QoS-driven user scheduling and power allocation method for energy harvesting wireless IoT

Through the theory of effective bandwidth and effective capacity, the random optimization problem in wireless Internet of Things communication is transformed into deterministic optimization problem. Combined with the Lagrangian method and branch delimiting method, the constraints on statistical delay service quality in wireless Internet of Things communication are solved under the constraints of statistical delay service quality in wireless Internet of Things communication. The user scheduling and power allocation problems have achieved the maximum arrival rate of non-priority user terminals and the arrival rate of priority user terminals, and improved network stability and efficiency.

CN115426690BActive Publication Date: 2024-11-19XI AN JIAOTONG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211071814.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-09-02
Publication Date
2024-11-19
Estimated Expiration
2042-09-02

AI Technical Summary

Technical Problem

In wireless Internet of Things communication, how to maximize the supportable arrival rate of non-priority user terminals while ensuring the quality of statistical delay service, and meet the arrival rate requirements of priority user terminals, especially provided by the energy harvesting module Insufficient and unstable energy supply becomes a challenge.

Method used

By introducing effective bandwidth and effective capacity theory, the data queue and energy queue characteristics are reconstructed, the random optimization problem is transformed into deterministic optimization problem, and the 0/1 planning problem of scheduling variables is solved using the Lagrangian equation and branch delimiting method. The sub-gradient method is used to achieve iterative updates, and the transmission power is reasonably controlled to meet the statistical delay service quality requirements.

Benefits of technology

On the basis of ensuring the quality of statistical delay service, it maximizes the supportable arrival rate of non-priority user terminals, and meets the arrival rate requirements of priority user terminals, improves the stability and efficiency of the energy-collecting wireless network, and reduces the battery Probability of low battery.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115426690B_ABST
    Figure CN115426690B_ABST
Patent Text Reader

Abstract

The invention discloses a user scheduling and power allocation method driven by QoS of an energy-harvesting wireless Internet of Things, comprising: data arrival and departure process under statistical delay service quality guarantee constraint; energy collection and consumption process under statistical battery service quality guarantee constraint; introduction of effective bandwidth and effective capacity theory; conversion of random optimization problem into deterministic optimization problem through feature reconstruction of data queue; conversion of random optimization problem into deterministic optimization problem through feature reconstruction of inverse energy queue; obtaining the two deterministic optimization problems through Lagrange equations, and then performing user scheduling and power allocation driven by QoS of an energy-harvesting wireless Internet of Things according to the solution results. The method can maximize the supportable arrival rate of non-priority user terminals under statistical delay service quality constraint, and meet the arrival rate requirements of priority user terminals at the same time.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the research field of wireless Internet of Things communications and relates to a user scheduling and power allocation method driven by QoS of an energy-harvesting wireless Internet of Things. Background Art

[0002] The Internet of Things (IoT) is a major driving force behind the future development of wireless communication technology and is expected to offer broad application prospects for communication systems. With the rapid growth of IoT applications expected in 2020 and beyond, such as mobile medicine, automotive networks, smart homes, industrial control, and environmental monitoring, hundreds of millions of IoT devices will be connected online, realizing the reality of "everything is connected," creating an unprecedented internet industry and bringing boundless vitality to wireless communications. With the exponential growth of devices in wireless networks, access and transmission demands pose a significant challenge to traditional energy supply. Energy harvesting can provide renewable green energy for communication devices, successfully alleviating the pressure on future wireless communications. Typical renewable energy sources include solar, wind, electromechanical, and hotspot phenomena. Recently, energy harvesting has been recognized as a key environmentally friendly enhancement for wireless networks, with broad application prospects, and has attracted significant interest from both industry and academia.

[0003] However, ensuring the statistical delay quality of service (QoS) requirements of energy-harvesting wireless communication systems presents a significant challenge. On the one hand, traffic data buffered by the transmitter must be transmitted promptly to the receiver, necessitating increased transmit power to ensure the QoS requirements are met even under poor channel conditions. On the other hand, ensuring a relatively stable energy supply for data transmission, as the energy harvesting module provides insufficient and unstable energy for data transmission, presents a more difficult challenge. It is important to note that arbitrarily increasing transmit power can lead to battery depletion or interruption, which can degrade the QoS requirements for data transmission. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art and provide an energy-harvesting wireless Internet of Things QoS-driven user scheduling and power allocation method, which can maximize the supportable arrival rate of non-priority user terminals under the statistical delay service quality constraint while meeting the arrival rate requirements of priority user terminals.

[0005] To achieve the above objectives, the energy harvesting wireless Internet of Things QoS-driven user scheduling and power allocation method described in the present invention includes:

[0006] Statistical delay quality of service constraints under the data arrival and departure process;

[0007] Statistical analysis of energy collection and consumption under battery quality of service constraints;

[0008] Introducing the theory of effective bandwidth and effective capacity;

[0009] By reconstructing the features of the data queue, the random optimization problem is transformed into a deterministic optimization problem;

[0010] By reconstructing the features of the inverse energy queue, the stochastic optimization problem is transformed into a deterministic optimization problem;

[0011] The two deterministic optimization problems are solved by the Lagrange equation, and then energy harvesting wireless Internet of Things QoS-driven user scheduling and power allocation are performed based on the solution results.

[0012] Also includes:

[0013] Normalized channel power gain to describe the channel state information of the corresponding user;

[0014] All channel state information is accurately estimated at the access point and fed back to each user terminal without delay.

[0015] The relationship between the data arrival or departure process and the statistical delay service quality assurance is:

[0016] Determine the data cache queue length based on the data arrival process and data service process;

[0017] Ensure quality of service by statistically analyzing delays and constraining the length of data cache queues.

[0018] The relationship between the energy harvesting or consumption process and the statistical battery service quality assurance is:

[0019] The energy harvesting process and the energy consumption process jointly affect the remaining energy in the battery;

[0020] By statistically guaranteeing the quality of battery service, the remaining energy in the battery is constrained, the probability of low battery is limited, and the probability of low battery is controlled within the preset target value.

[0021] The effective bandwidth and effective capacity theory introduced are:

[0022] The queue index θ is determined by the arrival process and the departure process;

[0023] To ensure the constraint of the queue length distribution probability, the effective bandwidth of the arrival process is guaranteed to be no greater than the effective capacity of the departure process at the target queue index.

[0024] By reconstructing the characteristics of the data queue, we transform the stochastic optimization problem into a deterministic optimization problem. We use the concept of queue index to obtain the buffer index corresponding to each user terminal. This buffer index is used to constrain the data arrival and service processes, transforming the stochastic optimization problem into a deterministic optimization problem.

[0025] When the queue length threshold is the same, the larger the buffer index is, the smaller the target delay violation probability is, and the stricter the statistical delay service quality guarantee is.

[0026] By reconstructing the characteristics of the inverse energy queue, we transform the random optimization problem into a deterministic optimization problem. We then use virtual inverse queuing technology to transform the unstable battery queue into a stable queuing system. We also use the concept of queue index to obtain the energy index corresponding to each user terminal. This energy index is used to constrain the energy collection and consumption processes, transforming the random optimization problem into a deterministic one.

[0027] When the remaining energy threshold is the same, the larger the energy index is, the smaller the target battery outage probability is, and the stricter the statistical delay service quality guarantee is.

[0028] In the specific process of solving the two optimization problems by Lagrange equations, the random optimization problem is transformed into a deterministic convex optimization problem, and the Lagrange equation of the optimization problem is obtained;

[0029] According to the branch and bound method, the above 0 / 1 planning problem of scheduling variables is solved by a deterministic convex problem in two stages: user terminal and access point.

[0030] The subgradient method is used to achieve iterative updates.

[0031] The present invention has the following beneficial effects:

[0032] The energy-harvesting wireless Internet of Things QoS-driven user scheduling and power allocation method described in the present invention assumes that multiple user terminals communicate with an access point through a wireless channel. Each user terminal is equipped with an energy harvesting block that can obtain renewable green resources from the surrounding environment. The access point can use the diversity of channels and energy states to schedule user terminals in a timely manner. The scheduled user terminals need to reasonably control the transmission power based on the channel and energy state to meet the statistical delay service quality requirements of data transmission. Unlike other early works, based on the user's demand for statistical delay service quality assurance, the user terminals are divided into priority and non-priority through the inverse sorting technology and effective bandwidth or effective capacity theory, and a deterministic 0 / 1 planning problem is specified. The goal is to maximize the supportable arrival rate of non-priority user terminals under the statistical delay service quality constraint, while meeting the arrival rate requirements of priority user terminals, and achieve optimal control in the energy-harvesting wireless network. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a model diagram of a single-cell uplink multi-user energy harvesting wireless network system in the present invention;

[0034] Figure 2is a performance comparison curve diagram of the effective capacity of UE1 and the delay violation probability of UE2 in the present invention;

[0035] Figure 3 This is a performance comparison curve diagram of the effective capacity and average energy collection rate of UE1 in the present invention;

[0036] Figure 4 is a graph showing the relationship between the effective capacity and energy index of UE1 in the present invention;

[0037] Figure 5 This is a curve diagram showing the relationship between the delay violation probability and the queue threshold in the present invention. DETAILED DESCRIPTION

[0038] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only embodiments of a part of the present invention, not all embodiments, and are not intended to limit the scope of the present invention. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion of the concepts disclosed in the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present invention.

[0039] The accompanying drawings illustrate schematic diagrams of the structures of the disclosed embodiments of the present invention. These figures are not drawn to scale; for the purpose of clarity, some details are exaggerated and some details may be omitted. The shapes of the various regions and layers shown in the figures, as well as their relative sizes and positional relationships, are merely exemplary and may deviate in practice due to manufacturing tolerances or technical limitations. Those skilled in the art may design regions / layers with different shapes, sizes, and relative positions as needed.

[0040] refer to Figure 1 and Figure 2 In the energy-harvesting wireless IoT QoS-driven user scheduling and power allocation method described in this invention, access points can schedule user terminals using channel and energy status. Scheduled user terminals are required to reasonably control their transmit power to meet the statistical delay quality of service (QoS) requirements for data transmission. Based on these requirements, user terminals are classified into priority and non-priority categories using inverse sorting techniques and effective bandwidth or effective capacity theory. A deterministic 0 / 1 programming problem is then formulated to maximize the supportable arrival rate of non-priority user terminals while meeting the arrival rate requirements of priority user terminals, subject to statistical delay QoS constraints.

[0041] refer to Figure 1 and Figure 2The energy harvesting wireless Internet of Things QoS-driven user scheduling and power allocation method of the present invention comprises the following steps:

[0042] 1) Construct a single-cell uplink multi-user energy harvesting wireless network system model:

[0043] The specific process of step 1) is:

[0044] 1a) Modeling a single-cell uplink multi-user energy harvesting wireless network, where M users communicate with an access point using time division multiple access, all user terminals and access points have single antennas, and the channel between each transceiver fades over the transmission frame;

[0045] 1b) Each user is equipped with an energy harvesting module that can collect renewable energy from the surrounding environment, convert this raw energy into electrical energy, and store it in a battery to provide energy for wireless transmission;

[0046] 1c) In the kth transmission frame, the scheduling variable of the mth user terminal is expressed as d m,k ∈{0,1}, where d m,k =1 indicates that the user terminal works in this transmission frame; d m,k =0 indicates idle state. Scheduling variable d m,k , The following restrictions apply:

[0047]

[0048] 1d) In the kth transmission frame, the normalized channel power gain between the mth user terminal and the access point is Used to describe channel state information. Where L is the large-scale path fading, assuming that all UEs in all frames are the same; h m,k is a small-scale fading, obeys the Rayleigh distribution, and is independent and identically distributed in different transmission frames during the entire transmission process; N0 is the power spectral density of additive white Gaussian noise;

[0049] 1e) Let all channel state information γ m,k , k can be accurately estimated at the access point, and the user terminal can obtain perfect channel state information and dynamically adjust the transmit power μ m,k .

[0050] 2) Data arrival and departure process under statistical delay quality of service constraints:

[0051] The specific process of step 2) is:

[0052] 2a) Data traffic flows at a constant rate v m(bits / s / Hz) arrives and is stored in a data buffer of finite size. Therefore, the data arrival process of the mth user is {TW vm}, where W (Hz) is the uplink bandwidth of the system and T (s) is the time interval between each transmission frame;

[0053] 2b) According to Shannon's formula, using adaptive modulation and coding (AMC), the achieved throughput of the mth UE in the kth transmission frame is:

[0054] R m,k =TW log2(1+μ m,k γ m,k )bits / frame

[0055] Therefore, the data service process of the mth user is expressed as {d m,k R m,k};

[0056] 2c) The change of data buffer queue length on the transmission frame is:

[0057] D m,k+1 =max{D m,k +TW vm -d m,k R m,k ,0}

[0058] Among them, D m,k is the data buffer queue length of the mth user terminal at the beginning of the kth transmission frame;

[0059] 2d) Use statistical delay quality of service to constrain queue length, that is, control queuing delay and keep the delay violation probability within a small target value, that is,

[0060]

[0061] Among them, ξ m is the target delay violation probability, D yh The delay violates the threshold.

[0062] 3) Statistical analysis of energy collection and consumption under battery quality of service constraints:

[0063] The specific process of step 3) is:

[0064] 3a) The energy collection process of the mth user in the kth transmission frame is: Energy values ​​of all user terminals are all independent and equal to κ or 0 with probability ρ or 1-ρ;

[0065] 3b) The energy consumption process of the mth user in the kth transmission frame is: Energy consumption value for:

[0066]

[0067] Where η is the constant circuit power consumption;

[0068] 3c) The change of the remaining battery power during the transmission frame is:

[0069]

[0070] Among them, B m,k is the remaining battery capacity of the mth user terminal at the beginning of the kth transmission frame;

[0071] 3d) Ensure the continuity of energy supply by statistically guaranteeing the battery service quality and limiting the probability of low battery to a small target value, i.e.

[0072]

[0073] Among them, ∈ m is the target’s low battery probability, B th The low battery threshold.

[0074] 4) Establishment and defects of stochastic optimization problem:

[0075] The specific process of step 4) is:

[0076] 4a) Assume there is only one non-priority user terminal, represented by

[0077] 4b) The goal of the problem is to establish the priority of the user terminal The minimum rate requirement and statistical delay service quality guarantee are used to maximize the data arrival rate. Therefore, the optimization problem is expressed as:

[0078]

[0079]

[0080]

[0081] D m,k+1 =max{D m,k +TW vm -d m,k R m,k ,0}

[0082]

[0083]

[0084]

[0085] Among them, C m Represents the supportable data arrival rate required by the priority user terminal;

[0086] 4c) For this stochastic optimization problem, the data queue and battery queue in each transmission frame must be dynamically tracked to optimize the control power resources. This can be expressed as a Markov decision process. However, this algorithm requires a large amount of memory and a complex learning process, and the state value function iteration process converges slowly. Therefore, the present invention utilizes the effective bandwidth and effective capacity theory to transform this stochastic optimization problem into a deterministic optimization problem.

[0087] 5) Introducing the theory of effective bandwidth and effective capacity to transform the random optimization problem into a deterministic optimization problem:

[0088] The specific process of step 5) is:

[0089] 5a) In the theory of effective bandwidth and effective capacity, for a stable queue system, when the arrival and departure processes of the queue are independent and time-independent, the queue length Q k The distribution probability of is approximately expressed as:

[0090]

[0091] Among them, θ is a positive real number, called the queue index, from the queue to the process Q in and leaving process Q out Determine the equation Ψ(Q in ,θ)=Φ(Q out ,θ) where

[0092]

[0093] Among them, Ψ(Q in ,θ) is the effective bandwidth of the arrival process, Φ(Q out ,θ) is the effective capacity of the leaving process.

[0094] 5b) To ensure the constraint of queue length distribution probability, it is necessary to ensure that the effective bandwidth of the arrival process is no greater than the effective capacity of the departure process under the target queue index, that is,

[0095] Ψ(Q in ,θ)≤Φ(Q out ,θ)

[0096] 6) Based on the theory of effective bandwidth and effective capacity, the data queue characteristics are reconstructed:

[0097] The specific process of step 6) is:

[0098] 6a) Re-express the target queue index θ as:

[0099]

[0100] in, is the buffer index of the mth user terminal, with the same threshold D th hour, The larger is the target delay violation probability ξ m The smaller the value, the stricter the statistical delay service quality assurance is.

[0101] 6b) The effective capacity function of the data service process {d m,k R m,k} can be calculated as:

[0102]

[0103] 6c) Due to the data arrival rate {TW vm} is constant, then we have

[0104]

[0105] 7) Based on the effective bandwidth and effective capacity theory, the inverse energy queue characteristics are reconstructed:

[0106] The specific process of step 7) is:

[0107] 7a) Using virtual reverse queuing technology to transform the unstable battery queue into a stable queuing system, where the energy collection and consumption processes play the role of the departure process and arrival process respectively, the statistical battery service quality guarantee can be reformulated as:

[0108]

[0109] in, is the inverse energy queue length, B e is the battery capacity.

[0110] 7b) Re-express the target queue index θ as

[0111]

[0112] in, is the energy index of the mth user terminal, with the same threshold B th hour, The larger the target battery outage probability ∈ m The smaller the value, the stricter the statistical delay service quality assurance is.

[0113] 7c) The effective capacity function of the energy harvesting process is expressed as The effective bandwidth function of the energy consumption process is expressed as Then there is

[0114]

[0115] 8) Reframe the stochastic optimization problem P1 as a deterministic optimization problem and simplify it:

[0116] The specific process of step 8) is:

[0117] 8a) Combining the equivalent constraints of statistical delay and battery service quality assurance, P1 is re-planned as:

[0118]

[0119]

[0120]

[0121]

[0122] 8b) Due to For a given fixed value of the energy harvesting process, the function log(·) increases monotonically, and the problem P2 can be reformulated as:

[0123]

[0124]

[0125]

[0126]

[0127] in, is the normalized buffer index,

[0128] To determine the value, it is obvious that problem P3 is a convex optimization problem and can be solved by the Lagrangian method.

[0129] 9) Solve the optimization problem through the Lagrange equation and obtain the best solution:

[0130] The specific process of step 9) is:

[0131] 9a) The Lagrangian equation for the optimization problem P3 is:

[0132]

[0133] Among them, the vector μ=(μ1,…,μ M) contains the M-tuple transmit power; vector α=(α1,…,α M ) contains the M-tuple Lagrange multipliers associated with the first constraint, and for convenience we define Vector β=(β1,…,β M ) contains the M-tuple Lagrange multipliers associated with the second constraint;

[0134] 9b) According to the branch and bound method, the above scheduling variable d m,k The 0 / 1 programming problem can be solved by a two-stage problem:

[0135] ① User terminal problem: When the mth user terminal is scheduled, that is, d m =1, then the transmit power must be determined At this time, the Lagrangian function is simplified to:

[0136]

[0137] Then, take L m (μ m ,α,β) about μ m The derivative of , let it be 0, that is

[0138]

[0139] The optimal power can be solved as:

[0140]

[0141] Where Ω(X) is the wrightOmega function, which is the solution of the equation Y+log Y=X.

[0142]

[0143] At the same time, the optimal power Limited by the remaining charge in the battery:

[0144]

[0145] ②Access point problem: In each frame, the AP selects the mth * UE, can achieve the minimum target Lagrangian function value Right now

[0146]

[0147] 9c) Adopting adaptive joint user scheduling and power allocation algorithm

[0148] To ensure that the Lagrange multipliers (α1,…,α M ) and (β1,…,βM ) converges to the optimal value, and the subgradient method is used to implement the iterative update of the transmission frame. The Lagrange multiplier α m , β m , Updated as follows:

[0149]

[0150]

[0151] Among them, Δ and Λ are the update step sizes, so the power μ m and the Lagrange multiplier α m and β m can converge to the optimal state.

[0152] Simulation experiment

[0153] The non-priority user terminal is represented as UE1, the priority user terminal is represented as UE2, and the minimum supported rate is set to C2 = 1 bit / s / Hz.

[0154] Benchmark solutions: (1) MDP solution: a solution based on Markov decision process; (2) Round Robin solution: the access point schedules the user terminal, and the specific scheduling ratio should ensure the minimum rate requirement of UE2.

[0155] Depend on Figure 2 As shown, consider two scenarios: ξ1=0.1 in scenario 1 and ξ1=0.001 in scenario 2. It can be seen that the effective capacity of UE1 in scenario 1 is greater than that in scenario 2, because in the same transmission process, The smaller ξ2, the greater the effective capacity. Furthermore, for all schemes, the effective capacity for UE1 increases with increasing ξ2, because the more relaxed latency QoS requirements for UE2 allow the system to schedule more frames for UE1. Note that the scheduling ratio for UE2 in the round-robin scheme decreases with increasing ξ2, to meet only the minimum rate requirement for UE2 in the simulation. Furthermore, this scheme achieves the maximum effective capacity in both cases, demonstrating its superiority.

[0156] Depend on Figure 3 , setting ξ1 = 0.1 in Scenario 1 and ξ1 = 0.001 in Scenario 2, and ξ2 = 0.001 in both scenarios, with ∈1 = ∈2 = 0.1. It can be seen that the effective capacity of UE1 in all scenarios increases with the average rate of energy harvesting, because all user terminals can allocate more power in a good energy state. Furthermore, the increase is more pronounced when the latency QoS constraint is relaxed, that is, in Scenario 1 with ξ1 = 0.1.

[0157] Depend on Figure 4As shown, set The simulations vary from 10 to 210. It can be seen that in all cases of ξ1 or ξ2, the effective capacity first increases with the energy exponent and then decreases. The results show that appropriate statistical battery quality of service assurance can reduce the occurrence of battery outages and help improve transmission stability and effective capacity.

[0158] Depend on Figure 5 As shown, set ∈1=∈2=0.1, the results show that under different thresholds, the actual delay violation probability always destroys the theoretical value, which indicates that all units meet the requirements of statistical delay service quality assurance.

Claims

1. A user scheduling and power allocation method driven by QoS of energy harvesting wireless Internet of Things, characterized in that: include: Statistical delay quality of service constraints under the data arrival and departure process; Statistics on energy collection and consumption process under battery quality of service guarantee constraints; Introducing the theory of effective bandwidth and effective capacity; By reconstructing the features of the data queue, the random optimization problem is transformed into a deterministic optimization problem; Through the feature reconstruction of the inverse energy queue, the random optimization problem is transformed into a deterministic optimization problem; The two deterministic optimization problems are solved by Lagrange equations, and then energy harvesting wireless Internet of Things QoS-driven user scheduling and power allocation are performed according to the solution results; According to the theory of effective bandwidth and effective capacity, the data queue characteristics are reconstructed. The specific process is as follows: Re-express the target queue index θ as: in, is the buffer index of the mth user terminal, with the same threshold D th hour, The larger the target delay violation probability ζ is, the m The smaller the value, the stricter the statistical delay service quality guarantee is. The effective capacity function of the data service process {d m,k R m,k } can be calculated as: Since the data arrival rate {TWv m } is constant, then we have According to the theory of effective bandwidth and effective capacity, the specific process of reconstructing the inverse energy queue characteristics is: The virtual reverse queuing technology is used to change the unstable battery queue to a stable queuing system, where the energy collection and consumption processes play the role of the departure process and the arrival process respectively. The statistical battery service quality guarantee is re-expressed as: in, is the inverse energy queue length, B e is the battery capacity; Re-express the target queue index θ as in, is the energy index of the mth user terminal, with the same threshold B th hour, The larger the target battery power failure probability ∈ m The smaller it is, the stricter the statistical delay service quality guarantee is; 7c) The effective capacity function of the energy harvesting process is expressed as The effective bandwidth function of the energy consumption process is expressed as Then there is The optimization problem is expressed as: in, is the normalized buffer index, To determine the value, it is a convex optimization problem, which is solved by the Lagrangian method. T is the time interval of each transmission frame. Assume that there is only one non-priority user terminal, which is expressed as Priority user terminal In the kth transmission frame, the scheduling variable of the mth user terminal is represented by d m,k ∈{0,1}, where d m,k =1 indicates that the user terminal works in this transmission frame; d m,k =0 indicates idle state, scheduling variable The achieved throughput of the mth UE in the kth transmission frame is R m,k , the energy collection process of the mth user in the kth transmission frame is Energy values ​​of all user terminals are independent, the energy consumption process of the mth user in the kth transmission frame is: Energy consumption value C m represents the supportable data arrival rate required by the priority user terminal. In the kth transmission frame, the dynamically adjusted transmission power μ of the mth user terminal m,k .

2. The method for user scheduling and power allocation driven by QoS of energy harvesting wireless Internet of Things according to claim 1 is characterized in that: Also includes: Normalize the channel power gain to describe the channel state information of the corresponding user; All channel state information is accurately estimated at the access point and fed back to each user terminal without delay.

3. The method for user scheduling and power allocation driven by QoS of energy harvesting wireless Internet of Things according to claim 1 is characterized in that: The relationship between the data arrival or departure process and the statistical delay service quality assurance is: Determine the data cache queue length based on the data arrival process and the data service process; The quality of service is guaranteed by statistical delay and the length of data cache queue is constrained.

4. The method for user scheduling and power allocation driven by QoS of energy harvesting wireless Internet of Things according to claim 1, characterized in that: The relationship between the energy collection or consumption process and the statistical battery service quality assurance is: The energy harvesting process and the energy consumption process together affect the remaining energy in the battery; By statistically guaranteeing the quality of battery service, the remaining energy in the battery is constrained, the probability of low battery is limited, and the probability of low battery is controlled within the preset target value.

5. The method for user scheduling and power allocation driven by QoS of energy harvesting wireless Internet of Things according to claim 1, characterized in that: The effective bandwidth and effective capacity theory introduced are: The queue index θ is determined by the arrival process and the departure process; To ensure the constraint of queue length distribution probability, the effective bandwidth of the arrival process is guaranteed to be no greater than the effective capacity of the departure process at the target queue index.

6. The method for user scheduling and power allocation driven by QoS of energy harvesting wireless Internet of Things according to claim 1, characterized in that: In the process of transforming the random optimization problem into a deterministic optimization problem through feature reconstruction of the data queue, the buffer index corresponding to each user terminal is obtained through the concept of queue index. The buffer index is used to constrain the data arrival process and service process, and the random optimization problem is transformed into a deterministic optimization problem. When the queue length threshold is the same, the larger the buffer index is, the smaller the target delay violation probability is, and the stricter the statistical delay service quality guarantee is.

7. The method for user scheduling and power allocation driven by QoS of energy harvesting wireless Internet of Things according to claim 1, characterized in that: Through the feature reconstruction of the inverse energy queue, in the process of converting the random optimization problem into a deterministic optimization problem, the virtual inverse queuing technology is used to change the unstable battery queue into a stable queuing system; through the concept of queue index, the energy index corresponding to each user terminal is obtained, and the energy index is used to constrain the energy collection process and consumption process, and the random optimization problem is converted into a deterministic optimization problem; When the remaining energy threshold is the same, the larger the energy index is, the smaller the target battery outage probability is, and the stricter the statistical delay service quality guarantee is.