Resource allocation method for data security transmission in high-reliability low-latency internet of things

Abstract

The application is a resource allocation method for data security transmission in high-reliability and low-latency Internet of Things, comprising the following steps: step 1: constructing an uplink system supporting ultra-reliable and low-latency communication; step 2: constructing an optimization problem based on the sum of the QoE of the communication between all the ground nodes and the base station to maximize the security data rate; step 3: constructing an optimization model based on game theory, modeling the optimization problem in step 2 as a game problem with QoE evaluation strategy; step 4: using a distributed learning algorithm to iteratively obtain the resource allocation strategy between the ground nodes and the base station until the optimization problem in step 2 reaches the optimal value, and the channel resource allocation method is determined. The application approximately transforms the original complex problem into an easy-to-handle problem, and proposes an efficient distributed iterative algorithm to converge to the NE solution.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of wireless communication, and particularly relates to a resource allocation method for data security transmission for realizing ultra-reliable and low-latency communication (URLLC) services in a 6G Internet of Things (IoT) uplink network. BACKGROUND

[0002] Since the 1980s, with the continuous progress of wireless communication technology, a development process from fixed to mobile, analog to digital, circuit switching to cloud network integration, narrowband to wideband, and now to the Internet of Everything has been experienced, which continuously provides new impetus for the progress of human society and the development of economy, and drives the development of the entire related industry. However, with the increasing number of service access points and the continuous improvement of related service demand, the communication industry is urged to continue to meet new challenges. With the advent of the 5G era, 5G is integrated with technologies such as artificial intelligence, cloud computing, big data, Internet of Things / industrial Internet of Things, and edge computing, and has a huge impact in various industries.

[0003] The concept of ultra-reliable and low-latency communication (URLLC) comes from the classification of 5G service types by 3GPP: enhanced mobile broadband (eMBB), such as augmented reality / virtual reality (AR / VR); massive machine type communication (mMTC), which is mainly used in machine-centered scenarios; and ultra-reliable and low-latency communication (URLLC), which is mainly used in scenarios where people and machines are the center of communication, also known as critical machine type communication. Since the delay and reliability requirements of the URCR communication scenario are strict, it is often used in vehicle-to-vehicle communication with safety requirements, wireless control of industrial equipment, remote surgery, and large-scale virtual reality games with ultra-low latency. These scenarios do not cover all possible application cases, but only provide a general situation that can be foreseen. However, the current 5G access capabilities cannot well solve the problem in some cases, and the new wireless interface must have higher flexibility to support new application environments that may appear in the future, so a large amount of research has been done in both academia and industry. The IDC FutureScape report states that by 2024, 40% of cities will integrate the physical world and the digital world through the Internet of Things (IoT) and artificial intelligence, and improve the remote management of key infrastructure and digital services, thereby realizing the development of smart cities. Ultra-reliable and low-latency communication (URLLC) is an important standard for supporting 6G-based Internet of Things networks, which uses short data packet communication to meet its strict requirements for reliability and delay. SUMMARY

[0004] In order to solve the above problems, the application designs an uplink Internet of Things network end-to-end communication resource allocation strategy, so as to meet the ultra-reliable low-latency communication service quality requirements and improve the ultra-reliable low-latency communication service quality in the Internet of Things.

[0005] In order to achieve the above purpose, the application is realized by the following technical scheme:

[0006] The application is a resource allocation method for data security transmission in high-reliability low-latency Internet of Things, which comprises the following steps:

[0007] Step 1: Construct an uplink system supporting ultra-reliable low-latency communication, which comprises a plurality of communication base stations, ground nodes and communication eavesdroppers, each of which comprises a plurality of ground nodes and a service eavesdropper within the communication range of the communication base station;

[0008] Step 2: Construct an optimization problem based on the sum of the security data rate-based QoE of communication between all ground nodes and base stations.

[0009] The optimization function in step 2 is:

[0010]

[0011] Wherein:

[0012]

[0013]

[0014] Where C represents the Shannon information capacity under infinite block length, V represents the channel dispersion, ∈ represents the decoding error probability, Q -1 represents the inverse function of the Gaussian Q function, represents the signal-to-interference-and-noise ratio (SINR) between the ground node m b and the base station b, represents the signal-to-interference-and-noise ratio of the eavesdropper deployed at the eavesdropped ground node m b of the cell b, b represents the channel resource occupied by the ground node related to the base station b, R in the MOS function is the data rate, is the minimum acceptable rate, and θ k is the optimal data rate.

[0015] Step 3: Construct an optimization model based on game theory, model the optimization problem in step 2 as a game problem with QoE evaluation strategy, specifically: each base station BS is regarded as a participant in the game process, and then the optimization problem is modeled as a strategy game, which is a precise potential game, for a set of game players, denotes the action space of player b, u b is the utility function of player b, in each element in the action space of each player, is subject to only one ground node occupies the channel s∈S, and the ground node is associated with the player, thus, denotes the joint action space that all players can choose, denotes the strategy distribution of all players except b, where × denotes the Cartesian product, and the expression of the utility function is as follows

[0016]

[0017] Suppose any one player unilaterally changes its action strategy from a b to a′ b According to the above formula, the following equation is established

[0018]

[0019] where is a defined potential function.

[0020] Step 4: A distributed learning algorithm is used to iteratively obtain the resource allocation strategy between the ground node and the base station until the optimization problem in step 2 is optimal, and the channel resource allocation method is determined.

[0021] Further improvement of the application is that the algorithm proposed in step 4 is a distributed algorithm, and in the inner loop of the algorithm, the proposed dynamic channel resource allocation strategy is based on a learning algorithm (SLA) of a random learning automaton and runs in an iterative manner, wherein each player is regarded as a learning automaton, and each player selects an available channel for data transmission in each iteration according to its current allocation strategy, and the changing communication environment is fed back to update the strategy. Through repeating the above process, each player continuously interacts with the environment to adjust its channel strategy to achieve the best QoE. In the outer loop of the algorithm, the constraint on the optimization variable is relaxed to [0, 1], the feasible point is updated by using the interior approximation method, and the SLA is stimulated to generate a series of feasible solutions so that the optimization target value monotonically increases.

[0022] Further improvement of the application is that the channel resource allocation strategy is an approximate optimal solution of the original problem, and specifically includes the following steps:

[0023] Step 4-1: Initialization: Set the initial value of iteration t=1, and the initial strategy is

[0024] Step 4-2: Loop iteration t = 1, 2,..., update

[0025] Step 4-3: Loop each iteration, t0 = 1, 2,...

[0026] Step 4-4: Initialization: Let The initial value of iteration t0 = 1 is set for the selection probability of GN about channel s associated with cell b, and the initial selection probability vector is

[0027] Step 4-5: At the beginning of t0 time slot, each participant b selects a ground node a using channel s according to the latest selection probability vector

[0028] Step 4-6: Each participant receives a signal from the ground terminal occupying channel s and obtains channel information, and after t0 time slot, each participant Receives the utility function

[0029] Step 4-7: All participants update the channel selection probability vector according to the following rules:

[0030]

[0031] Where η ∈ (0, 1) is the step size, if there is a Approximately 1, stop iteration and output the optimal strategy, Otherwise, return to step 1;

[0032] Step 4-8: End the loop;

[0033] Step 4-9: If the increase of the objective function is less than a given threshold, stop the algorithm and end the loop.

[0034] The beneficial effects of the present application are: the present application studies a resource allocation method of a 6G supported Internet of Things uplink communication network based on a limited block length communication theory, the problem is expressed as a non-convex problem, in order to provide URLLC service, the Shannon formula under the traditional infinite block length assumption is no longer applicable, the present application needs to use short data packet transmission to reduce delay; based on the safety rate formula of limited block length, the present application also designs an optimization target for evaluating communication reliability; the non-convex and composite expression of the optimization target leads to the problem being difficult to solve, the present application designs a game theory based model to analyze the problem, and proposes a distributed channel allocation algorithm to converge to the Nash equilibrium solution of the model; meet the communication service quality requirements of ultra-reliability and low latency. BRIEF DESCRIPTION OF DRAWINGS ​

[0035] Figure 1 is a system model diagram of the present application.

[0036] Figure 2 is a MOS function diagram of the present application.

[0037] Figure 3 is a channel allocation strategy optimization algorithm flow chart of the present application.

[0038] Figure 4 is a curve of the change of the channel s selected by the GN in a cell of the present application.

[0039] Figure 5 is the convergence behavior of the algorithm under different sizes of the set parameters of the present application.

[0040] Figure 6 is a comparison result of the MOS of the GN occupying the channel in each cell of the present application. DETAILED DESCRIPTION

[0041] The embodiments of the present application will be described below with reference to the drawings. Many practical details will be described in the following description for the purpose of clear illustration. However, it should be understood that these practical details are not used to limit the present application. That is, these practical details are not necessary in some embodiments of the present application.

[0042] The present application proposes a resource allocation method for a 6G-supported Internet of Things uplink communication network based on the limited block length communication theory, which specifically includes the following steps:

[0043] Step 1: First, an uplink system supporting ultra-reliable low-latency communication is constructed, which includes multiple base stations, ground nodes, and communication eavesdroppers. Each base station communication range includes multiple ground nodes and an eavesdropper.

[0044] As shown in Figure 1 , consider a 6G-supported Internet of Things uplink network, in which B base stations (BSs) are deployed, denoted as b∈B={1, 2, …, B}, N ground nodes (GNs), denoted as There is an eavesdropper in each cell range. It is assumed that each ground node, base station, and eavesdropper is equipped with only one antenna, and each GN only accesses one base station, denotes the set of all ground nodes within the service range of base station b, and then, and This network employs an orthogonal frequency division multiple access (OFDM) strategy with a reuse factor of 1. This means all cells utilize the entire spectrum resource, which is divided into S channels. These channels are orthogonal to each other and belong to the set S. Furthermore, the bandwidth of each channel is less than the coherence bandwidth, meaning each channel experiences flat fading. Within each cell, each channel is assigned to a GN associated with the base station, and this GN can occupy multiple channels. Therefore, interference does not exist within a single cell, but it does exist between cells.

[0045] To facilitate the analysis of the system model, this invention provides the following definition: Vector m = [m1, m2, ..., m B [ is defined as the number of channels s∈S occupied by ground nodes GNs in all cells, where] It is the GN associated with base station b and occupying channel s, notation s is omitted, let It is a constraint of m, let Indicates from GN The channel coefficients to BS c Indicates GN m b The signal power meets the requirements. Therefore, the received signal at BSb is

[0046]

[0047] in Additive white Gaussian noise

[0048] Based on the above analysis, from GN m b The signal-to-interference-plus-noise ratio (SINR) to BSb is

[0049]

[0050] in, Indicates from GN m c The channel gain to BS b.

[0051] Similarly, eavesdroppers deployed in cell b eavesdrop on signals from GN m b The signal-to-interference-plus-noise ratio (SINR) of the information is

[0052]

[0053] in, For noise power, Indicates from GN m b The channel gain to the eavesdropper.

[0054] To guarantee ultra-reliable low-latency communication, a short channel transmission mode is adopted, which leads to the decoding error rate not becoming 0 even if the SINR is very high, so the Shannon formula is no longer applicable. We need to use a new method to depict the relationship between the achievable rate, decoding error probability and transmission delay of secure Internet of Things communication under the condition of limited block length transmission. For a given channel block length L (L < 100), the channel coding rate is:

[0055]

[0056] where C is based on the Shannon capacity under the condition of infinite block length, V represents the channel dispersion, ∈ represents the expected decoding error probability, Q -1 is the inverse function of the Gaussian Q function,

[0057] Therefore, m b The approximate decoding rate under the given is expressed as

[0058]

[0059] where and Obviously, this approximation shows that compared with the channel capacity, the rate penalty is increased to maintain the maximum channel error probability ∈ on the limited block length L, which is proportional to .

[0060] Since it is unknown which GN in each cell b is eavesdropped by the eavesdropper, in order to avoid any information leakage of any GN, the secure communication rate of each GN in the cell should be considered. Therefore, the maximum secure communication rate of GN m b occupying the channel s can be approximated as:

[0061]

[0062] where and are the maximum decoding error probabilities, respectively.

[0063] Step 2: Construct an optimization problem to maximize the sum of the QoE based on the secure data rate of the communication between all ground nodes and base stations.

[0064] In order to evaluate the QoE of GN in uplink data transmission, the present application proposes a mean opinion score (MOS) standard

[0065]

[0066] where R in the MOS function is the data rate, is the minimum acceptable rate, theta k is the optimal data rate, R k represents the secure communication rate of GN k. The value of MOS varies from 1 to 5, when MOS = 1 represents the unacceptable QoE of GNs, and MOS = 5 is a very good QoE result.

[0067] In order to meet the communication requirements of secure URLLC in the Internet of Things, the application maximizes the optimization problem of MOS by considering channel resource allocation. Since the resource optimization problems on different channels are independent of each other, the single-channel problem proposed is:

[0068]

[0069] The above problem P is a non-convex discrete variable problem, and it is very difficult to find its solution.

[0070] Step 3: Construct an optimization model based on game theory, and model the problem in step 2 as a game problem with QoE evaluation strategy.

[0071] Specifically, it includes:

[0072] 3.1: Problem transformation

[0073] Since the optimization objective in P is an extremely complex composite function, it is difficult to handle. Therefore, first consider the relaxation of the function log2(1+x), as follows

[0074]

[0075] The equality is strictly established when x = x0, and

[0076] In order to simplify the optimization objective, the secure communication rate R sec Rewritten in the following form:

[0077]

[0078] Where

[0079] By formula substitution, The lower bound at the feasible point x t is:

[0080]

[0081] Where And Verification shows that psi(x, y) is a concave function.

[0082] Since MOS is a difficult-to-handle piecewise function, the present invention defines a continuous function to replace it

[0083]

[0084] where is a clear monotone non-increasing concave function, let The partial derivatives of f(x, y) with respect to x and y are as follows:

[0085]

[0086]

[0087] By using the first-order Taylor expansion of f(x, y) at the reachable point (x t , y t ), we can get

[0088]

[0089] It is obvious that for and the equality holds when x = x t , y = y t .

[0090] 3.2: Game-theoretic analysis

[0091] Each base station is considered as a player, and the optimization problem is modeled as a strategic game model, where is the set of players, denotes the action space of player b, and u b denotes the utility function of player b. In , each element in the action space of each player (BS) is subject to the condition that only one ground node occupies the channel s ∈ S, and the ground node is associated with the player. Therefore, denotes the joint action space that all players can choose, denotes the strategy profile of all players except b, where × denotes the Cartesian product, and the expression of the utility function is as follows

[0092]

[0093] Step 4: A distributed learning algorithm is adopted to iteratively obtain the resource allocation strategy between the ground nodes and the base stations until the optimization problem in Step 2 reaches the optimal, the channel resource allocation strategy is determined, and the algorithm flowchart is shown in Figure 3 .

[0094] (1) Initialization: Set the initial value of iteration t = 1, and the initial strategy is

[0095] (2) Loop iteration t = 1, 2,..., according to the optimization strategy, update

[0096] (3) Loop each iteration, t0= 1, 2,...

[0097] (4) Initialization: Let be the selection probability of GN associated with cell b with respect to channel s, set the initial value of iteration t0= 1, and the initial selection probability vector is

[0098] (5) At the beginning of the t0time slot, each participant b selects a ground node a using channel s according to the latest selection probability vector , that is, the selection strategy

[0099] (6) Then, each participant receives a signal from the ground terminal occupying channel s and obtains channel information, and after the end of the t0time slot, each participant receives the utility function

[0100] (7) All participants update the channel selection probability vector according to the following rules:

[0101]

[0102] where η ∈ (0, 1) is the step size, if there exists a approximating 1, stop iteration and output the optimal strategy, otherwise, return to step 1.

[0103] (8) End the loop.

[0104] (9) If the increase of the objective function value is less than a given threshold, stop the algorithm. End the loop.

[0105] Theorem 1: is an exact potential game

[0106] Proof: Let be the potential function of . Then the potential function can be written in the following form,

[0107]

[0108] where and Assume that any participant unilaterally changes its action policy from a b to a' b According to the above formula, the following equation holds

[0109]

[0110] Therefore, is a precise potential game.

[0111] Numerical simulations are performed by Matlab software to evaluate the communication performance of the proposed algorithm in 6G-based IoT. It is assumed that there are 3 cells, each of which is deployed with 1 BS in the center, and each cell is randomly distributed with 1 eavesdropper and 3 GNs. The system bandwidth is set to B = 5 MHz, which is divided into S = 4 channels. The noise power spectral density is -174 dBm / Hz, and the large-scale path loss model is 35.3 + 37.6 log 10 d b / d eav dB, where d b / d eav denotes the distance between GN / eavesdropper and base station. In the following simulation, the simulation results are obtained by 300 independent experiments, in which the involved parameters are optimized by experiments. Figure 4 The probability curve of the channel s selected by GN in one cell is shown as the algorithm iterates. It is noted that at about the 100th iteration, the selection probability vector of GN changes from to {1, 0, 0}. In Figure 5 , the convergence speed and performance of the algorithm are affected by the step size, and it can be seen from Figure 5 that when the step size η increases, the convergence speed of the algorithm is accelerated, but the cost is to get a non-NE solution with poor performance, and the convergence speed and performance of the algorithm are affected by the step size η. In addition, the position of the eavesdropper affects the QoE of all GNs. From Figure 5 it can also be seen that if the distance d eav between the eavesdropper and the BS decreases, the values of MOS and obtained by the optimal channel allocation strategy decrease. The reason for this is that when the eavesdropper moves towards the BS, the communication performance becomes worse to ensure secure communication. Figure 6 The MOS performance of each GN in the algorithm is shown, and in Figure 6 , in each cell, the histograms with the same mark represent the channels occupied by the same GN. It is worth noting that the algorithm can maintain better GN fairness in terms of MOS performance by optimizing the channel allocation strategy.

[0112] The present application studies the resource allocation strategy of 6G supported Internet of Things uplink communication network based on the limited block length communication theory, and the problem is expressed as a non-convex problem. To this end, we approximately transform the original problem into an easily handled problem, and propose an efficient distributed iterative algorithm to converge to the NE solution.

[0113] The above merely describes the embodiments of the present application and is not intended to limit the present application. Various modifications and changes can be made to the present application by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of the claims of the present application.

Claims

1.A resource allocation method for data security transmission in a high-reliability and low-latency Internet of Things, characterized in that: The resource allocation method comprises the following steps: Step 1: constructing an uplink system supporting ultra-reliable low-latency communication, the uplink system comprising a plurality of communication base stations, ground nodes and communication eavesdroppers, each of the communication base stations comprising a plurality of ground nodes and a service eavesdropper within a communication range; Step 2: constructing an optimization problem aiming to maximize the sum of QoEs of all the ground nodes and base stations based on a secure data rate of communication; Step 3: constructing an optimization model based on game theory, modeling the optimization problem in step 2 as a game problem with QoE evaluation strategies; Step 4: adopting a distributed learning algorithm to iteratively obtain resource allocation strategies between the ground nodes and the base stations until the optimization problem in step 2 reaches an optimum, and the channel resource allocation method is determined; wherein: the algorithm proposed in step 4 is a distributed algorithm, in the inner loop of the algorithm, the dynamic channel resource allocation strategy proposed is a learning algorithm (SLA) based on a random learning automaton and runs in an iterative manner, wherein each participant is regarded as a learning automaton, each participant selects an available channel for data transmission in each iteration according to the current allocation strategy, and a changing communication environment is fed back to update the strategy; through repeating the above process, each participant continuously interacts with the environment to adjust the channel strategy to achieve the best QoE; in the outer loop of the algorithm, the constraint on the optimization variable is relaxed to [0, 1], and an inner approximation method is used to update the feasible point, and the SLA is stimulated to generate a series of feasible solutions, so that the optimization target value monotonically increases; The channel resource allocation strategy is an approximate optimal solution of the original problem, and specifically comprises the following steps: Step 4-1: Initialization: Set the initial value of iteration t = 1, and the initial policy is where is a set of game participants; Step 4-2: Loop iteration t = 1, 2,..., update Step 4-3: loop each iteration, t0=1, 2, …; Step 4-4: Initialization: Let be the selection probability of the GN associated with the base station b with respect to the channel s, set the iteration initial value t0=1, and the initial selection probability vector is Step 4-5: At the beginning of the slot t0, each base station b selects a channel s according to the most recent selection probability vector selecting a channel s for use by a ground node a, i.e. a selection policy Step 4-6: Each participant receives the signal from the ground terminal occupying the channel s and acquires the channel information. After the end of the to time slot, each participant receiving utility function Step 4-7: all participants update the channel selection probability vector according to the following rules: where η ∈ (0, 1) is the step size, if there exists a Approximately 1, stop iteration and output the optimal strategy, otherwise return to step 1; Step 4-8: end the loop; Step 4-9: if the increase of the target function is less than a given threshold, stop the algorithm, end the loop, and the optimization function in step 2 is: Wherein: where R k represents the secure communication rate of GNk, C denotes the Shannon information capacity under infinite block length, V denotes the channel dispersion, ∈ denotes the decoding error probability, Q -1 denotes the inverse function of Gaussian Q function, denotes the signal-to-interference-and-noise ratio (SINR) between ground node m b and base station n, denotes the SINR of a eavesdropper deployed at base station b eavesdropping ground node m b , m b denotes the channel resource occupied by ground nodes associated with base station b, R in the MOS function is the data rate, is the minimum acceptable rate, θ k is the optimal data rate, L is the channel block length, and the step 3 is specifically: regarding each base station BS as a participant of a game process, and then the optimization problem is modeled as a strategy game, which is an exact potential game, is the set of game participants, denotes the action space of base station b, u b is the utility function of base station b, and in , each element in the action space of each participant is subject to the condition that only one ground node occupies the channel s∈S, and the ground node is associated with the participant, and therefore, denotes the joint action space that all participants can choose, denotes the strategy distribution of all participants except b, wherein × denotes the Cartesian product, and the expression of the utility function is as follows Assume that any base station unilaterally changes its action policy from a b to a' b According to the above formula, the following equation holds Γ(a b′ ,a -b )-Γ(a b ,a -b )=u b (a b ,a -b )-u b (a b ,a -b ) Wherein Γ is a defined potential function.

Images