URLLC pilot length and allocation optimization method based on differential evolution algorithm

Abstract

The application discloses a pilot length and allocation optimization method for ultra-reliable and low-latency communication (URLLC) based on a differential evolution algorithm, and comprises the following steps: converting a pilot length and pilot allocation joint optimization problem of the ultra-reliable and low-latency communication in a network-assisted full-duplex non-cell massive MIMO system into an optimization problem of maximizing a lower bound of a system ergodic data rate, and adopting a differential evolution algorithm based on an adaptive guide strategy to obtain an optimal solution of the optimization problem; firstly, encoding and initializing a population set of the pilot length and the pilot allocation scheme as optimization variables, and taking the lower bound of the system ergodic data rate as a fitness function to calculate fitness values of each individual in the population set; then, when the number of population iterations does not reach a set upper limit, performing a population evolution operation; finally, when the number of iterations reaches the upper limit, outputting a pilot length and a pilot allocation scheme represented by an optimal individual. Simulation results show that the application can effectively improve system performance.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of wireless communication, in particular to an ultra-reliable low-latency communication (URLLC) pilot length and allocation optimization method based on a differential evolution algorithm. BACKGROUND

[0002] Ultra-reliable low-latency communication is an important part of the fifth generation mobile communication technology and has been widely concerned in the fields of industrial automation and automatic driving. Network-assisted full-duplex (NAFD) has emerged in recent years and can complete data transmission and reception on the same time-frequency resource block. Compared with the traditional TDD / FDD mode, the spectrum efficiency of the mobile communication link can be doubled, and compared with the CCFD mode, the self-interference problem caused by full-duplex can be effectively alleviated, so that the system performance is effectively improved.

[0003] There are still many problems to be solved in implementing ultra-reliable low-latency communication in a network-assisted full-duplex non-cell large-scale MIMO system. Since ultra-reliable low-latency communication requires the system to complete data transmission within a given time threshold, the latency is proportional to the block length, that is, the user needs to complete channel estimation and data transmission within a given block length threshold. Different pilot allocation methods will cause different degrees of pilot pollution and interference problems, which will affect the performance of the system. Therefore, it is very important to jointly optimize the pilot length and the pilot allocation. SUMMARY

[0004] The purpose of the application is to provide an ultra-reliable low-latency communication (URLLC) pilot length and allocation optimization method based on a differential evolution algorithm to effectively improve the system performance.

[0005] To achieve this purpose, the application adopts the following technical scheme:

[0006] The ultra-reliable low-latency communication (URLLC) pilot length and allocation optimization method based on the differential evolution algorithm comprises the following steps:

[0007] Step 1, the pilot length and pilot allocation joint optimization problem of ultra-reliable low-latency communication in a network-assisted full-duplex non-cell large-scale MIMO system is converted into an optimization problem of maximizing the lower bound of the system sum ergodic data rate. Since the problem is non-convex, a differential evolution algorithm based on an adaptive guide strategy is used to obtain the optimal solution. First, the block length and allocation scheme in the optimization problem are encoded and initialized, and the lower bound of the system sum ergodic data rate is used as the fitness function to calculate the fitness value of each individual in the initial population set.

[0008] Step 2, when the number of iterations does not reach the upper limit, perform population evolution operations, including mutation, crossover and selection, in the mutation process, an adaptive guidance strategy is used to ensure that the evolution process is always in a favorable direction;

[0009] Step 3, when the number of iterations reaches the upper limit, output the pilot length and allocation scheme represented by the optimal individual.

[0010] In step 1, in a network-assisted full-duplex non-cell massive MIMO system, there are M remote antenna units (RAUs), wherein M U RAUs work in uplink reception mode, M D RAUs work in downlink transmission mode, all RAUs are distributed in the cell according to Poisson distribution and are connected to a central processing unit (CPU) through a low-delay backhaul link, each RAU is equipped with N antennas, there are K U uplink users and K D downlink users in the cell, and the total number of users is K.

[0011] In step 1, the fitness function is set according to the expression of the sum-ergodic rate lower bound of the network-assisted full-duplex non-cell massive MIMO system, and the fitness function fit(x a of individual x a is defined as:

[0012]

[0013] wherein, is the sum-ergodic rate lower bound of the i-th uplink user in the network-assisted full-duplex non-cell massive MIMO system, is the sum-ergodic rate lower bound of the j-th downlink user in the network-assisted full-duplex non-cell massive MIMO system, and individual x a =[x a (1), x a (2),..., x a (KL+1)], L is the number of pilots used by each user. Wherein is the pilot length used for channel estimation in ultra-reliable low-latency communication, when (i-1)L≤j≤iL+1, x a (j) is the serial number of the pilot used by the i-th user, wherein the elements in each position need to satisfy 1≤x a (j)≤x a (1), if x a (j) takes the value k, it means that the i-th user uses the k-th pilot for channel estimation.

[0014] In step 2, the mutation operator adopts an adaptive guiding strategy, selects the optimal individual and the worst individual in the population to guide the mutation of the individual in the current population, so that it can evolve in a favorable direction; the selection operator selects the individual with high fitness in the current population and the population after crossover to be inherited to the population of the next generation based on the fitness of the individual.

[0015] In step 3, it is checked whether the current iteration number i has reached the maximum iteration number L m If i < L m , return to step 2, otherwise, output the pilot length and pilot allocation scheme corresponding to the current optimal individual of the algorithm.

[0016] Beneficial effects: The method of the present application converts the pilot length and pilot allocation into a vector to be optimized, converts the joint optimization problem of the pilot length and pilot allocation into an optimization problem of maximizing the lower bound of the system ergodic data rate, and proposes a differential evolution algorithm based on an adaptive guiding strategy to solve the problem. The pilot length and pilot allocation method provided by the present application can effectively improve the system performance compared with the orthogonal pilot algorithm and the distributed estimation algorithm. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 The algorithm flowchart of the embodiment of the present application is shown in the figure;

[0018] Figure 2 The network-assisted full-duplex non-cell massive MIMO system schematic diagram of the embodiment of the present application is shown in the figure;

[0019] Figure 3 The system and ergodic data rate lower bound schematic diagram of the pilot length and allocation method based on the distributed estimation algorithm, the orthogonal pilot method and the pilot length and allocation method under different maximum allowed block lengths of the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0020] The present application will be further described below in combination with the drawings.

[0021] Firstly, the following is explained: considering a typical single-cell network-assisted full-duplex non-cell massive MIMO system, there are M remote antenna units (RAU) distributed, of which M U RAU works in the uplink receiving mode, M D RAU works in the downlink sending mode, all RAU positions are distributed in the cell according to the Poisson distribution and are connected with the central processing unit (CPU) through a low-delay backhaul link, each RAU is equipped with N antennas, and there are KU K D uplink users and K

[0022] Let the channel vector from the kth uplink user to all antennas of the lth uplink RAU be denoted as

[0023]

[0024] where denotes the large-scale fading coefficient from the kth uplink user to the lth uplink RAU, denotes the small-scale fading vector from the kth uplink user to the lth uplink RAU.

[0025] Let the channel vector from the kth downlink user to all antennas of the lth downlink RAU be denoted as

[0026]

[0027] where denotes the large-scale fading coefficient from the kth downlink user to the lth downlink RAU, denotes the small-scale fading vector from the kth downlink user to the lth downlink RAU.

[0028] Similarly, let the channel vector from the kth uplink user to all antennas of the lth downlink RAU be denoted as Let the channel vector from the kth downlink user to all antennas of the lth uplink RAU be denoted as

[0029] Let the channel gain from the ith uplink user to the jth downlink user be denoted as

[0030]

[0031] where and denote the large-scale fading and small-scale fading from the ith uplink user to the jth downlink user, respectively.

[0032] Since the network-aided full-duplex system can transmit and receive signals in the same time-frequency resource block, the signals transmitted by the downlink RAU will be received by the uplink RAU and thus cause self-interference problems. Let be the self-interference matrix between the ith uplink RAU and the jth downlink RAU, then

[0033]

[0034] ​​​In network-assisted full-duplex non-cellular massive MIMO systems, each data block can be divided into two parts: m c symbols are used for channel estimation, and m d symbols are used for data transmission. In order to meet the latency requirement of ultra-reliable low-latency communication, it is necessary to meet

[0035] In the channel estimation phase, all users send signals to the RAU. Since the number of orthogonal pilots is equal to the length of the pilot, there are m c orthogonal pilots, defined as and their set is denoted as P. Assuming that each user uses L pilots, the pilot sequence should satisfy and where P c represents the unit pilot power. For ease of description, define as the index of the pilot used by the kth uplink (downlink) user, and define as the index of the uplink (downlink) user using the kth pilot.

[0036] The pilot signal received by the lth uplink RAU can be represented as

[0037]

[0038] wherein represents a Gaussian white noise matrix, wherein each component obeys In order to estimate the channel The lth RAU multiplies the above formula by L / (m c P c )p m , and adds them to obtain

[0039]

[0040] wherein each component of n k,l =L / (m c P c )N l p m obeys The channel estimated by MMSE can be obtained as which is

[0041]

[0042] obeys wherein the expression is

[0043]

[0044] Similarly, the channel estimate with MMSE can be obtained as Subject to where The expression is

[0045]

[0046] In the data transmission phase, M U uplink RAUs receive data from K U uplink users, and M D downlink RAUs serve K D downlink users.

[0047] The received signal at the lthuplink RAU can be expressed as

[0048]

[0049] where, and denote the transmit power of the kthuplink and downlink user, respectively, and denote the transmit symbol of the kthuplink and downlink user, respectively. denotes the precoding vector for i downlink users at the jthdownlink RAU. denotes the Gaussian white noise vector, where each component is subject to

[0050] Similarly, the received signal at the jthdownlink user can be written as

[0051]

[0052] where n d denotes the downlink noise, which is subject to

[0053] In the ultra-reliable low-latency communication scenario, the lower bound of the ergodic achievable rate is approximated as

[0054]

[0055] where η = m c / m, m denotes the block length. The function f(x) is expressed as

[0056]

[0057] where Q -1 (x) is the inverse of the Gaussian Q function, defined as ​​

[0058] To obtain the signal of the ith uplink user, suppose that the signal received at the lth RAU is filtered by a filter

[0059]

[0060] wherein, is a filter for decoding the signal of the ith uplink user, is a channel vector of the ith uplink user to all uplink RAUs, is a self-interference matrix between all uplink RAUs and all downlink RAUs. For the ith uplink user, the first term in equation (14) is an expected signal, denoted as the second term is a leakage signal, denoted as the third term is interference caused by other uplink users, and interference caused by the kth uplink user is denoted as the fourth term is interference caused by the downlink RAU when sending signals to downlink users, and the inter-RAU self-interference caused by sending signals to the kth downlink user is denoted as the fifth term is noise.

[0061] On this basis, the signal-to-interference and noise radio (SINR) of the ith uplink user can be obtained as

[0062]

[0063] The decoding vector and the precoding vector used in the application satisfy the following relationship

[0064]

[0065]

[0066] In addition, for the convenience of description, define and as channel estimation error vectors. Since the estimation vector and the error vector of the channel are independent of each other, it can be obtained that

[0067]

[0068]

[0069]

[0070]

[0071]

[0072] ​According to the calculation result, for the i-th uplink user, there is

[0073]

[0074] wherein

[0075]

[0076] Similarly, for the i-th downlink user, there is

[0077]

[0078] wherein

[0079]

[0080] On this basis, the lower bound of the ergodic data rate of the i-th uplink (downlink) user is

[0081] The pilot length and pilot allocation have a direct impact on the data rate of the network-assisted full-duplex non-cell large-scale MIMO system. A reasonable pilot length and pilot allocation scheme is crucial to improve the data rate of the system. Therefore, the present application optimizes the pilot length and pilot allocation to maximize the lower bound of the ergodic data rate of the system. The joint optimization problem can be abstracted as the following mathematical problem

[0082]

[0083] Based on the analysis of the uplink and downlink data rates of the system, the URLLC pilot length and allocation optimization method based on the differential evolution algorithm provided by the present application can be divided into the following steps:

[0084] Step one, first encode the actual solution of the optimization problem shown in formula (27). Define the vector as an individual in the solution space, the individual x a =[x a (1),x a (2),...,x a (KL+1)], L is the number of pilots used by each user. The first element x a (1) is the pilot length used for channel estimation in ultra-reliable low-latency communication, and x a (j) is the serial number of the pilot used by the i-th user when (i-1)L≤j≤iL+1, wherein the element in each position needs to satisfy 1≤x a (j)≤x a (1), if x a (j) takes the value k, it means that the i-th user uses the k-th pilot for channel estimation.

[0085] Step two, set fitness function according to the system uplink and downlink spectral efficiency expression, define individual x a The fitness function fit(x a ) of the individual x

[0086]

[0087] Wherein, is the lower bound of the ergodic data rate of the ith uplink user in the network-assisted full-duplex non-cell large-scale MIMO system, is the lower bound of the ergodic data rate of the jth downlink user in the network-assisted full-duplex non-cell large-scale MIMO system.

[0088] Step three, set the population size N, crossover probability P c , the maximum iteration L m , and randomly initialize the initial population set X1, calculate the fitness value of each individual in the initial population;

[0089] Step four, if the current population iteration number does not reach the upper limit set, execute the population evolution operation, including mutation, crossover and selection three operation operators. In the evolution process, the adaptive guidance strategy is used to ensure that the algorithm can develop in a better direction, and at the same time, the algorithm has a faster convergence speed and is not easy to converge to a local optimal solution. The adaptive guidance strategy refers to that in each evolution process, the current best individual and the worst individual are selected to perform mutation operation on the individuals in the population, so that they develop in the direction of the best individual and away from the worst individual. If the current population iteration number has reached the upper limit set, go to step five;

[0090] Step five, output the pilot length and pilot allocation scheme corresponding to the current optimal individual.

[0091] The above only describes the preferred embodiments of the present application, and it should be noted that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be considered as the protection scope of the present application.

Claims

1. A method for optimizing the pilot length and allocation of URLLC based on differential evolution algorithm, characterized in that: The method comprises the following steps: Step 1, the pilot length and pilot allocation of the ultra-reliable and low-latency communication in the network-assisted full-duplex non-cell massive MIMO system are jointly optimized into an optimization problem of maximizing the lower bound of the system sum ergodic data rate, and a differential evolution algorithm based on an adaptive guide strategy is used to obtain an optimal solution; firstly, the pilot length and the pilot allocation scheme in the optimization problem are coded and a population set is initialized, and the lower bound of the system sum ergodic data rate is used as a fitness function to calculate the fitness value of each individual in the initial population set; Step 2, when the number of population iterations does not reach the upper limit, population evolution operations are performed, including three operation operators of mutation, crossover and selection, in the mutation process, the adaptive guide strategy is used to ensure that the evolution process is always in a favorable direction; Step 3, when the number of iterations reaches the upper limit, the pilot length and the pilot allocation scheme represented by the optimal individual are outputted; In step 1, in a network-assisted full-duplex non-cellular massive MIMO system, there are remote antenna units (RAUs), wherein some of the RAUs work in uplink reception mode, some of the RAUs work in downlink transmission mode, the positions of all the RAUs are distributed in a cell according to a Poisson distribution and communicate with a central processing unit (CPU) through a low-delay backhaul link, each RAU is equipped with antennas, there are uplink users and downlink users in the cell, and the total number of users is denoted as . In step 1, the fitness function is set according to the expression of the network-aided full-duplex non-cellular massive MIMO system and the lower bound of the ergodic data rate, and the individual is defined as: The fitness function of the individual is: ​ ； in, For network-assisted full-duplex non-cellular massive MIMO systems, the first The lower bound of the traversal data rate for each uplink user. For network-assisted full-duplex non-cellular massive MIMO systems, the first The lower bound of the traversal data rate for each downstream user, individual , The number of pilots used per user; where The pilot length used for channel estimation in ultra-reliable low-latency communication. hour, For the first The sequence number of the pilot used by each uplink user, where each element at each position needs to satisfy... ,like Values Then it means the first The first uplink user used the first Channel estimation is performed using pilot signals. 2.The method of ULLRC pilot length and allocation optimization based on differential evolution algorithm according to claim 1, characterized in that: In the step 2, the mutation operator adopts the adaptive guide strategy, the optimal individual and the worst individual in the population are selected to guide the mutation of the individuals in the current population, so that the individuals can evolve in a favorable direction; the selection operator selects the individuals with high fitness in the current population and the population after the crossover to be inherited to the population of the next generation based on the fitness of the individuals. 3.The method of ULLRC pilot length and allocation optimization based on differential evolution algorithm according to claim 1, characterized in that: In step 3, it is checked whether the current iteration round has reached the maximum iteration number If the current iteration round is less than , then the method returns to step 2, otherwise, the pilot length and allocation scheme corresponding to the current optimal individual of the algorithm are output.

Images