digital BS/UE beamforming solutions under the constraints of different hardware resources. However, considering the large-scale antenna array employed, in millimeter wave communication, accurately estimating the channel (i.e., all elements of the channel matrix) by existing solutions is itself a difficult task. Another possible approach to beam alignment at millimeter waves is beam training through spatial scanning, where the base station and user jointly train base station/user beam forming pairs in a pre-designed codebook representing the beam search space to find their main path. As described above, the millimeter wave all-digital massive MIMO system will be the best choice for B5G and 6G, but its disadvantages of high cost, high complexity, large power consumption, etc. restrict its development and application.

In the related art, a concept of an asymmetric millimeter wave large-scale MIMO system capable of reducing the cost, complexity and power consumption of a millimeter wave all-digital multi-beam array is proposed: the full-digital multi-beam transmitting and receiving array is designed asymmetrically, and a large-scale full-digital multi-beam transmitting array and a small-scale full-digital multi-beam receiving array are adopted at a base station end, so that narrower transmitting multi-beams and wider receiving multi-beams are generated; the user terminal can not only keep the traditional symmetrical form, but also adopt the asymmetrical form. The asymmetric millimeter wave large-scale MIMO system has two design ideas: the first is that the transmitting and receiving array scale and the number of radio frequency channels are different, and the second is that the transmitting and receiving array is the same but the number of radio frequency channels is different.

However, the applicant finds that, unlike a symmetric MIMO system, the conventional beam alignment method requires uplink and downlink training separately for uplink and downlink in order to obtain uplink and downlink transmit-receive beam pairs under an asymmetric system, which requires higher training overhead.

Disclosure of Invention

The present application is directed to solving, at least to some extent, one of the technical problems in the related art.

Therefore, a first objective of the present application is to provide a method for aligning uplink and downlink beams of multiple users facing asymmetric millimeter wave large-scale MIMO, where the method uses full-digital multi-directional beams to perform beam training on multiple users simultaneously for a downlink, and completes beam alignment between a base station and multiple users with less training overhead and higher reliability; aiming at the uplink, by means of a deep learning method, under the condition of not consuming extra training overhead, the optimal transmission narrow beam of the uplink user end is directly deduced according to the observation signal of downlink training, and the downlink transmission beam of the base station is widened to generate an uplink receiving wide beam by utilizing the partial reciprocity of the uplink and downlink beam domains, so that the training overhead is reduced, and the training process is simplified.

The second purpose of the present application is to provide a multiuser uplink and downlink beam alignment apparatus facing asymmetric millimeter wave large-scale MIMO;

a third object of the present application is to propose a non-transitory computer-readable storage medium.

To achieve the above object, an embodiment of a first aspect of the present application is to provide a multi-user uplink and downlink beam alignment method for asymmetric millimeter wave massive MIMO, where the method includes the following steps:

constructing a multi-directional beam codebook, and constructing a full-digital multi-directional beam for detecting multiple directions through optimized code words in the codebook;

controlling the base station end to send the multi-directional beams, controlling the user end to receive the multi-directional beams by using the receiving beams, and calculating the matched filtering output of the receiving signals of the user end so as to perform multiple rounds of downlink beam training from the base station end to the user end;

controlling the user terminal to carry out beam judgment according to the received signals, and determining a target downlink transmitting-receiving beam pair with maximum receiving power from the received signals received in the multiple rounds of downlink beam training;

carrying out data processing on received signals received by the user side under different signal-to-noise ratios, generating training data for predicting narrow beams sent by the user side to the base station side in an uplink mode, and carrying out offline training on a preset convolutional neural network through the training data to obtain trained network parameters;

performing online real-time signal detection based on the trained network parameters and the received signals of the user side to predict target uplink transmission narrow beams from the user side to the base station side;

and controlling the user side to feed back the label of the target downlink transmission narrow beam of the base station end to the base station end, and widening the target downlink transmission narrow beam into a target uplink receiving beam of the base station end.

Optionally, in an embodiment of the present application, constructing a multi-directional beam codebook includes: constructing a minimum mean square error function for optimizing the multi-directional beam; and solving the minimum mean square error function through a Riemannian manifold optimization algorithm to obtain optimized code words emitted to different directions, and combining the code words to generate the codebook.

Optionally, in an embodiment of the present application, the minimum mean square error function is represented by the following formula:

wherein, s.t.w^Hw＝1，

Where, w is a single directional beam,in order to be a vector of the response of the array,for the departure angle, g is the ideal gain vector, L_TIs the size of the codebook, g, sent by the base station during downlink transmission^*Is the conjugate transpose of the ideal gain vector;

constructing an all-digital multi-directional beam for detecting multiple directions by the following formula:

wherein M [. C]Is the direction of the multi-directional beam, M is the number of directions of the multi-directional beam,represents the optimized codeword transmitted to either direction.

Optionally, in an embodiment of the present application, in each round of downlink beam training from the base station to the user terminal, the base station transmits multiple training symbols, and determines a combination of directions for collecting an observation signal in any training symbol, where the round of downlink beam training from the base station to the user terminal includes: in each round of downlink beam training, the base station end sends a preset number of pilot frequency sequences; determining a combined signal corresponding to any training symbol in the current round of downlink beam training received by the user terminal; and calculating the matched filtering output of the combined signal corresponding to any training symbol.

Optionally, in an embodiment of the present application, the processing data of the received signals received by the user terminal under different signal-to-noise ratios to generate training data for predicting that the user terminal transmits a narrow beam to the base station in an uplink includes:

combining the matched filtered outputs of the received signals received by the user terminal under different signal-to-noise ratios by the following formula:

wherein the content of the first and second substances,r is the training round, n is any user, k^*Is a receiving beam of a user terminal, L_RReceiving codebook size for a user terminal, L ═ L_T/M，L_TTransmitting the codebook size for the base station; sequentially normalizing the received signals of each round of downlink beam training and then combining the normalized received signals; and taking the real part and the imaginary part of the normalized signal and a receiving beam of the user end as the training data.

Optionally, in an embodiment of the present application, widening the target downlink transmission narrow beam into a base station target uplink reception beam includes: based on the asymmetric characteristic of the uplink and downlink beams, the central angle of the target uplink receiving beam of the base station end is kept consistent with the target downlink sending narrow beam, and the width of the target uplink receiving beam of the base station end is widened to be the preset beam width.

In order to achieve the above object, a second aspect of the present application further provides a multiuser uplink and downlink beam alignment apparatus for asymmetric millimeter wave massive MIMO, including the following modules:

the system comprises a construction module, a detection module and a control module, wherein the construction module is used for constructing a multi-directional beam codebook and constructing a full-digital multi-directional beam for detecting multiple directions through optimized code words in the codebook;

the training module is used for controlling the base station end to send the multi-directional beams, controlling the user end to receive the multi-directional beams by using the receiving beams, and calculating the matched filtering output of the receiving signals of the user end so as to carry out multi-round downlink beam training from the base station end to the user end;

a decision module, configured to control the user side to perform beam decision according to the received signal, and determine a target downlink transmit-receive beam pair with the maximum received power from the received signals received in the multiple rounds of downlink beam training;

the data processing module is used for carrying out data processing on the received signals received by the user side under different signal-to-noise ratios, generating training data for predicting narrow beam uplink transmission from the user side to the base station side, and carrying out offline training on a preset convolutional neural network through the training data to obtain trained network parameters;

the prediction module is used for carrying out online real-time signal detection on the basis of the trained network parameters and the received signals of the user side so as to predict the target uplink transmission narrow beam from the user side to the base station side;

and the broadening module is used for controlling the user side to feed back the label of the target downlink transmission narrow beam of the base station side to the base station side, and broadening the target downlink transmission narrow beam into a target uplink receiving beam of the base station side.

Optionally, in an embodiment of the present application, the building module is specifically configured to: constructing a minimum mean square error function for optimizing the multi-directional beam; and solving the minimum mean square error function through a Riemannian manifold optimization algorithm to obtain optimized code words emitted to different directions, and combining the code words to generate the codebook.

Optionally, in an embodiment of the present application, the building module is further configured to: the minimum mean square error function is expressed by the following formula:

wherein, s.t.w^Hw＝1，

Where, w is a single directional beam,in order to be a vector of the response of the array,for the departure angle, g is the ideal gain vector, L_TIs the size of the codebook, g, sent by the base station during downlink transmission^*Is the conjugate transpose of the ideal gain vector;

constructing an all-digital multi-directional beam for detecting multiple directions by the following formula:

wherein M [. C]Is the direction of the multi-directional beam, M is the number of directions of the multi-directional beam,represents the optimized codeword transmitted to either direction.

The technical scheme provided by the embodiment of the application at least has the following beneficial effects: aiming at a downlink, the method and the device use full-digital multi-directional beams to perform beam training on multiple users simultaneously, complete beam alignment of a base station and the multiple users with less training overhead and higher reliability, and overcome the defects of high overhead and low reliability of the traditional method; aiming at an uplink, by means of a deep learning method, under the condition of not consuming extra training overhead, an optimal sending narrow beam of an uplink user end is directly deduced according to an observation signal of downlink training, and a downlink sending beam of a base station is widened to generate an uplink receiving wide beam by utilizing partial reciprocity of uplink and downlink beam domains, so that the selection or generation of an uplink receiving and sending beam is completed, the training overhead is reduced, the training process is simplified, the uplink sending narrow beam and the receiving wide beam obtained by widening can reach better spectral efficiency through predicting the uplink sending narrow beam and the widening through a neural network, and the uplink transmission rate of an asymmetric system can be ensured under the condition of not needing extra overhead.

In order to implement the foregoing embodiments, an embodiment of the third aspect of the present application further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the asymmetric millimeter wave massive MIMO-oriented multi-user uplink and downlink beam alignment method in the foregoing embodiments.

Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.

Drawings

The foregoing and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:

fig. 1 is a flowchart of a multi-user uplink and downlink beam alignment method for asymmetric millimeter wave massive MIMO according to an embodiment of the present application;

fig. 2 is a flowchart of a specific asymmetric millimeter wave massive MIMO-oriented multi-user uplink and downlink beam alignment method according to an embodiment of the present application;

fig. 3 is a schematic view of an application scenario of a multi-user uplink and downlink beam alignment method for asymmetric millimeter wave massive MIMO according to an embodiment of the present application;

fig. 4 is a schematic diagram of beam training according to an embodiment of the present application;

fig. 5 is a schematic diagram of a downlink beam training misalignment rate performance curve according to an embodiment of the present application;

fig. 6 is a schematic diagram of another misalignment rate performance curve for downlink beam training according to an embodiment of the present application;

fig. 7 is a schematic diagram of a CDF performance curve for uplink transmission according to an embodiment of the present disclosure;

fig. 8 is a schematic diagram of another CDF performance curve for uplink transmission according to an embodiment of the present application;

fig. 9 is a schematic structural diagram of a multi-user uplink and downlink beam alignment apparatus for asymmetric millimeter wave massive MIMO according to an embodiment of the present application.

Detailed Description

Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.

The following describes a multi-user uplink and downlink beam alignment method and apparatus for asymmetric millimeter wave massive MIMO proposed in the embodiments of the present invention with reference to the accompanying drawings.

Fig. 1 is a flowchart of a multi-user uplink and downlink beam alignment method for asymmetric millimeter wave massive MIMO according to an embodiment of the present application, and as shown in fig. 1, the method includes the following steps:

step 101, constructing a multi-directional beam codebook, and constructing a full-digital multi-directional beam for detecting multiple directions through optimized codewords in the codebook.

The uplink and downlink beams in the embodiment of the present application include a downlink transmit-receive beam pair from the base station end to the user end, an uplink transmit narrow beam from the user end to the base station end, and an uplink receive beam from the base station end.

In the embodiment of the present application, downlink beam training from a base station to a user side is performed first, wherein a full-digital multi-directional beam corresponding to multiple directions of multiple users is constructed for an application scenario of the multiple users, so as to perform beam training for the multiple users at the same time.

The beam training codebook in the application adopts digital beam codewords, and can be optimally designed according to the center pointing angle and the beam width of each codeword.

In an embodiment of the present application, constructing the multi-directional beam codebook includes constructing a minimum mean square error function for optimizing the multi-directional beam, solving the minimum mean square error function through a riemann manifold optimization algorithm, obtaining optimized codewords transmitted to different directions, and combining the codewords to generate the multi-directional beam codebook.

It should be noted that the gain of an ideal beam is constant within its coverage angle, while the gain of the range outside its coverage angle is 0, and all the ideal beams with different central angles can cover the whole search space without overlapping. To approximate the beam constructed in the present application to an ideal beam, in one embodiment of the present application, the interval [ -1,1 ] may be set]Lattice points are formed by delta stepThen whenOtherwise g_j0. Further, the optimization problem is modeled using minimum mean square error, and in this embodiment, the minimum mean square error function is represented by the following formula:

wherein, s.t.w^Hw＝1，

Where, w is a single directional beam,in order to be a vector of the response of the array,for the departure angle, g is the ideal gain vector, L_TIs the size of the codebook, g, sent by the base station during downlink transmission^*Is the conjugate transpose of the ideal gain vector. It should be understood that solving the minimum mean square error is a riemann manifold optimization problem, and in the embodiment of the present application, the solving is performed by a riemann manifold optimization algorithmWhen solving the minimum mean square error function, as an example, the optimization and solution of the beam can be accomplished by a correlation algorithm of MATLAB.

Further, after solving the least mean square error, obtaining a plurality of optimized code words for useThe optimized codeword, representing transmission in direction α (j), constitutes simultaneous transmission in M directions (assumed to be M [1 ]],M[2],...,M[M]) The detected multi-directional beam can be expressed by the following formula in one embodiment of the present application:

wherein M [. C]Is the direction of the multi-directional beam, M is the number of directions of the multi-directional beam,represents the optimized codeword transmitted to either direction.

Therefore, the full-digital multi-directional beam for detecting multiple directions is constructed according to the optimized code words in the codebook.

And 102, controlling the base station end to send the multi-directional beams, controlling the user end to receive the multi-directional beams by using the receiving beams, and calculating the matched filtering output of the receiving signals of the user end so as to perform multiple rounds of downlink beam training from the base station end to the user end.

Specifically, in the embodiment of the present application, when performing downlink beam training, the base station performs multiple rounds of beam training from the base station to the user terminal, and when transmitting beams, the multi-directional beams constructed in step 101 are used to reduce the number of observations, and the user terminal receives beams using a wider beam with non-full precision.

In an embodiment of the present application, in each round of downlink beam training from a base station to a user terminal, the base station sends a plurality of training symbols, and determines a combination of directions in which an observation signal is collected in any training symbol, where the observation signal is a received signal received by the user terminal, that is, the application determines a combination of downlink sending directions of the base station corresponding to the received signal in each round of training, and in each round of downlink beam training, the base station sends a preset number of pilot sequences, and determines a signal of the combination corresponding to any training symbol in the round of downlink beam training received by the user terminal; a matched filtered output of the combined signal corresponding to any one of the training symbols is calculated.

Specifically, the present application first establishes a channel matrix between the user terminal and the base station terminal, and the established channel matrix between the base station and the user n can be represented by the following formula:

wherein M is_nFor the number of component paths, phi, from base station to user n_m,n,ψ_m,nRespectively, the departure angle and arrival angle, γ, of the mth component path from the base station to user n_m,nThe channel gain of the mth component path of the channel from the base station to the user n, v and u are respectively the sending end T_xAnd a receiving end R_xThe array response vector of (1).

And the number of detected directions satisfies M2^R，R∈Z，L_T＝ML。

The downlink training process of the base station end is as follows: this stage includes 1+ log₂And M rounds of beam training are carried out, and in each round, the BS transmits a plurality of training symbols. For each round r e {1,2, …, log₂M +1, and G (r, l) denotes a combination of directions (in ascending order) in which the observed signals are collected in the l-th training symbol. Assume a total training overhead of E_totBy K_rE_totTo represent the training overhead of the r-th round, and for K_r，r∈{1,2,…,log₂M +1} has:

wherein, each round of beam training process can be expressed as: sending a message containing n_sOne symbol pilot sequenceThe signal of the ith round and the ith combination received by user n using beam k can be expressed as:

further, the matched filter output is r_n,r,k,l＝y_n,r,k,ls^H＝Eh_n,r,k,l+z_n,r,k,l，

Wherein the content of the first and second substances,for the ith combined code word of the r round, f_n,kIs the k-th codeword for user n,for the effective channel after beamforming by the transmitting end and the receiving end,the energy required for training.Is a noise matrix.

The following is a detailed description of the first round and each subsequent round of beam training:

first, for r 1, the base station will train L multi-directional beams capable of simultaneously detecting M directions: g (1, L) { L, L + L, …, L + (M-1) L },in the downlink training process of the user terminal, the codebook L is received_RSmaller and therefore may be directly traversed.

Training all candidate beam pairs by the first wheel set, and observing L x L_RNext, each observation is assigned an overhead of K₁E_tot/(L*L_R) The matched filter output is:

r_n,1,k,l＝K₁E_toth_n,1,k,l/(L*L_R)+z_n,1,k,l，

and selecting the optimal beam pair, thereby obtaining the optimal receiving beam of the user terminal.

Second, at the subsequent log₂In the M rounds of beam training, the base station side helps each user to determine the optimal beam direction by using different combinations of the transmission beam directions, and the user side receives the optimal beam determined in the first round. For r e [2,3, …, log₂M+1]Each beam scanning is combined by L/2The composition, expressed as:

G(r,l)＝{[G(1,l)]_1:a(r),[G(1,l+L/2)]_a(r)+1:a(r),[G(1,l)]_{2a(r)+1:3a(r)},[G(1,l+L/2)]_{3a(r)+1:4a(r)},…,[G(1,l)]_{M-2a(r)+1:M-a(r)},[G(1,l+L/2)]_M-a(r)+1:M},

wherein a (r) M/(2)^r-1)。

And further determining the matched filtering output of the r round as:

r_n,r,k,l＝K_rE_toth_n,r,k,l/(L/2)+z_n,r,k,l。

therefore, multiple rounds of downlink beam training from the base station end to the user end are carried out by determining the combined signal corresponding to any training symbol in the current round of downlink beam training received by the user end and calculating the matched filtering output mode of the combined signal corresponding to any training symbol.

And 103, controlling the user side to carry out beam judgment according to the received signals, and determining a target downlink transmitting-receiving beam pair with the maximum receiving power from the received signals received in the multiple rounds of downlink beam training.

Specifically, after the beam training, each user can independently determine the best beam pair according to the receiving power of the received signal obtained by the user in the downlink training process, make a beam decision, and select the best downlink transmit-receive beam pair.

In one embodiment of the present application, let P be for any user n_n(r,k,l)＝||r_n,r,k,l||²Representing the power received by user n with beam k from the i-th combination of the r-th round of beam scanning,and a candidate set representing the optimal transmission beam direction of the base station end after the r-th beam scanning. For the first round and each subsequent round of beam training mode explained in step 102, the following explanation is made to correspondingly determine the target downlink transmit-receive beam pairs of each round:

first, for r ═ 1:

P_n(1,k,l)＝||r_n,1,k,l||²，

and then selecting the transmitting-receiving beam pair with the maximum receiving power:

wherein k is^*I.e. the best receiving beam of the user terminal,representing the multi-beam combination of maximum power after the first round of beam scanning.

And the candidate direction set of the best beam direction after the first round of beam scanning is:

wherein, defineIs shown andthere are combinations of common directions. The expected received power from the best direction of coverage/non-coverage for the multiple beams is approximated (ignoring receiver noise and inter-beam interference) respectivelyAnd 0. For judging G_n(r,l_n(r)) whether or not the binary decision threshold contains the received power of the best beam direction is:

that is, the present application may determine the best downlink transmission beam from the base station to the user from the candidate direction set by using the binary decision threshold.

Second, for subsequent log₂M wheels, r ∈ {2,3, …, log₂M +1, and the binary decision is compared withIn combination, a new set of round candidate directions may be determined as follows:

after obtaining the set of candidate directions, for subsequent logs₂In M rounds, each round can narrow the range of the candidate set according to the formula of the new candidate direction set until the target downlink transmitting-receiving beam pair with the maximum receiving power is determined finally.

And 104, processing data of the received signals received by the user terminal under different signal-to-noise ratios, generating training data for predicting narrow beams sent by the user terminal to the base station terminal in an uplink manner, and performing offline training on a preset convolutional neural network through the training data to obtain trained network parameters.

Specifically, after determining the optimal downlink transmit-receive beam pair from the base station to the user terminal, the prediction of the uplink transmission narrow beam from the user terminal to the base station is performed. During specific implementation, training data are generated firstly, a known channel matrix is generated according to a channel model, the optimal beam direction is determined, the base station end performs beam training according to the step 102, and the user end obtains received signals under different signal-to-noise ratios according to a millimeter wave mobile communication system model and obtains the training data after data processing.

In an embodiment of the present application, performing data processing on received signals received by a user terminal under different signal-to-noise ratios to generate training data for predicting uplink transmission of a narrow beam from the user terminal to a base station, includes: combining the matched filtering outputs of the received signals received by the user terminal under different signal-to-noise ratios by the following formula:

wherein the content of the first and second substances,r is the training round, n is any user, k^*Is a receiving beam of a user terminal, L_RReceiving codebook size for a user terminal, L ═ L_T/M，L_TTransmitting the codebook size for the base station; sequentially normalizing the received signals of each round of downlink beam training and then combining the normalized received signals; and taking the real part and the imaginary part of the normalized signal and a receiving beam of the user end as training data.

Specifically, for any user n, the matched filtered outputs of the received signals of the r-th round of user n are combined as:

because the output dynamic range of the matched filtering of the received signals is large, the received signals are complex signals and are not suitable for being directly used as input, the received signals of multiple rounds are combined after being normalized according to each round:

and the real part and the imaginary part of the normalized signal are used as two characteristic inputs of the designed neural network. In addition, the user end receives the wide beam k^*The prediction of its narrow beam also has a large impact, and therefore is also input as a feature:

the label is the true best transmitting beam direction of the user terminal. By varying the total training overhead E in the embodiments of the present application_totTraining data with different signal-to-noise ratios are obtained.

Further, a preset convolutional neural network is trained offline through the obtained training data, namely, a convolutional neural network structure based on data driving and set network parameters are set, and the generated training data is used for training the network offline to obtain trained network parameters.

As an example, the convolutional neural network main body is composed of a preprocessing module, a convolution module and an output module, in the training process, data with different signal-to-noise ratios are mixed and trained, when the loss function is not reduced any more, namely the prediction accuracy of the verification set is not improved any more, the training is stopped, and the training model is stored, so that the training process is completed on line.

And 105, performing online real-time signal detection based on the trained network parameters and the received signals of the user side to predict the target uplink transmission narrow beam from the user side to the base station side.

Specifically, the trained network parameters are stored for online real-time signal detection, and online real-time signal detection is performed according to the received signals of the user side, namely the observation signals received by the user side in the downlink beam training, so as to predict the uplink transmission narrow beams.

And 106, controlling the user side to feed back the label of the target downlink transmission narrow beam of the base station end to the base station end, and widening the target downlink transmission narrow beam into a target uplink receiving beam of the base station end.

It should be noted that, due to the asymmetric nature of the base station uplink and downlink beams, the base station downlink transmission beam is narrower than the uplink reception beam, and the optimal transmission beam at the base station end can already be determined in step 103, so the present application makes use of the partial reciprocity of the uplink and downlink beam domains to broaden the base station downlink transmission beam and generate the uplink reception wide beam.

In specific implementation, as an example, after the user feeds back the label of the target downlink transmission narrow beam to the base station, based on the asymmetric characteristic of the uplink and downlink beams, the user keeps the central angle of the target uplink reception beam at the base station consistent with the target downlink transmission narrow beam, and widens the width of the target uplink reception beam at the base station to a preset beam width. Therefore, the generation of the target uplink receiving beam at the base station end is realized.

To sum up, the method for aligning uplink and downlink beams of multiple users facing asymmetric millimeter wave large-scale MIMO according to the embodiment of the present application performs beam training on multiple users simultaneously by using all-digital multi-directional beams for a downlink, and completes beam alignment between a base station and multiple users with less training overhead and higher reliability, thereby overcoming the disadvantages of high overhead and low reliability of the conventional method; aiming at an uplink, by means of a deep learning method, under the condition of not consuming extra training overhead, an optimal sending narrow beam of an uplink user end is directly deduced according to an observation signal of downlink training, and a downlink sending beam of a base station is widened to generate an uplink receiving wide beam by utilizing partial reciprocity of uplink and downlink beam domains, so that the selection or generation of an uplink receiving and sending beam is completed, the training overhead is reduced, the training process is simplified, the uplink sending narrow beam and the receiving wide beam obtained by widening can reach better spectral efficiency through predicting the uplink sending narrow beam and the widening through a neural network, and the uplink transmission rate of an asymmetric system can be ensured under the condition of not needing extra overhead.

In order to more clearly illustrate a specific implementation process of the asymmetric millimeter wave massive MIMO-oriented multi-user uplink and downlink beam alignment method according to the embodiment of the present application, several implementation manners in a scene in practical application are described below.

Fig. 3 is a schematic view of an application scenario of a multi-user uplink and downlink beam alignment method for asymmetric millimeter wave massive MIMO according to an embodiment of the present application, where Tx is a signal transmitting end, for example, may be a base station end, Rx is a receiving end, and may include multiple user equipments UE, and Tx transmits a receiving end corresponding to a signal value through multiple channels. Considering the typical scene of millimeter wave outdoor mobile communication, one base station covers N users, and N users are equipped at the base station_TEach antenna has an independent radio frequency link, and the user side is N_RA root antenna. Time base station end sending codebook size L in downlink transmission_TThe receiving codebook size of the user terminal is L_RThe size of the receiving codebook at the base station end is L during uplink transmission_R', the size of the user side transmission codebook is L_T' (according to the characteristics of the asymmetric millimeter wave massive MIMO system, there is L_T>L′_R，L_T′>L_R)。

For example, consider an asymmetric millimeter wave communication system operating at 73GHz and having a coherence bandwidth of 100MHz, where a base station covers N {2,4,8} users. The base station is provided with 64 antennas, each antenna is provided with an independent radio frequency link, and the user side is provided with 8 antennas.

The application proposes two example scenarios, respectively: firstly, 64 antennas are used by a base station end in downlink transmission, the size of a sending codebook is 32, 4 antennas are used by a user end, the size of a receiving codebook is 4, 16 antennas are used by the base station end in uplink transmission, the size of the receiving codebook is 16, 8 antennas are used by the user end, and the size of the sending codebook is 8; secondly, 64 antennas are used by the base station end in downlink transmission, the size of the sending codebook is 32, 8 antennas are used by the user end, the size of the receiving codebook is 4, 64 antennas are used by the base station end in uplink transmission, the size of the receiving codebook is 16, 8 antennas are used by the user end, and the size of the sending codebook is 8. When the system is used for beam alignment, the application provides a specific multiuser uplink and downlink beam alignment method facing asymmetric millimeter wave large-scale MIMO, as shown in FIG. 2, the method comprises the following steps:

(1) downlink beam training from a base station end to a user end:

(1.1) constructing a multi-directional beam codebook:

by usingThe optimized codeword, representing transmission in direction α (j), is constructed in the following manner while going in 4 directions (assuming M [1 ]],M[2],...,M[4]) Multi-directional beam of detection:

(1.2) downlink beam training:

channel modeling: based on the measurements in the existing literature, the millimeter wave channel is modeled with a limited number of multipath components (from different aoas and aods). In line-of-sight scenarios, the channel is modeled as a rice channel with a main path, with the rice factor set to 13.2 dB.

Since each multi-directional beam simultaneously detects M-4 directions, a total of 1+ log is required₂M is 3 rounds (the multi-beam coverage direction of each round is shown in fig. 4), and the training overhead ratio of each round is {0.65,0.175,0.175}, respectively.

1) r is 1: the first round base station end needs to perform L-L traversing the whole space_TThe first set of multi-directional beam coverage directions is G (1, l) { l, l +8, …, l +24},(see fig. three)) The ue needs to traverse 4 directions, so that L × L is needed_RFor 32 observations, the matched filter output for each observation is:

r_n,1,k,l＝K₁E_toth_n,1,k,l/(L*L_R)+z_n,1,k,l，

select the best (| r)_n,1,k,lMaximum) beam pair, the corresponding base station end is the candidate set of the best transmit beam, and the corresponding user end is the best receive beam.

2) r is 2: the second round of base station side sends beam directions to recombine the first round of multi-directional beam directions, and the coverage direction of the first multi-directional beam is set as G (2, l) { [ G (1, l) { ]]_1:2,[G(1,l+4)]_3:4}，(see fig. three), it only needs to observe four times, covering half of the departure angle space, and the ue receives the best narrow beam in the first round of discrimination.

3) r is 3: the third round of base station side transmission beam direction is obtained by the first round of multi-directional beam direction recombination, and the coverage direction set of the first multi-directional beam is G (3, l) { [ G (1, l) { ]]₁,[G(1,l+4)]₂，[G(1,l)]₃,[G(1,l+44}，(see fig. three), observe four times as the second round, cover half of the angular space of departure, the user uses the best narrow beam that the first round differentiates to receive.

(1.3) beam decision:

after the beam training, each user can independently determine the optimal beam pair according to the receiving power of the user in the downlink training process.

1)r＝1：

P_n(1,k,l)＝||r_n,1,k,l||²，

Picking out the transmit-receive beam pair with the maximum receive power:

k^*i.e. the best receiving beam of the user terminal,representing the multi-beam combination of maximum power after the first round of beam scanning.

The candidate direction set of the best beam direction after the first round of beam scanning is:

wherein, defineIs shown andthere are combinations of common directions. The expected received power from the best direction of coverage/non-coverage for the multiple beams is approximated (ignoring receiver noise and inter-beam interference) respectivelyAnd 0. For judging G_n(r,l_n(r)) whether or not the binary decision threshold contains the received power of the best beam direction is:

2) r is 2: make binary decision withIn conjunction, user n may determine a new round candidate direction as follows:

3) r is 3: make binary decision withIn combination, user n can finally determine the best transmit beam direction of the best base station:

(2) and (3) predicting narrow beams sent by the user side to the base station side in an uplink mode:

(2.1) generating training data:

combining the matched filtered outputs of the received signals of the r-th round of the user n into:

because the output dynamic range of the matched filtering of the received signals is large, the received signals are complex signals and are not suitable for being directly used as input, the received signals of multiple rounds are combined after being normalized according to each round:

and the real part and the imaginary part of the normalized signal are used as two characteristic inputs of the designed neural network. In addition, the user end receives the wide beam k^*The prediction of its narrow beam also has a large impact, and therefore is also input as a feature:

the label is the true best transmitting beam direction of the user terminal. Varying the signal-to-noise ratio E_totAnd 200000 groups of data under each signal-to-noise ratio of 11dB to 16dB are collected in the training set, and 20000 groups of data under each signal-to-noise ratio of the verification set.

And (2.2) setting and training network parameters:

the network main body consists of three modules of preprocessing, convolution and output, the convolution module consists of four convolution layers and four ReLU active layers in an alternating mode, the convolution module is connected with a pooling layer, and the output module consists of a full connection layer and Softmax. And (3) mixing and training data with different signal-to-noise ratios, optimizing model parameters by using an Adam optimizer, stopping training when the loss function is not reduced any more (the prediction accuracy of the verification set is not improved any more), storing the training model, and completing the training process on line.

(2.3) online detection:

and storing the trained network parameters for online real-time signal detection, performing online real-time signal detection according to the received signals of the user side, and predicting the uplink transmission narrow beam.

(3) Generation of uplink receiving beams at a base station:

because the uplink and downlink beams of the base station are asymmetric (the downlink transmission beam of the base station end is narrower than the uplink reception beam), the optimal transmission beam of the base station end can be determined in the step (1), and after the user end feeds back the information to the base station, the base station keeps the central angle of the reception wide beam consistent with the transmission narrow beam, and the width is directly widened to the corresponding beam width.

In order to more clearly illustrate the technical effects of the asymmetric millimeter wave massive MIMO-oriented multi-user uplink and downlink beam alignment method, performance curves in the two example scenarios are calculated.

Fig. 5 and 6 are downlink beam training misalignment rate performance curves of an example scenario (i) and an example scenario (ii), respectively, and it can be found from fig. 5 and 6 that in both of these example scenarios, the performance of the training scheme (MBT) of the present application is superior to that of the baseline algorithm Hierarchical Search (HS), and when the number of users increases, the performance of the hierarchical search is further deteriorated.

Fig. 7 and 8 are diagrams illustrating CDF performance curves of uplink transmission achievable in an example scenario (i) and an example scenario (ii), respectively, and it can be obtained from fig. 7 and 8 that a neural network provided by the present application predicts that a received wide beam obtained by uplink transmission of a narrow beam and broadening can achieve better spectral efficiency, and can ensure an uplink transmission rate of an asymmetric system without additional overhead.

In order to implement the foregoing embodiments, the present application further provides a multiuser uplink and downlink beam aligning device for asymmetric millimeter wave massive MIMO, and fig. 9 is a schematic structural diagram of the multiuser uplink and downlink beam aligning device for asymmetric millimeter wave massive MIMO provided in the embodiments of the present application, and as shown in fig. 9, the device includes a constructing module 100, a training module 200, a deciding module 300, a data processing module 400, a predicting module 500, and a widening module 600.

The constructing module 100 is configured to construct a multi-directional beam codebook, and construct an all-digital multi-directional beam for detecting multiple directions through optimized codewords in the codebook.

The training module 200 is configured to control the base station to transmit the multi-directional beams, control the user to receive the multi-directional beams using the receive beams, and calculate a matched filtering output of the receive signal of the user to perform multiple rounds of downlink beam training from the base station to the user.

The decision module 300 is configured to control the user side to perform beam decision according to the received signal, and determine a target downlink transmit-receive beam pair with the maximum received power from the received signals received in the multiple rounds of downlink beam training.

The data processing module 400 is configured to perform data processing on received signals received by the user side under different signal-to-noise ratios, generate training data for predicting that the user side transmits a narrow beam to the base station side in an uplink manner, and perform offline training on a preset convolutional neural network through the training data to obtain trained network parameters.

The prediction module 500 is configured to perform online real-time signal detection based on the trained network parameters and the received signal of the user end, so as to predict an uplink narrow beam sent by the user end to the target of the base station end.

The broadening module 600 is configured to control the user side to feed back a label of a target downlink transmission narrow beam at the base station side to the base station side, and broaden the target downlink transmission narrow beam into a target uplink reception beam at the base station side.

Optionally, in an embodiment of the present application, the building module 100 is specifically configured to: constructing a minimum mean square error function for optimizing the multi-directional beam; and solving a minimum mean square error function through a Riemannian manifold optimization algorithm to obtain optimized code words emitted to different directions, and combining the code words to generate the codebook.

Optionally, in an embodiment of the present application, the building module 100 is further configured to:

the minimum mean square error function is expressed by the following formula:

wherein, s.t.w^Hw＝1，

Where, w is a single directional beam,in order to be a vector of the response of the array,for the departure angle, g is the ideal gain vector, L_TIs the size of the codebook, g, sent by the base station during downlink transmission^*Is the conjugate transpose of the ideal gain vector;

constructing an all-digital multi-directional beam for detecting multiple directions by the following formula:

wherein M [. C]Is the direction of the multi-directional beam, M is the number of directions of the multi-directional beam,represents the optimized codeword transmitted to either direction.

Optionally, in an embodiment of the present application, the training module 200 is further configured to control the base station to send a preset number of pilot sequences in each round of downlink beam training; determining a combined signal corresponding to any training symbol in the current round of downlink beam training received by the user terminal; a matched filtered output of the combined signal corresponding to any one of the training symbols is calculated.

Optionally, in an embodiment of the present application, the data processing module 400 is further configured to combine the matched filtered outputs of the received signals received by the user side under different signal-to-noise ratios according to the following formula:

wherein the content of the first and second substances,r is the training round, n is any user, k^*Is a receiving beam of a user terminal, L_RReceiving codebook size for a user terminal, L ═ L_T/M，L_TTransmitting the codebook size for the base station; sequentially normalizing the received signals of each round of downlink beam training and then combining the normalized received signals; and taking the real part and the imaginary part of the normalized signal and a receiving beam of the user end as training data.

Optionally, in an embodiment of the present application, the widening module 600 is further configured to keep a central angle of a target uplink receiving beam at the base station end consistent with the target downlink transmitting narrow beam based on an asymmetric characteristic of the uplink and downlink beams, and widen a width of the target uplink receiving beam at the base station end to a preset beam width.

To sum up, the device for aligning uplink and downlink beams of multiple users facing asymmetric millimeter wave large-scale MIMO according to the embodiments of the present application performs beam training on multiple users simultaneously by using all-digital multi-directional beams for a downlink, and completes beam alignment between a base station and multiple users with less training overhead and higher reliability, thereby overcoming the disadvantages of large overhead and low reliability of the conventional method; aiming at an uplink, by means of a deep learning method, under the condition of not consuming extra training overhead, an optimal sending narrow beam of an uplink user end is directly deduced according to an observation signal of downlink training, and a downlink sending beam of a base station is widened to generate an uplink receiving wide beam by utilizing partial reciprocity of uplink and downlink beam domains, so that the selection or generation of an uplink receiving and sending beam is completed, the training overhead is reduced, the training process is simplified, the uplink sending narrow beam and the receiving wide beam obtained by widening can reach better spectral efficiency through predicting the uplink sending narrow beam and the widening through a neural network, and the uplink transmission rate of an asymmetric system can be ensured under the condition of not needing extra overhead.

In order to achieve the above embodiments, the present application further proposes a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the asymmetric millimeter wave massive MIMO-oriented multi-user uplink and downlink beam alignment method according to any of the above embodiments.

In the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.

Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present application, "plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.

Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing steps of a custom logic function or process, and alternate implementations are included within the scope of the preferred embodiment of the present application in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present application.

The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.

It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.

It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.

In addition, functional units in the embodiments of the present application may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.

The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc. Although embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application, and that variations, modifications, substitutions and alterations may be made to the above embodiments by those of ordinary skill in the art within the scope of the present application.