Resource allocation method and system for UAV-assisted Internet of Things device network

By establishing a stochastic optimization model for maximum EE and using the Dinkelbach method and Gaussian theory, the energy efficiency and timeliness balance of UAV-assisted IoT devices under the imperfect CSI, and efficient resource allocation in 6G emergency communication scenarios are achieved , improves the energy efficiency and timeliness of the network.

CN117915482BActive Publication Date: 2024-11-15WUHAN INTEGRATED INTELLIGENT SYST INTEGRATION CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410075325.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-01-18
Publication Date
2024-11-15
Estimated Expiration
2044-01-18

AI Technical Summary

Technical Problem

In 6G wireless communication, UAV assisted IoT devices are difficult to maximize network energy efficiency (EE) and ensure different efficiency requirements when CSI is incomplete. Especially in emergency communication scenarios, the balance between timeliness and energy efficiency is difficult to achieve. .

Method used

Build a system model of UAV assisted IoT device network, establish a random optimization model with the goal of EE maximization, consider the conversion of AoI requirements into rate constraints, and combine the Dinkelbach method and Gaussian theory to transform the imperfect optimization model of CSI into deterministic convex Question, propose the AEE-ICSI resource allocation plan.

Benefits of technology

In the case of incomplete CSI, the energy efficiency and timeliness of the drone-assisted IoT network are effectively improved, ensuring the timeliness requirements of different services, and improving the interference efficiency of the system and the timeliness of data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117915482B_ABST
    Figure CN117915482B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of electrical communication technology, and specifically discloses a resource allocation method and system for a UAV-assisted Internet of Things device network. In order to solve the problem that the existing limited research only focuses on energy efficiency and information age (AoI) in an ideal environment, but does not consider the impact of imperfect channel state information (CSI) in an actual UAV-assisted Internet of Things device (IoT) network, the UAV-assisted Internet of Things device network is firstly systematically modeled, and then a random optimization model with the goal of maximizing EE is established based on the AoI model, and then the random optimization model is simplified, and finally the simplified random optimization model is solved to obtain the optimal solution of the optimization model. Simulation results show that this method has good interference efficiency (IE), AoI and EE (energy efficiency) in the UAV-assisted Internet of Things.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of electrical communication technology, and in particular to a resource allocation method and system for a UAV-assisted Internet of Things device network. Background Art

[0002] In the future sixth generation (6G) wireless communication, Internet of Things (IoT) devices are rapidly increasing and carrying a large number of various services. These devices mainly rely on ground base stations (BS) for information transmission. However, in emergency communication scenarios such as natural disasters, emergencies or network failures, when the target area lacks BS coverage, IoT devices often cannot work properly, and ultra-reliable low-latency communication (URLLC) of 6G wireless IoT services is difficult to guarantee. Therefore, in order to solve this problem, a study proposed a solution to use drone-assisted communication, that is, drones act as aerial base stations to collect IoT device data and transmit it to other IoT devices or directly to base stations to ensure that IoT devices can operate reliably under any circumstances. This solution has the advantage of high flexibility, but the limited battery power of drones remains a key challenge. At the same time, with the increase of IoT devices, frequent data interactions between different devices will lead to a rapid increase in data traffic, which will cause a sharp increase in drone energy consumption and seriously affect the performance of IoT networks. Therefore, how to maximize network energy efficiency (EE) has become a problem that must be solved for drone-assisted IoT networks.

[0003] In addition, due to the real-time changes in the network environment of drone-assisted IoT devices, the timeliness of the collected information is limited, especially in time-sensitive services such as disaster relief and environmental monitoring, the timeliness requirement becomes more and more critical. In order to quantify the timeliness of data, some studies have proposed the concept of Age of Information (AoI) and applied it to various wireless networks, which quantifies the time it takes from generation to receiving the destination. Good AoI performance can reduce the average time for data update and ensure the timeliness of information. Therefore, in the 6G drone-assisted IoT data transmission network, how to maximize the network EE while ensuring AoI has become an important issue that needs to be solved urgently.

[0004] In existing research, drone-assisted IoT networks have been studied in many aspects. Most of the research focuses on two aspects: (i) the relationship between energy and transmission efficiency; (ii) the impact of AoI on service quality.

[0005] In order to make the transmission efficiency of drone-assisted IoT networks higher, most scholars have considered the energy problem. For example, some literature has studied the problem of minimizing the total energy of drone communication with energy harvesting, which not only meets the user's minimum data transmission requirements, but also minimizes the total energy consumption of drones. Some literature has studied a full-duplex drone confidential communication scheme based on energy efficiency optimization. This scheme jointly optimizes the drone track and transmission power within a limited flight time to maximize the energy efficiency of confidential communication. At the same time, some literature has considered the EE problem under drone-assisted IoT networks. Some literature has studied the energy-saving design of drone data collection in large-scale machine communication networks, and proposed the problem of minimizing the total energy consumption of data collection, which effectively improved the overall energy efficiency. Some literature has proposed a dual UAV-assisted IoT based on non-orthogonal multiple access, which reduces the energy consumption of UAVs while ensuring a specific throughput. Some literature has proposed a joint UAV positioning and resource allocation algorithm for D2D communication in UAV-assisted industrial IoT networks, which maximizes the EE of the system under limited transmission rate and transmission power. Some literature has studied drone-assisted IoT, in which multiple rotor drones are used to provide services for multiple ground IoT devices, which can effectively improve the throughput and energy efficiency of the system. Although the above studies have made a lot of contributions to UAV-assisted networking, they rarely consider the timeliness of data, which may cause the UAV to be unable to receive fresh signals, affecting the timeliness performance.

[0006] In addition, some studies consider using AoI to reflect time-sensitive data. High-quality communication usually pursues higher timeliness. Therefore, some scholars are committed to minimizing AoI so that each IoTD can obtain better communication quality. Some literature studies the AoI minimization problem of unauthorized large-scale access in IoT networks to minimize the total AoI. Some literature considers cellular-based medical IoT systems with imperfect channel state information (CSI) and derives the average AoI closed-form expression that satisfies the transmission data interruption probability and EH interruption probability. Some literature studies the AoI minimization problem in IoT networks, including random identical and different multi-channel settings. Some literature studies the average AoI under traditional protocols and automatic retransmission request protocols to improve the freshness of target information in reality. In addition, since AoI can measure the timeliness of data well, some researchers have considered applying AoI to drone services. Some literature studies drone-assisted intelligent transportation systems, which enhance the use of vehicle networks, and AoI is used to measure the freshness of vehicle data packets. Some literature considers the task-aware trajectory design problem of cellular-connected drones, with the goal of minimizing the worst-case AoI. However, there are relatively few studies on the combination of UAVs and AoI, and the limited research mainly focuses on AoI guarantee under ideal environments without considering the impact of imperfect channel state information (CSI).

[0007] In the 6G emergency communication environment, complex terrain or dense buildings lead to randomness of imperfect CSI. Since channel estimation cannot accurately obtain the current channel state in real time, channel errors will occur. This channel uncertainty will affect the reliability of data transmission, causing uncertainty in transmission delay and timeliness. In addition, different services have different requirements for timeliness. Therefore, in the case of imperfect CSI, how to ensure the timeliness requirements of different services is a challenge. Summary of the invention

[0008] The present invention provides a resource allocation method and system for a UAV-assisted Internet of Things device network, and solves the technical problem of how to maximize network EE while ensuring different timeliness requirements when CSI is imperfect.

[0009] In order to solve the above technical problems, the present invention provides a resource allocation method for a UAV-assisted Internet of Things device network, comprising the steps of:

[0010] S1. Construct a system model of a UAV-assisted IoT device network; the system model includes C unmanned aerial vehicles (UAVs), D IoT devices (IoTDs), and J subcarriers; the D IoT devices include M AoI-tolerant devices (ATDs) and N AOI-sensitive devices (ASDs), where AoI refers to the information age; the UAV acts as an air base station to provide services for the IoT devices in a 6G emergency scenario, and the IoT devices share the spectrum resources of the UAV;

[0011] S2. Based on the system model, a stochastic optimization model with EE maximization as the goal is established:

[0012] max R

[0013]

[0014]

[0015]

[0016]

[0017] Among them, R represents the total system rate, max represents the maximum, and st represents the conditions that need to be met. It means arbitrary; It represents the allocation factor of the nth ASN, i.e. ASD n, on the jth subcarrier, i.e. subcarrier j. It represents the allocation factor of the mth ATD, i.e. ATD m, on the jth subcarrier. When , it indicates that subcarrier j is allocated to ASD n or ATD m; represents the transmit power of ASD n to UAV on subcarrier j, represents the transmit power of ATD m to UAV on subcarrier j; represents the channel gain between ASD n and the cth UAV, i.e., UAV c, on subcarrier j, represents the channel gain between ATD m and UAVc on subcarrier j; represents the achievable transmission rate from ASD n to UAV on the jth subcarrier, represents the achievable transmission rate from ATD m to UAV on the jth subcarrier; is the minimum rate threshold of ASD n, The minimum rate threshold of I th represents the predetermined interference threshold; ξ represents the rate probability threshold of ATD, ψ represents the rate probability threshold of ASD; Pr{·} represents the probability of event violation in {·}, constraint C1 represents the constraint condition for spectrum sharing between IoT devices and UAVs; constraint C2 represents the minimum rate threshold constraint for all ATDs under AoI-tolerant service; constraint C3 represents that the spectrum of a UAV can be shared with multiple IoTDs; constraint C4 represents the rate constraint of all ASDs under AoI-sensitive service requirements;

[0018] S3, simplifying the random optimization model to obtain a simplified optimization model;

[0019] S4. Solving the simplified optimization model.

[0020] Furthermore, the total system rate R is calculated by the following formula:

[0021]

[0022] Calculated by the following formula:

[0023]

[0024] Where W represents the bandwidth of each channel, They represent the SINR, or signal to interference plus noise ratio, of ASD n and ATD m respectively.

[0025] Further, Calculated by the following formula:

[0026]

[0027] in, is the channel gain from ASD n to UAV on subcarrier j, is the channel gain from ATD m to UAV on subcarrier j, σ2 represents the sum of noise power and UAV interference;

[0028] Calculated by the following formula:

[0029]

[0030] Further, in step S3, the simplified optimization model is expressed as:

[0031] max R-μE,

[0032]

[0033]

[0034]

[0035]

[0036] Where μ represents the auxiliary variable introduced by the Dinkelbach method, represents the total power of the system, p cir Indicates the circuit power of all mobile devices in the system; They are defined as:

[0037]

[0038]

[0039] in, τ t represents the channel estimation error of ATD, δ s represents the channel estimation error of ASD, Q -1 (ψ) represents the inverse Gaussian function; It is estimated and

[0040] Furthermore, in the simplified optimization model, the objective function R-μE is expanded as follows:

[0041]

[0042] Among them, the intermediate parameter c t ,q s 、c s ,q t They are defined as:

[0043]

[0044]

[0045]

[0046]

[0047] Furthermore, the step S4 specifically includes the steps of:

[0048] S41. According to the duality theory, the Lagrangian dual function of the simplified optimization model is obtained as follows:

[0049]

[0050] Among them, α, l m , χ and are non-negative multipliers of various terms;

[0051] S42, based on Γ(x), the simplified optimization model is converted into a Lagrangian dual model:

[0052]

[0053]

[0054] S43. According to Lagrange's KKT condition, the transmission power is obtained and Solution for:

[0055]

[0056] Among them, the auxiliary parameters and

[0057] Furthermore, the step S43 specifically includes the steps of:

[0058] S431, Initialize system parameters: ξ,ψ, N, M, α, l m , χ, Convergence accuracy Θ, step size λ k ;

[0059] S432, set the transmission power and The starting point and Set the number of iterations k = 0, the gradient function

[0060] S433, calculate function value and According to the gradient Calculate the gradient value gk ;

[0061] S434, when || g k ||>Θ, go to step S435, otherwise end, || || represents the norm;

[0062] S435, Settings And according to Calculating λ k , λ is the data arrival rate;

[0063] S436, update iteration point And calculate the function Γ(p k+1 ) value;

[0064] S437, judge if ||Γ(p k+1 )-Γ(p k )||>Θ, then go to step S438, otherwise output the solution of the optimization model

[0065] S438. Set k=k+1 and jump to step S434.

[0066] The present invention also provides a resource allocation system for a UAV-assisted Internet of Things device network, the key of which is that it includes a system model construction module, an optimization model construction module, an optimization model simplification module and an optimization model solving module; the system model construction module, the optimization model construction module, the optimization model simplification module and the optimization model solving module are respectively used to execute steps S1, S2, S3 and S4 described in the above method.

[0067] The resource allocation method and system of the UAV-assisted Internet of Things device network provided by the present invention first systematically models the UAV-assisted Internet of Things device network, and then establishes a random optimization model with the goal of maximizing EE based on the AoI model. Due to the uncertainty of the UAV channel, it is difficult for the random optimization model to obtain the optimal solution. Therefore, the random optimization model is simplified. Finally, the simplified random optimization model is solved to obtain the optimal solution of the optimization model. The present invention proposes a resource allocation algorithm for UAV-assisted Internet of Things devices, which takes into account AoI and energy efficiency (AEE-ICSI) under imperfect CSI conditions. Simulation results show that the algorithm has good interference efficiency (IE), AoI and EE in the UAV-assisted Internet of Things, proving the effectiveness of the scheme. The main contributions of the present invention are summarized as follows:

[0068] 1) In order to ensure the timeliness requirements of different services, the AoI requirements are converted into rate constraints so that the resource allocation model can ensure the AoI requirements by constraining the rate threshold.

[0069] 2) In order to improve the network efficiency of drone-assisted IoT, an optimization problem based on the AoI model is proposed to improve the network efficiency. Combining the transmission rate requirements for different AoI services and the interference constraints between drones and IoT devices, a stochastic optimization model based on the AoI model is established. This model solves the problem of maximizing the network EE while ensuring the needs of different AoI services.

[0070] 3) In view of the complex model environment, the channel estimation cannot accurately obtain the current channel state, which makes the proposed optimization model have the problem of imperfect CSI, resulting in the inability of the optimization model to obtain the optimal solution. This paper uses the Dinkelbach method, quadratic transformation method and Gaussian theory to transform the imperfect CSI optimization model into a deterministic convex problem. Finally, an AEE-ICSI resource allocation scheme based on drone-assisted IoT communication scenarios is proposed to solve the optimization model and effectively improve the EE of the network. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 is a flow chart of a resource allocation method for a UAV-assisted Internet of Things device network provided by an embodiment of the present invention;

[0072] Figure 2 is an architecture diagram of a UAV-assisted Internet of Things device network provided by an embodiment of the present invention;

[0073] Figure 3 is an example diagram of AoI provided by an embodiment of the present invention;

[0074] Figure 4 is a relationship diagram between IE of IoT and channel estimation errors δ and τ under different channel conditions provided by an embodiment of the present invention;

[0075] Figure 5 is the maximum AoI threshold value of the AoI sensitive service provided by the embodiment of the present invention Relationship diagram with IoT throughput;

[0076] Figure 6 is a diagram showing the relationship between the bandwidth W and the total EE of IoT devices provided in an embodiment of the present invention;

[0077] Figure 7 is a diagram showing the relationship between the bandwidth W and the total IE of IoT devices provided in an embodiment of the present invention;

[0078] Figure 8 is a relationship diagram between bandwidth W and IoT device energy provided by an embodiment of the present invention;

[0079] Fig. 9 4 is a diagram showing the relationship between the bandwidth W and the throughput of IoT provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0080] The following specifically illustrates the implementation mode of the present invention in conjunction with the accompanying drawings. The embodiments are provided for illustrative purposes only and cannot be understood as limiting the present invention. The accompanying drawings are provided for reference and illustration only and do not constitute a limitation on the scope of patent protection of the present invention, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0081] The embodiment of the present invention provides a resource allocation method for a UAV-assisted Internet of Things device network, such as Figure 1 As shown, the steps include:

[0082] S1. Construct a system model of UAV-assisted IoT device network;

[0083] S2. Based on the system model, a stochastic optimization model with EE maximization as the goal is established:

[0084] S3, simplifying the random optimization model to obtain a simplified optimization model;

[0085] S4. Solve the simplified optimization model.

[0086] (1) Step S1: Building a system model

[0087] Figure 2 A drone-assisted IoT network architecture is considered in this paper, in which drones act as aerial base stations to provide services for IoT devices in 6G emergency scenarios, and IoT devices share the spectrum resources of drones. The network has C UAVs (including UAV 1 to UAV C) and D IoTDs, and the D IoTDs include M devices carrying services with AoI tolerance requirements (including ATD 1 to ATD M) and N devices carrying services with AoI sensitive requirements (including ASD 1 to ASD N), where IoT devices carrying services with low timeliness requirements are called AoI tolerant devices (ATDs), and IoTDs carrying services with high timeliness requirements are called AoI sensitive devices (ASDs). In addition, there are J subcarriers in the network.

[0088] (2) Step S2: Constructing a stochastic optimization model

[0089] The present invention uses queue theory to construct AoI models for different services. On this basis, a random optimization model with energy efficiency maximization as the goal is established.

[0090] A. Average AOI of M / M / 1 model

[0091] In drone-assisted IoT, information changes over time, and the validity of data is closely related to timeliness. Therefore, we introduce the concept of AoI, which describes the timeliness of data received by the BS to measure the performance of the system. In the 6G emergency scenario, the main task of the drone-assisted IoT network is that the IoT device sends the collected data packets to the drone that receives the data packets and sends the data packets to the base station or other devices. In the drone-assisted IoT network, we assume that the source is the IoT device and the destination is the drone. t k is the time when the IoT device sends the IoT data packet, t k ' is the time at which the packet is received at the destination. At any time t, the index of the latest packet received by the drone is:

[0092] k=max{i|t k '≤t} (1)

[0093] Max means to maximize.

[0094] The latest update timestamps received by the drone are as follows:

[0095] U(t)=t k (2)

[0096] The AoI at time t is defined as:

[0097] Δ(t)=tU(t) (3)

[0098] Figure 3 A schematic diagram showing the change of AoI over time in a UAV-assisted IoT network. The time curve in the graph represents the state age time A t , which increases linearly and then drops to a smaller value when a packet is received.

[0099] In drone-assisted IoT, we assume that the process of IoT devices sending data packets is an M / M / 1 queuing model. In the M / M / 1 model, the data arrival rate of IoT devices follows a Poisson distribution, and the data service time follows an exponential distribution. λ is the data arrival rate, and is the achievable transmission rate from ASD n (nth ASD) to UAV on subcarrier j (jth subcarrier). Y is the exponential distribution of data inter-arrival time, and IoTD utilization is defined as therefore, and T i is the system time, that is, the queue waiting time W i and service time S i The sum is as follows:

[0100] Ti =W i +S i (4)

[0101] If the queue is empty, then W i = 0. At the same time, when the UAV-assisted IoT network reaches a stable state, the system time T i is random and consistent, the system time T of the M / M / 1 model i The probability density function of is:

[0102]

[0103] According to formula (4), it is expected that Calculated by the following formula:

[0104]

[0105] in expect It can be obtained by the following formula:

[0106]

[0107] The average AoI can be obtained as:

[0108]

[0109] Where Q represents Figure 3 We substitute equations (6) and (7) into (8), and the average AoI of M / M / 1 can be obtained as follows:

[0110]

[0111] Through the derived average AoI, we can observe that the average AoI in the M / M / 1 model is related to the processing rate and data arrival rate of IoTD.

[0112] B. AoI Model for AoI-Sensitive Services

[0113] In the drone-assisted IoT device network, there are two different services for IoT devices: AoI-sensitive and AoI-tolerant. AoI-sensitive services need to ensure the timeliness of data transmission. However, AoI-tolerant services have no requirements for timeliness. Therefore, the services for which we build the AoI model have different AoI requirements.

[0114] To meet the timeliness requirements of AoI-sensitive services, the average AoI must be less than the maximum allowed in the queue. The threshold value is expressed as follows:

[0115]

[0116] According to Cardano’s approach, the rate requested by the user for ASD is obtained in the following way:

[0117]

[0118] in is the solution of equation (10), which represents the minimum rate threshold of ASD n. Equation (11) transforms the AoI requirement of ASD n into a rate requirement and obtains the minimum rate constraint required to ensure the AoI requirement of ASD n. According to the rate formula (11), the signal to interference plus noise ratio (SINR) under ASD is expressed as:

[0119]

[0120] in, is the transmit power of ASD n to UAV on subcarrier j, is the transmit power of ATD m to UAV on subcarrier j, is the channel gain from ASD n to UAV on subcarrier j, is the channel gain from ATD m to UAV on subcarrier j. At the same time, the mobility and complex behavior of UAVs make the communication environment have imperfect CSI, so and is random. 2 represents the sum of noise power and UAV interference.

[0121] C. AoI Model for AoI Tolerant Services

[0122] Since AoI tolerant services have no timeliness requirements, we only need to ensure the minimum rate required by all ATDs, namely:

[0123]

[0124] Among them, Pr{·} represents the probability of violation of the event in {·}, Represents arbitrary. Similarly, the SINR of ATD is expressed as:

[0125]

[0126] According to equations (12) and (14), the corresponding user rates of ATD and ASD are obtained in the following way:

[0127]

[0128] Where W represents the bandwidth of each channel. Therefore, according to formula (15), we can obtain the total system rate as:

[0129]

[0130] Therefore, the total power of the system is:

[0131]

[0132] where p cir Represents the circuit power of all mobile devices in the system.

[0133] In the UAV-assisted IoT network, the premise of using the spectrum sharing mode is that the interference caused by the IoT device to the UAV must be lower than the predetermined interference threshold, otherwise it will affect the normal communication of the UAV. Therefore, the interference must be constrained, as shown in formula (18):

[0134]

[0135] in Indicates the subcarrier allocation factor for ASD n, Indicates the subcarrier allocation factor of ATD m. , it indicates whether subcarrier j is allocated to ASD n or ATD m. The allocation of subcarriers to ASDs is determined by the indicator variable and express. is the channel gain between ASD n and UAV c (the cth UAV) on subcarrier j, I represents the channel gain between ATD m and UAV c on subcarrier j. th is the predetermined interference threshold.

[0136] D. Problem Formulation

[0137] In the UAV-assisted Internet of Things, in order to solve the problem of maximizing the EE in the network while ensuring the AoI requirements of different services, a resource allocation model considering imperfect CSI is established. This model comprehensively considers the UAV transmission power constraint, inter-device interference constraint and channel selection constraint to maximize energy efficiency. The mathematical model is shown in formula (19):

[0138]

[0139] Among them, constraint C1 represents the constraint condition for spectrum sharing between IoT devices and UAVs. Constraint C2 represents the minimum rate threshold constraint for all ATDs under AoI tolerance service. Constraint C3 indicates that the spectrum of a UAV can be shared with multiple IoTs. Constraint C4 represents the rate constraint of all ASDs based on the transformation of formula (10) under AoI-sensitive service requirements.

[0140] (3) Step S3: Simplifying the stochastic optimization model

[0141] From the random optimization model, we can see that the random parameter h of the imperfect CSI is introduced. n s,j and ht m,j , making the model a mixed integer nonlinear programming model. Due to the introduction of random parameters, traditional optimization methods cannot solve the global optimal solution of this model. Therefore, this paper uses the idea of ​​Gaussian theory to transform the random probability constraints into deterministic convex constraints and applies them to the solution of the optimization model. Moreover, another challenge in solving this optimization model is that the objective function is a fractional problem, which is not easy to solve.

[0142] Since the optimization model is a fractional objective function, it is difficult to solve. Therefore, the Dinkelbach method is used to introduce the auxiliary variable μ and transform the objective function into the form of parameter subtraction:

[0143] F(μ)=R-μE (20)

[0144] Expand the objective function into (21):

[0145]

[0146] At the same time, due to the complexity of the rate constraints, it is difficult to obtain an analytical solution to EE. Therefore, we simplify the rate equation according to the quadratic transformation method. The rate equation under ATD is transformed into:

[0147]

[0148] Among them, the intermediate parameters are simplified. and Therefore, using a similar approach, the rate formula under ASD can be written as:

[0149]

[0150] in and We can obtain the transformed rate constraints with ATD and ASD through equations (22) and (23).

[0151] because and The uncertainty of leads to errors in the channel, so the rate expression constraint and The channel uncertainty can be modeled as:

[0152]

[0153] in, and is the estimated channel gain and therefore, and Represents the corresponding channel estimation error, and CN() indicates that it obeys Gaussian probability distribution.

[0154] Due to the influence of random parameters, the rate constraint of ATD becomes an uncertain constraint that is difficult to solve. Therefore, in order to reduce the influence of random parameters, we transform the rate constraint into equation (25):

[0155]

[0156] in, and According to the established model (24), (25) can be rewritten as:

[0157]

[0158] In the formula,

[0159] According to Gauss's theorem, we can get:

[0160]

[0161] Among them, Q -1 (ξ) represents the inverse Gaussian function with ξ as the independent variable.

[0162] Since the formula is relatively complex, it is difficult to solve the optimization model. Therefore, according to the inequality and Formula (27) can be rewritten as:

[0163]

[0164] in and Therefore, the rate constraint (13) becomes:

[0165]

[0166] in On this basis, the random parameters in the ATD rate constraint are converted into deterministic constraints.

[0167] In the rate constraint with ASD, random parameters also exist, which makes it difficult to solve the optimization model. We convert the rate constraint into:

[0168]

[0169] Based on the same method, formula (30) is rewritten as:

[0170]

[0171]

[0172] Among them, at m, At n, Based on model (24), formula (31) becomes:

[0173]

[0174] In the formula and Based on the same method, formula (32) can be written as follows:

[0175]

[0176] Because (33) is very complicated, it is not easy to get an analytical expression. After simplification, we can get:

[0177]

[0178] in and Q -1 (ψ) represents the inverse Gaussian function with ψ as the independent variable. The rate constraint in the optimization model (19) with ASD becomes:

[0179]

[0180] In the formula, Based on this transformation, the rate constraint under ASD is transformed into a deterministic constraint. Based on this series of transformations, model (19) is transformed into a deterministic constraint model. As shown in formula (36):

[0181]

[0182] The objective function is expressed as (37):

[0183]

[0184] (4) Step S4: Solving the model

[0185] Since the objective function in the optimization model (37) contains the influence of random parameters, the optimization problem does not belong to the category of convex optimization problems. For non-convex optimization problems, since their dual problems belong to the category of convex optimization problems, we convert them into dual problems. Based on this, step S4 specifically includes the steps:

[0186] S41. According to the duality theory, the Lagrangian dual function Γ(x) of the simplified optimization model is obtained;

[0187] S42, based on Γ(x), the simplified optimization model is transformed into a Lagrangian dual model;

[0188] S43. According to Lagrange's KKT condition, the transmission power is obtained and Solution

[0189] According to the duality theory, the Lagrange dual function of the optimization problem is as follows:

[0190]

[0191] Among them, α, l m , χ and is a non-negative multiplier of each term. After simplification, we can get:

[0192]

[0193] in:

[0194]

[0195] Therefore, the optimization problem (36) is replaced by the Lagrangian dual problem. Then, the Lagrangian dual problem can be formulated as:

[0196]

[0197] Therefore, according to Lagrangian KKT conditions, the transmission power can be obtained and for:

[0198]

[0199] Formula (42) can be transformed into:

[0200]

[0201] Therefore, it is equal to:

[0202]

[0203] in and Therefore, formula (44) can be restated as:

[0204]

[0205] Based on equation (45), the power function solution of the optimization model is obtained. According to equations (16) and (17), the EE of IoTDS service can be obtained.

[0206] Step S43 specifically includes the following steps:

[0207] S431, Initialize system parameters: ξ,ψ, N, M, α, l m , χ, Convergence accuracy Θ, step size λ k ;

[0208] S432, set the transmission power and The starting point and Set the number of iterations k = 0, the gradient function

[0209] S433, calculate function value and According to the gradient Calculate the gradient value g k ;

[0210] S434, when || g k ||>Θ, go to step S435, otherwise end, |||| represents the norm;

[0211] S435, Settings And according to Calculating λ k , λ is the data arrival rate;

[0212] S436, update iteration point And calculate the function Γ(p k+1 ) value;

[0213] S437, judge if ||Γ(p k+1 )-Γ(p k )||>Θ, then go to step S438, otherwise output the solution of the optimization model

[0214] S438. Set k=k+1 and jump to step S434.

[0215] We use an iterative algorithm to find and To obtain the optimal solution of the optimization model. First, initialize the system parameters and set the initial point and Convergence accuracy Θ and step size λ k Then calculate the initial value and gradient of the function, and judge whether the iteration stops by comparing the gradient norm and the convergence precision Θ. If the norm of the gradient is less than the convergence precision Θ, the iterative algorithm stops and obtains and If the norm of the gradient is greater than the convergence accuracy Θ, the step size is calculated and the iteration point is updated. Then, we determine whether it can continue iteratively. Finally, we find and The solution is substituted into formula (15) to obtain the global optimal solution of the optimization model.

[0216] The pseudo code of AEE-ICSI resource allocation algorithm is shown in Table 1.

[0217] Table 1: Algorithm pseudo code

[0218]

[0219]

[0220] According to the pseudo code in Table 1, the UAV transmission power can be obtained in each iteration process. and The solution is substituted into equation (38) to obtain the global optimal solution of the optimization model, which requires O(MNJ) operations. In addition, the number of outer loops of the proposed AEEICSI resource allocation algorithm is k. Only some simple addition and subtraction operations are involved in the outer loop process, and the constant complexity is ignored. Therefore, the total complexity of the proposed AEE-ICSI resource allocation algorithm is O(MNJ). In addition, if the appropriate step size and the initial value of the Lagrange multiplier are selected, the parameter k can be very small.

[0221] Based on the above method, this embodiment also provides a resource allocation system for a UAV-assisted Internet of Things device network, including a system model construction module, an optimization model construction module, an optimization model simplification module and an optimization model solving module; the system model construction module, the optimization model construction module, the optimization model simplification module and the optimization model solving module are respectively used to execute steps S1, S2, S3, and S4 in the above method.

[0222] The simulation analysis is carried out below.

[0223] The present invention provides numerical and simulation results in 6G emergency situations to show the performance of the proposed algorithm and verify the efficiency of the optimization we proposed. The simulation scenario is a drone-assisted IoT network in a 6G emergency scenario. The number of drones is 30, the number of IoT devices is 20, and the base station bandwidth is 5MHz. The rest of the simulation parameter settings are shown in Table 2.

[0224] Table 2: Simulation parameter settings

[0225]

[0226]

[0227] Figure 4The relationship between the IE of IoTD and the channel estimation errors δ and τ under different channel conditions is shown. Figure 4 The channel condition of IoTD 1 is Figure 4 The channel conditions of IoTD 2 are like Figure 4 As shown in Figure 1, in the 6G emergency scenario, different channel conditions will lead to different IEs of IoT devices, that is, the smaller the channel estimation errors δ and τ are, the smaller the IE of IoT devices is. As shown in equation (16), when the channel estimation errors δ and τ are large, the channel gain deviates more from the accurate value. In addition, we can see that the IE of IoTD 2 with good channel conditions is higher than that of IoTD 1 with poor channel conditions.

[0228] Figure 5 The maximum AoI threshold for AoI sensitive services is shown The relationship between the throughput of IoT and Figure 5 It can be seen that as the maximum AoI threshold increases, the throughput of IoT decreases. This is because in AoI-sensitive services, when the maximum AoI threshold is When increases, the timeliness of IoT data will decrease, which reduces the communication quality between devices,and causes a decrease in network throughput.,In addition, we can also see that the data arrival rate λ also affects the throughput.,When the data arrival rate increases, the throughput increases.

[0229] Figure 6 The relationship between bandwidth W and total EE of IoT is shown. Figure 6 It can be seen that in the 6G emergency scenario, the proposed AEE-ICSI resource allocation algorithm has a higher EE than the comparison algorithm. We can also see that when the bandwidth W increases, the corresponding EE also increases rapidly. At the same time, we can also see the efficiency of the proposed AEEICSI resource allocation algorithm under different channel estimation errors. When the channel estimation error τ is large, the total EE of IoTD is small.

[0230] Figure 7 The relationship between bandwidth W and total IE of IoT is shown. Figure 7 It can be seen that the proposed AEE-ICSI resource allocation algorithm has a lower IE than the comparison algorithm. When the bandwidth W increases, the corresponding IE also increases, but the growth curve is relatively flat. We can also see the IE of the proposed AEE-ICSI resource allocation algorithm under different channel estimation errors. When the channel estimation error τ is more significant, the total IE of the device is larger.

[0231] Will Figure 6 and Figure 7By comparison, it can be seen that the total EE in the UAV-assisted IoT network is higher than the total IE, which indicates that the proposed AEE-ICSI resource allocation algorithm has better performance than the comparison algorithm. In the 6G emergency scenario, it can effectively guarantee the EE under the condition of imperfect CSI and the timeliness requirements of different AoI services.

[0232] Figure 8 The relationship between bandwidth W and IoT energy is shown. Figure 8 It can be seen that the proposed AEE-ICSI resource allocation algorithm has higher energy than the comparison algorithm. When the bandwidth W increases, the corresponding energy also increases rapidly. We can also see the energy of the proposed AEE-ICSI resource allocation algorithm for different channel estimation errors. When the channel estimation error τ is large, its IoTD EE is lower.

[0233] exist Figure 6 , Figure 7 and Figure 8 We can observe that in the 6G emergency scenario, when the estimation errors δ and τ are smaller, the channel gain is closer to the exact value, the total rate of the UAV-assisted IoT network increases, and the corresponding EE and IE become smaller.

[0234] Fig. 9 The relationship between bandwidth W and IoT throughput is shown. Fig. 9 It can be seen that the throughput (bits / J) of the proposed AEE-ICSI resource allocation algorithm IoTD has a higher throughput than the comparison algorithm. We can also see the throughput of the proposed AEE-ICSI resource allocation algorithm under different channel estimation errors. When the channel estimation error is small, the throughput of IoTD is large. Similarly, when the channel estimation error τ is small, the total system rate increases, that is, the throughput of IoT increases.

[0235] From the above simulation results, it can be seen that in the 6G emergency scenario, when the channel conditions change, it can be observed that the IE and EE of the proposed AEE-ICSI resource allocation algorithm will increase. Under the same environment, as the channel estimation error increases, the throughput will decrease. At the same time, it can be seen that in AoI-sensitive services, the increase in AoI will also reduce the network throughput. This shows that the AEE-ICSI resource allocation algorithm proposed in the drone-assisted IoT network can effectively reduce interference and improve the EE of the network.

[0236] In summary, the embodiment of the present invention provides a resource allocation method for a UAV-assisted IoT device network, and studies the problem of maximizing network EE while ensuring different AoI requirements in a UAV-assisted IoT network. First, considering the different requirements of different services for data timeliness in the 6G emergency communication scenario, an AoI model for different services is established based on the average AoI under the M / M / 1 system, and different AoI requirements are converted into rate expressions. On this basis, an AEE-ICSI resource allocation algorithm is proposed, and an optimization model that maximizes resource efficiency while ensuring data timeliness is established. Finally, in the optimization model, we use the Dinkelbach method to introduce auxiliary variables to simplify the optimization objective function. At the same time, a simplified method based on Gaussian theory is proposed for the channel uncertainty problem in the UAV-assisted IoT. In addition, we also analyze the complexity of the proposed AEE-ICSI resource allocation algorithm. The simulation results show that the proposed AEE-ICSI resource allocation algorithm has good IE and EE performance, proving the effectiveness of the scheme.

[0237] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A method for allocating resources in a UAV-assisted IoT device network, characterized in that: Includes steps: S1. Construct a system model of a UAV-assisted IoT device network; the system model includes C unmanned aerial vehicles (UAVs), D IoT devices (IoTDs), and J subcarriers; the D IoT devices include M AoI-tolerant devices (ATDs) and N AOI-sensitive devices (ASDs), where AoI refers to the information age; the UAV acts as an aerial base station to provide services for the IoT devices in a 6G emergency scenario, and the IoT devices share the spectrum resources of the UAV; S2. Based on the system model, a stochastic optimization model with EE maximization as the goal is established: maxR Among them, R represents the total system rate, max represents the maximum, and st represents the conditions that need to be met. It means arbitrary; It represents the allocation factor of the nth ASN, i.e. ASD n, on the jth subcarrier, i.e. subcarrier j. It represents the allocation factor of the mth ATD, i.e. ATD m, on the jth subcarrier. When , it indicates that subcarrier j is allocated to ASD n or ATD m; represents the transmission power of ASDn to UAV on subcarrier j, represents the transmit power of ATD m to UAV on subcarrier j; represents the channel gain between ASD n and the cth UAV, i.e., UAV c, on subcarrier j, represents the channel gain between ATD m and UAV c on subcarrier j; represents the achievable transmission rate from ASD n to UAV on the jth subcarrier, represents the achievable transmission rate from ATD m to UAV on the jth subcarrier; is the minimum rate threshold of ASD n, Indicates the minimum rate threshold of ATD m; I th represents the predetermined interference threshold; ξ represents the rate probability threshold of ATD, ψ represents the rate probability threshold of ASD; Pr{} represents the probability of event violation in {}, constraint C1 represents the constraint condition for spectrum sharing between IoT devices and UAVs; constraint C2 represents the minimum rate threshold constraint for all ATDs under AoI tolerant service; constraint C3 represents that the spectrum of a UAV can be shared with multiple IoTDs; constraint C4 represents the rate constraint of all ASDs under AoI sensitive service requirements; S3, simplifying the random optimization model to obtain a simplified optimization model; in step S3, the simplified optimization model is expressed as: maxR-μE, Where μ represents the auxiliary variable introduced by the Dinkelbach method, represents the total power of the system, p cir Indicates the circuit power of all mobile devices in the system; They are defined as: Where W represents the bandwidth of each channel, δ s represents the channel estimation error of ASD, δ t represents the channel estimation error of ATD, Q -1 () represents the inverse Gaussian function; It is estimated and is the channel gain from ASD n to UAV on subcarrier j, is the channel gain from ATD m to UAV on subcarrier j, σ 2 represents the sum of noise power and UAV interference; In the simplified optimization model, the objective function R-μE is expanded as: Among them, the intermediate parameter c t ,q s 、c s ,q t They are defined as: S4, solving the simplified optimization model; The step S4 specifically comprises the following steps: S41. According to the duality theory, the Lagrangian dual function of the simplified optimization model is obtained as follows: Among them, α, m , χ and are non-negative multipliers of various terms; S42, based on Γ(x), the simplified optimization model is converted into a Lagrangian dual model: S43. According to Lagrange's KKT condition, the transmission power is obtained and Solution for: Among them, the auxiliary parameters and 2. The resource allocation method for a UAV-assisted IoT device network according to claim 1, characterized in that: The total system rate R is calculated by the following formula: Calculated by the following formula: in, They represent the SINR, or signal to interference plus noise ratio, of ASD n and ATD m respectively.

3. The resource allocation method for a UAV-assisted IoT device network according to claim 2, characterized in that: Calculated by the following formula: Calculated by the following formula:

4. The resource allocation method for a UAV-assisted IoT device network according to claim 1, characterized in that: The step S43 specifically includes the following steps: S431, Initialize system parameters: ξ,ψ, N, M, α, m , χ, Convergence accuracy Θ, step size λ k ; S432, set the transmission power and The starting point and Set the number of iterations k = 0, the gradient function S433, calculate function value and According to the gradient Calculate the gradient value g k ; S434, when g k >Θ goes to step S435, otherwise ends, indicating the norm; S435, Settings And according to Calculating λ k , λ is the data arrival rate; S436, update iteration point And calculate the function Γ(p k+1 ) value; S437, judge if Γ(p k+1 )-Γ(p k )>Θ, then go to step S438, otherwise output the solution of the optimization model S438. Set k=k+1 and jump to step S434.

5. A resource allocation system for a UAV-assisted IoT device network, characterized by: It includes a system model construction module, an optimization model construction module, an optimization model simplification module and an optimization model solving module; the system model construction module, the optimization model construction module, the optimization model simplification module and the optimization model solving module are respectively used to execute steps S1, S2, S3, S4 described in any one of claims 1 to 4.