CN115087085A - Power distribution method for short packet communication superposition pilot frequency - Google Patents

Abstract

The invention discloses a power distribution method for short packet communication superposition pilot frequency, which comprises the following steps: 1) performing LMMSE channel estimation on the superposed pilot frequency (SP) signals; 2) eliminating pilot frequency interference in SP signals by utilizing the estimated channel and carrying out MRC data detection; 3) obtaining a lower boundary of a URLLC reachable speed expression of the SP signal by utilizing the statistical channel state information; 4) determining an optimal power allocation for the pilot and data by maximizing the weight sum rate to an optimization objective; 5) the optimization problem is solved by an iterative algorithm. The power distribution method provided by the invention enables the short packet transmission based on the superimposed pilot frequency to have higher transmission rate than that based on the traditional pilot frequency, and can more accurately depict the transmission time delay.

Description

A power allocation method for superimposed pilot frequency for short packet communication

technical field

The invention belongs to the field of communication resource allocation, and in particular relates to a power allocation method for superimposed pilot frequency for short packet communication

Background technique

With the explosive growth of IoT devices with stringent latency and reliability requirements, ultra-reliable low-latency communication (URLLC) has become an important part of fifth-generation mobile communications (5G) and 6G. URLLC is one of the most innovative technical solutions in 5G mobile networks. In practical applications, the reliability requirements of factory automation and remote surgery are 1-10 ^-9 , and the end-to-end delay is less than 1ms. Other services, such as smart grids, intelligent transportation systems and process automation, have more relaxed requirements for reliability. The 3rd Generation Partnership Project (3GPP) proposes a general URLLC requirement as follows: to transmit a data packet with a size of 32 bytes, with a reliability of 99.999% (block error rate (BLER) of 10 ^-5 ) and a delay of 1 ms. URLLC relies on short packet transmission or finite block length transmission, where the encoding length of the data or the number of times the channel is used is very limited. Therefore, the transmission is no longer error-free, and Shannon's formula is no longer applicable. In addition, because the current frame structure based on regular pilot (RP) transmits pilot and data separately, when the data packet is a short packet, the overhead caused by the pilot is not negligible, which will make the transmission efficiency of the system significantly decreased. Although some studies have improved the transmission rate in finite block lengths through resource allocation, the performance improvement is very limited.

As an alternative to RP, superimposed pilot (SP) has also attracted extensive attention in recent years. In SP, pilot and data are superimposed and transmitted on the same frequency. Compared with RP, SP does not need to reserve additional block length for pilot, so pilot can have the same sequence length as data, which not only improves the quality of channel estimation, but also improves transmission efficiency. In addition, using SP can also increase the number of orthogonal pilots to reduce pilot pollution. However, SP will inevitably degrade the channel estimation quality due to the introduction of data interference, which limits the performance of the system. The usual solution is to suppress interference by finding a fixed pilot data power allocation scaling factor for users to maximize the total throughput, but this scaling factor is not necessarily optimal for each user. Similarly, pilot symbols are also regarded as interference when performing data detection. Since pilot symbols are known at the base station side, they can be subtracted from the received signal, but since the channel estimation is imperfect, pilot symbols interference cannot be completely eliminated.

SUMMARY OF THE INVENTION

In order to solve the technical problems mentioned in the above background art, the present invention proposes a power allocation method for superimposed pilot frequency for short packet communication, which provides the optimal SP power allocation for all URLLC users in a single cell, thereby improving the The channel estimation quality is better than the traditional RP transmission. In order to realize the above-mentioned technical purpose, the technical scheme of the present invention is:

A power allocation method for superimposed pilot frequency for short packet communication, comprising the following steps:

Step 1: Perform LMMSE channel estimation on the superimposed pilot SP signal Y, and obtain the channel between the kth user and the base station from the K users

Step 2: Use the estimated channel to eliminate pilot interference in the SP signal and perform MRC data detection to obtain an estimate of the data sent by the kth user

Step 3: Obtain the statistical channel state information by measuring the propagation environment, and use the statistical channel state information to obtain the lower bound of the URLLC reachable rate expression of the SP signal

Step 4: Construct the weighted sum rate maximization problem to maximize the URLLC achievable rate weighted sum as the optimization goal;

Step 5: The weighted sum rate maximization problem is transformed into a geometric programming problem to solve through an iterative algorithm, and the optimal power allocation of all user pilots and data is obtained.

Preferably, the first step is as follows: K single-antenna users are randomly and uniformly distributed in a single cell, share the same system bandwidth B, and send superimposed pilot signals to the multi-antenna base station in the center of the cell in TDD mode, and the formula is expressed as follows :

和s_i分别表示第i个用户的正交导频序列和数据序列，两者具有相同的长度且叠加在一起同时同频进行传输，H表示共轭转置；Among them, q _i represents the pilot power allocated to the ith user, _pi represents the data power allocated to the _ith user, hi is the channel between the ith user and the base station, N is the superimposed white Gaussian noise matrix,

and s _i represent the orthogonal pilot sequence and data sequence of the i-th user, respectively, both of which have the same length and are superimposed together for transmission at the same frequency, and H represents the conjugate transpose;

The base station obtains the despread signal y _k after despreading the received signal, as shown in the following formula:

表示第k个用户的正交导频序列；对y_k信号进行LMMSE信道估计，如下式所示：Among them, τ _c represents the length of the transmission block, q _k represents the pilot power allocated to the k-th user, h _k represents the channel between the k-th user and the base station,

Represents the orthogonal pilot sequence of the kth user; performs LMMSE channel estimation on the _yk signal, as shown in the following formula:

Among them, β _k and β _i represent the large-scale fading coefficients of the kth and ith users, respectively.

Preferably, in the step 1, the superimposed pilot signal Y is transmitted in the form of a short data packet with a finite block length, and the achievable rate R _k of the kth user in the finite block length transmission is expressed as:

Among them, V _k =1-(1+γ _k ) ^-2 is the channel dispersion, ε is the decoding error probability, γ _k is the instantaneous signal-to-interference noise ratio of the kth user, and Q ^-1 (·) represents the Gaussian function inverse function of .

Preferably, in the second step, the data estimation of the kth user is obtained through MRC detection, and the formula is expressed as follows:

表示估计出的第k个用户的信道，q_i、p_i分别表示分配给第i个用户的导频功率和数据功率，q_k、p_k分别表示分配给第k个用户的导频功率和数据功率，

表示第i个用户的正交导频序列，h_k和h_i分别表示基站与第k个用户和第i个用户间的信道，s_k和s_i分别表示第k个用户和第i个用户的数据序列，N表示叠加高斯白噪声矩阵，||·||表示欧式范数，H表示共轭转置，E{·}表示取期望值，τ_c表示传输块的长度，σ_k的表达式为：in,

represents the estimated channel of the kth user, q _i and p _i represent the pilot power and data power allocated to the ith user, respectively, q _k , p _k represent the pilot power and the data power allocated to the kth user, respectively data power,

represents the orthogonal pilot sequence of the _i -th user, h _k and hi represent the channel between the base station and the k-th user and the i-th user, respectively, _sk and _si represent the k-th user and the i-th user, respectively The data sequence of , N represents the superimposed white Gaussian noise matrix, ||·|| represents the Euclidean norm, H represents the conjugate transpose, E{·} represents the expected value, τ _c represents the length of the transmission block, and the expression of σ _k for:

β _k and β _i represent the large-scale fading coefficients of the kth and ith users, respectively.

Preferably, in the step 3, the lower bound of the URLLC reachable rate expression is shown in the following formula:

Among them, E{·} represents the expected value, γ _k is the instantaneous signal-to-interference-noise ratio of the kth user, and the function f _k (·) is a monotonically decreasing convex function, and its expression is:

表示第k个用户在导频干扰不能被完全消除时的有效信干噪比，其表达式为：where x is the function argument, ε is the decoding error probability, τ _c is the length of the transmission block, Q ^-1 ( ) is the inverse function of the Gaussian function,

Represents the effective signal-to-interference-noise ratio of the kth user when the pilot interference cannot be completely eliminated, and its expression is:

M表示基站天线数，β_k、β_i分别表示第k个和第i个用户的大尺度衰落系数，p_k、p_i分别表示第k个和第i个用户的数据功率，q_k、q_i分别表示第k个和第i个用户的导频功率。in,

M represents the number of base station antennas, β _k and β _i represent the large-scale fading coefficients of the k th and i th users, respectively, p _k , p _i represent the data power of the k th and i th users, respectively, q _k , q _i represents the pilot power of the kth and ith users, respectively.

Preferably, in the fourth step, while maximizing the weighted sum of the URLLC achievable rates, all users must meet the minimum rate requirement R ^req and the maximum power constraint P _max , and the weighted sum rate maximization problem is expressed as follows:

表示URLLC可达速率表达式下界，w_k表示第k个用户的权重，p_k和q_k分别表示分配给第k个用户的数据功率和导频功率。in,

represents the lower bound of the URLLC achievable rate expression, w _k represents the weight of the kth user, and p _k and q _k represent the data power and pilot power allocated to the kth user, respectively.

The beneficial effects brought by the above technical solutions:

(1) This method can ensure that the SP power allocation of all users is optimal, thereby maximally suppressing the interference caused by the inaccuracy of channel estimation, and achieving a higher transmission rate;

(2) This method is suitable for more practical scenarios, that is, scenarios of limited block length transmission, and can accurately describe the delay and reliability in URLLC;

(3) This method performs power allocation based on the user's large-scale fading coefficient, and the change of the large-scale fading is very slow, so the calculation delay can be greatly reduced.

Description of drawings

Fig. 1 is a system model diagram of a single cell URLLC uplink scenario;

2 is a flow chart of a power allocation method for superimposing pilots for short packet communication;

3 is a structural diagram of a conventional pilot (RP) and a superimposed pilot (SP);

FIG. 4 is a diagram showing the relationship between a system and a rate and the number of transmit antennas provided by an embodiment of the present invention;

FIG. 5 is a relationship diagram of a system and rate and the number of system blocks provided by an embodiment of the present invention;

FIG. 6 is a relationship diagram of a system and a rate and the number of users according to an embodiment of the present invention.

Detailed ways

The technical solutions of the present invention will be described in detail below with reference to the accompanying drawings.

The embodiment of the present invention provides a power allocation method for short-packet communication superimposed pilot frequency, which provides optimal SP power allocation for all URLLC users in a single cell, thereby improving the quality of channel estimation, and obtaining better performance than traditional ones. RP transport for better performance.

The system model of this example is shown in Figure 1. It is assumed that in a single cell, there is a base station equipped with M antennas in the center of the cell, K users with single antennas are randomly distributed in the cell, and the channel between users and the base station is block fading , and obey the complex Gaussian distribution of CN(0, β _{k IM} ), where β _i represents the large-scale fading coefficient of the ith user. All users send short packet data to the base station (limited block length transmission), and the achievable rate expression of the kth user in finite block length transmission is:

where τ _c =BT represents the length of the transport block, ε is the decoding error probability, V _k =1-(1+γ _k ) ^-2 , and the data packet transmission time T is less than the channel coherence time.

Based on the above system and in conjunction with FIG. 2 , a power allocation method for short-packet communication superimposed pilot frequency provided by an embodiment of the present invention is described in detail, and the method includes the following steps:

Step 201: As shown in FIG. 3, an embodiment of the present invention provides a schematic diagram of a frame structure of a regular pilot (RP) and a superimposed pilot (SP). In RP, pilot frequency and data are transmitted separately; in SP, pilot frequency and data are superimposed together and transmitted on the same frequency and have the same length. All users in the cell share the same system bandwidth B, and send superimposed pilot signals to the multi-antenna base station in the center of the cell in TDD mode:

和s_i分别表示第i个用户的正交导频序列和数据序列，两者具有相同的长度且叠加在一起同时同频进行传输，H表示共轭转置。基站在对接收信号进行解扩后获得：Among them, q _i represents the pilot power allocated to the ith user, _pi represents the data power allocated to the _ith user, hi is the channel between the ith user and the base station, N is the superimposed white Gaussian noise matrix,

and s _i respectively represent the orthogonal pilot sequence and data sequence of the i-th user, both of which have the same length and are superimposed together and transmit on the same frequency at the same time, and H represents the conjugate transpose. The base station obtains after despreading the received signal:

Then perform LMMSE channel estimation on the signal to obtain the channel between the kth user and the base station

At this stage, the data is treated as interference that affects the quality of the channel estimation.

Step 202: Use the estimated channel to eliminate pilot interference in the SP signal and perform MRC data detection to obtain an estimate of the data sent by the kth user

表示估计出的第k个用户的信道，

由于信道估计是不完美的，所以导频干扰不能被完全消除，残留的导频干扰会影响数据的估计质量。in

represents the estimated channel of the kth user,

Since the channel estimation is imperfect, the pilot interference cannot be completely eliminated, and the residual pilot interference will affect the estimation quality of the data.

以下将对步骤203的具体流程作详细说明：Step 203: Obtain the lower bound of the URLLC reachable rate expression of the SP signal by using the statistical channel state information

The specific process of step 203 will be described in detail below:

(1) First calculate the effective signal-to-interference-noise ratio of the user. Statistical channel state information refers to the large-scale fading information of the channel, including path loss and shadow fading effect, thus obtaining the effective signal-to-interference-noise ratio of user k when the pilot interference cannot be completely eliminated:

in

(2) The lower bound of the URLLC achievable rate expression has nothing to do with the small-scale fading of the channel, and its lower bound is:

where the expression of the function f _k ( ) is:

is a monotonically decreasing convex function

其中速率是指URLLC可达速率，p_i表示分配给第i个用户的数据功率，q_i表示分配给第i个用户的导频功率，同时所有用户必须满足最小速率要求R^req和最大功率约束P_max，该加权和速率优化问题P1表示为：Step 204: Determine the optimal power allocation for each user pilot and data by maximizing the weighted sum rate as the optimization goal

The rate refers to the URLLC achievable rate, pi represents the data power allocated to the _ith user, qi represents the pilot power allocated to the _ith user, and all users must meet the minimum rate requirement R ^req and the maximum power constraint P _max , the weighted sum rate optimization problem P1 is expressed as:

P1:

是关于

的递增函数，该优化问题可转化为P2：An auxiliary variable χ _k is introduced, since

its about

, the optimization problem can be transformed into P2:

P2:

in,

Step 205: The optimization problem is transformed into a geometric programming (GP) problem through an iterative algorithm to solve. The specific process of step 205 will be described in detail below:

(1) Use the log function to approximate the objective function to obtain its lower bound:

是在第n+1次迭代中用于近似目标函数的参数，

和

为

的函数，其表达式为：in

are the parameters used to approximate the objective function in the n+1th iteration,

and

for

function whose expression is:

Replace the objective function of P2 with this lower bound and omit the constant term to get the new objective function:

Simplify further to get

是在第n次迭代中所有用户的最优功率分配集合，有效信干噪比

中c_i的分母是一个正项式，其表达式为：(2) Constraints that do not conform to the GP standard form are dealt with by the method of continuous convex approximation. The main idea is to approximate the positive term by constructing a series of monomial functions. Assumption

is the optimal power allocation set for all users in the nth iteration, the effective signal-to-interference-noise ratio

The denominator of c _i is a positive term, and its expression is:

In the n+1th iteration, this polynomial can approximate the following monomial:

和

的表达式为in,

and

The expression is

Thus P2 can be transformed into the standard form of the GP problem:

P3:

求解该问题的具体过程为：利用CVX的Mosek求解器对P3进行求解，获得第k个用户在第n+1次迭代的最优功率分配

利用该功率分配更新参数

并计算P2的目标函数值Obj⁽ⁿ⁺¹⁾，同时迭代次数加1，重复上述过程直到Obj⁽ⁿ⁺¹⁾收敛。in,

The specific process of solving this problem is: use the Mosek solver of CVX to solve P3, and obtain the optimal power distribution of the kth user at the n+1th iteration

Update parameters with this power allocation

And calculate the objective function value Obj ⁽ⁿ⁺¹⁾ of P2, at the same time increase the number of iterations by 1, and repeat the above process until Obj ⁽ⁿ⁺¹⁾ converges.

The technical scheme provided by the present invention is further elaborated below by specific embodiment:

和

Assuming that the number of iterations is n=1 and the error threshold ξ is 10 ^-5 , the initial value of the optimal power distribution is obtained by solving the following optimization problem

and

P4:

和

计算

和P2的目标函数初始值Obj⁽⁰⁾，否则，令Obj⁽⁰⁾＝0，并重新上述过程。无线通信网络参数的初始化如下：Using the same method as in P1, the above optimization problem can be transformed into a GP problem for solving. When θ≥1, use

and

calculate

and the initial value of the objective function of P2 Obj ⁽⁰⁾ , otherwise, let Obj ⁽⁰⁾ = 0, and repeat the above process. The initialization of wireless communication network parameters is as follows:

This example is a special case of the embodiment of the present invention, and can be extended to other similar cases.

Figure 4, Figure 5 and Figure 6 show the relationship between the sum rate and the number of antennas, block length and actuators of the central controller under long packet transmission (Shannon rate) and short packet transmission (URLLC rate) for RP and SP, respectively. It can be seen that under the optimal power allocation, even if the pilot interference is imperfectly eliminated, the performance of SP is still better than that of RP, and the perfect elimination of pilot interference is the ideal situation of SP, which can be regarded as its upper bound. in addition,

The performance at the Shannon rate is better than the URLLC rate, which means that in short packet transmission, we should use the URLLC reachable rate formula instead of the Shannon formula for transmission design, otherwise the transmission delay and reliability will be underestimated.

The embodiment is only to illustrate the technical idea of the present invention, and cannot limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention all fall within the protection scope of the present invention. .

Claims (6)

1. A power distribution method for short packet communication superimposed pilot is characterized by comprising the following steps:

the method comprises the following steps: LMMSE channel estimation is carried out on the superposed pilot SP signal Y, and the channel between the kth user and the base station is obtained from the K users

Step two: eliminating pilot frequency interference in SP signals by using the estimated channel and carrying out MRC data detection to obtain the estimation of data sent by the kth user

Step three: tong (Chinese character of 'tong')The statistical channel state information is obtained after the propagation environment is measured, and the URLLC reachable rate expression lower bound of the SP signal is obtained by utilizing the statistical channel state information

Step four: constructing a weighting and rate maximization problem by taking the maximized URLLC reachable rate weighted sum as an optimization target;

step five: and converting the weighting and rate maximization problem into a geometric programming problem through an iterative algorithm to solve, so as to obtain the optimal power distribution of all user pilot frequencies and data.

2. The method for allocating power to superposed pilots for short packet communication according to claim 1, wherein the first step specifically comprises: k single-antenna users are randomly and uniformly distributed in a single cell, share the same system bandwidth B, and send superposed pilot signals to a multi-antenna base station in the center of the cell in a TDD mode, wherein the formula is expressed as follows:

wherein q is _i Indicating the pilot power, p, allocated to the ith user _i Indicates the data power, h, allocated to the ith user _i Is the channel between the ith user and the base station, N is the superimposed gaussian white noise matrix,

and s _i Orthogonal pilot sequence and data sequence respectively representing the ith user, which have the same length and are superposed together for simultaneous co-frequency transmission, ^H represents a conjugate transpose;

the base station obtains a de-spread signal y after de-spreading the received signal _k As shown in the following formula:

wherein, tau _c Denotes the length of the transport block, q _k Denotes the pilot power, h, allocated to the k-th user _k Indicating the channel between the kth user and the base station,

an orthogonal pilot sequence representing a k-th user;

for y _k The signal is subjected to LMMSE channel estimation as shown in the following formula:

wherein, beta _k 、β _i Representing the large-scale fading coefficients of the kth and ith users, respectively.

3. The method as claimed in claim 2, wherein in step one, the superimposed pilot signal Y is transmitted in short data packets with limited block length, and the achievable rate R of the kth user in the limited block length transmission is _k The expression is as follows:

wherein, V _k ＝1-(1+γ _k ) ^-2 Is the channel dispersion, ε is the decoding error probability, γ _k Is the instantaneous signal to interference plus noise ratio, Q, of the k-th user ^-1 (. cndot.) represents the inverse of a Gaussian function.

4. The method as claimed in claim 1, wherein in the second step, the data estimation of the kth user is obtained through MRC detection, and the formula is expressed as follows:

wherein,

representing the estimated channel of the kth user, q _i 、p _i Respectively representing pilot power and data power, q, allocated to the ith user _k 、p _k Respectively representing the pilot power and data power allocated to the kth user,

orthogonal pilot sequence, h, representing the ith user _k And h _i Respectively representing the channels, s, between the base station and the k and i users _k And s _i Respectively representing data sequences of a kth user and an ith user, N representing a superimposed white Gaussian noise matrix, | | · | | representing a Euclidean norm, ^H denotes the conjugate transpose, E {. is } denotes the expectation, τ _c Indicating the length of the transport block, σ _k The expression of (a) is:

β _k 、β _i representing the large-scale fading coefficients of the kth and ith users, respectively.

5. The method for allocating power to superposition pilot for short packet communication according to claim 1, wherein in the third step, the lower bound of the expression of the URLLC achievable rate is shown as follows:

wherein E {. is expressed as an expectation value, γ _k Function f for the instantaneous SINR of the kth user _k (. cndot.) is a monotonically decreasing convex function expressed as:

where x is the function argument,. epsilon.is the decoding error probability,. tau. _c Indicating the length of the transport block, Q ^-1 (. cndot.) represents the inverse of a Gaussian function,

represents the effective signal-to-interference-and-noise ratio of the kth user when the pilot interference can not be completely eliminated, and the expression is:

wherein,

m represents the number of base station antennas, beta _k 、β _i Respectively representing the large-scale fading coefficients, p, of the kth and ith users _k 、p _i Respectively representing the data power of the k and i users, q _k 、q _i Indicating the pilot power of the kth and ith users, respectively.

6. The method of claim 1, wherein in step four, the minimum rate requirement R must be met by all users while maximizing URLLC sum of achievable rates ^req And maximum power constraint P _max The weighted sum rate maximization problem is expressed as followsShowing:

wherein,

represents the lower bound of the URLLC reachable Rate expression, w _k Weight, p, representing the kth user _k And q is _k Respectively representing the data power and pilot power allocated to the kth user.

Images