A Multi-Sensor Network Resource Scheduling Method Based on Multi-Radio Signal Energy Harvesting

Abstract

The present invention relates to a multi-sensor network resource scheduling method based on multi-radio frequency signal energy harvesting, specifically as follows: Multiple source nodes broadcast radio frequency signals with different transmission powers, and multiple passive sensors convert the radio frequency signals into available electrical energy for themselves through energy harvesting technology; Under the time division multiplexing protocol, the multi-sensors use the harvested energy to sequentially perform short packet transmissions to the same information processing terminal with a certain transmission code length, and jointly optimize the transmission powers of the multiple source nodes and the short packet transmission code lengths of the multi-sensors to minimize the maximum block error rate in all sensors' short packet transmissions. The present invention uses multi-radio frequency signal energy harvesting technology, does not need to rely on traditional power supply, and can achieve energy harvesting and resource allocation in places without power supply. In addition, the fairness scheduling algorithm of this method can flexibly adapt to multi-sensor networks with different service requirements, improving resource utilization and system efficiency while ensuring fairness of resource allocation.

Description

Technical Field

[0001] This application relates to the field of wireless communication, and particularly to a multi-sensor network resource scheduling method based on multi-radio frequency signal energy harvesting. Background Art

[0002] In the future sixth-generation (6G) communication system, extremely large-scale connectivity is one of the main performance indicators for realizing the so-called Internet of Everything (IoE). At the same time, the explosive number of connected devices will pose new requirements in two aspects. On the one hand, how to provide stable and low-cost energy supply for a large number of devices. On the other hand, how to allocate limited communication resources among a vast number of devices to achieve efficient communication. Traditional power supply methods, namely batteries and cables, have limitations in terms of lifespan, capacity, etc., and laying cables and replacing batteries (by humans or robots) will cause great inconvenience and costs, especially in a network of a vast number of devices. In this regard, radio frequency-based wireless power transfer (WPT) technology can convert the received radio frequency signals into power resources in a controllable and stable manner, thereby realizing instant charging of devices and meeting actual service requirements. More importantly, multiple devices can obtain energy from the broadcast radio frequency signals through energy harvesting (EH) technology, which is beneficial to the power supply in a large-scale device network.

[0003] Compared with the wireless energy transfer of a single source node, multiple source nodes can further improve the system energy efficiency. The channel diversity introduced by multiple source nodes can well compensate for the great negative impacts brought by some channel fades. However, the joint wireless energy transfer of multiple source nodes also poses great challenges to system analysis. On the one hand, mutual interference will occur between the radio frequency signals received by the receiving node from multiple source nodes, which will have an unpredictable impact on the WPT energy efficiency of the system. On the other hand, when the total available resources are limited, different resource scheduling methods will lead to different system performances, especially when the channel qualities and transmission requirements of different devices, source nodes, and terminal nodes are different. Traditional power and blocklength allocation schemes are difficult to achieve the goal of maximizing multi-user transmission reliability. Summary of the Invention

[0004] The present invention proposes a multi-sensor network resource scheduling method based on multi-radio frequency signal energy harvesting. Multiple sensors obtain the energy required for their short-packet transmission by harvesting the radio frequency signals of multiple source nodes. By optimizing the transmission power of the source nodes and the short-packet transmission blocklength of the sensors, the system fairness and reliability are improved.

[0005] The technical solution of the present invention is a multi-sensor network resource scheduling method based on multi-radio frequency signal energy harvesting, comprising the following steps:

[0006] Step 1: Determine the number of nodes in the multi-sensor wireless energy transfer system, including the location information of multi-source nodes, multi-sensors, and information processing terminals; determine the channel quality, including the channel state information between multi-source nodes and multi-sensors and the channel state information between multi-sensors and data processing terminals; determine the packet size of multi-sensor information transmission according to actual service requirements;

[0007] Step 2: Integrate and process the received multi-radio frequency signals using an actual non-linear energy harvesting model based on circuits for energy harvesting, and construct a mathematical model between the energy harvested by users and the transmission power of multi-source nodes based on this non-linear energy harvesting model to obtain the energy harvested by users;

[0008] Step 3: The multi-sensors sequentially transmit their data short packets to the same terminal in a time-division multiplexing manner by consuming the energy harvested in the first stage. Fairly consider the reliable transmission of all sensors, and use the maximum block error rate of all sensor short packet transmissions as a measure of system reliability. Based on this, construct a mathematical model between the block error rate of sensor short packet transmission, the code length of short packet transmission, and the transmission power of source nodes;

[0009] Step 4: Under the condition that the total power of multi-source nodes and the total code length of multi-sensor short packet transmission are limited, by jointly optimizing the transmission power of multi-source nodes and the code length of multi-sensor short packets, construct an optimization problem to minimize the maximum block error rate of sensor short packet transmission, and decompose the original problem into two sub-problems: power allocation problem and code length allocation problem;

[0010] Step 5: In the power allocation problem, fix the code length of sensor short packet transmission, decouple the direct association of source nodes with system reliability, introduce a relaxation variable, that is, the transmission power of sensors, and then use the successive convex approximation algorithm to shape the original problem into multiple convex local problems, and use the iterative algorithm to obtain a sub-optimal solution to the power allocation problem;

[0011] Step 6: In the code length allocation problem, fix the transmission power of multi-source nodes, perform successive convex approximation on the non-convex constraints in the constructed problem, and optimize the code length of each sensor through the iterative algorithm under the condition that the total code length of sensor short packet transmission is limited to obtain a sub-optimal solution to the code length allocation problem;

[0012] Step 7: Alternately execute the power allocation in Step 5 and the code length allocation in Step 6, and update the scheduling strategy of power and code length based on each round of iteration. If the difference between the minimized maximum transmission block error rate obtained by iterative optimization and the result of the previous round of iteration is less than the convergence condition, the iteration ends, and the power and code length resource allocation results of this wireless power supply communication network are obtained; otherwise, return to Step 5 and continue the alternating optimization of power and code length in the next round.

[0013] Further, in the above Step 1, the source node set and the sensor set are respectively defined as:

[0014] S = {S j}, j ∈ {1, ..., M}

[0015] U = {U i}, i ∈ {1, ..., N}

[0016] where j is the source node number, S j represents the source node numbered j, M represents the number of source nodes, i is the sensor node number, U i represents the sensor numbered i, N represents the number of sensor nodes, and the terminal node is represented as D.

[0017] Furthermore, in the channel gain described in step 1, the channel gain between the j-th source node S j and the i-th sensor U i is defined as z i,j , and the channel gain between the i-th sensor U i and the terminal D is defined as

[0018] The multi-sensor short packet data packet set described in step 1 is defined as

[0019] k = {k i}, i ∈ {1, ..., N}

[0020] where k i represents the amount of data in the short data packet transmitted from sensor i, in bits.

[0021] Furthermore, in step 2, during the energy transfer phase, M source nodes simultaneously transmit RF signals to N users in a broadcast manner. The transmission power set of the M source nodes is defined as:

[0022] P = {P j}, j ∈ {1, ..., M}

[0023] where P j represents the transmission power of the source node S j . Correspondingly, each user U i will receive a set of RF signals, and the power is defined as:

[0024]

[0025] where Q i,j (P j ) represents the power obtained from the j-th source node, i.e., S j , and it is defined as:

[0026] Q i,j (P j) = P j z i,j 。

[0027] Furthermore, the non - linear energy harvesting model constructed in step 2 is defined as:

[0028]

[0029] where P dc,i represents the DC signal power obtained by user U i through energy harvesting, and F eh (Q i ) represents the functional relationship between the DC signal power P i output during the energy harvesting process of user U dc,i and the RF signal power Q i . The constant a is defined as where represents the load impedance, represents the ideality factor, v t represents the thermal voltage, W0(·) represents the branch of the Lambert W function, which is the inverse function of f(x) = xe x , I s represents the reverse - bias saturation current, is the sum of monomials about Q i , that is, the positive polynomial of Q, and is defined as:

[0030]

[0031] where n0 represents the truncation constant, M represents the number of received RF signals, and the constant where is the constant coefficient, and is defined as I s represents the reverse - bias saturation current, is a constant value, represents the ideality factor, v t represents the thermal voltage; R ant represents the matched antenna impedance, and the expression represents any non - negative sequence whose sum is equal to the constant is defined as:

[0032]

[0033] where represents the constant waveform coefficient of unit power, and the expression reflects the interference between multiple received RF signals, represents the non - negative sequence the m-th value in, Q m represents the power of the m-th received radio frequency signal;

[0034] Based on the non-linear energy harvesting model, user U i The harvested energy is defined as:

[0035] E i (P) = P dc,i n0T s = F eh (Q i )n0T s

[0036] where E i (P) represents the energy harvested by U i in the first stage, n0 represents the code length of energy harvesting in the first stage, in symbols, T s represents the time length of each symbol, in seconds.

[0037] Furthermore, in step 3,

[0038] The multi-user short packet transmission code length set is defined as:

[0039] n = {n i},i ∈ {1,..., N}

[0040] The transmission power of user U i is defined as:

[0041]

[0042] where P dc,i represents the energy harvesting power of user U i n i represents the code length of short packet transmission of user U i Based on this, the transmission signal-to-noise ratio of user U i is defined as:

[0043]

[0044] where σ 2 represents the noise power. To ensure reliable communication, the constraint of the user short packet transmission signal-to-noise ratio is defined as follows:

[0045] γ i ≥ 1, i ∈ {1,…, N}

[0046] That is, the signal-to-noise ratio of all short packet transmissions is not less than 1;

[0047] The transmission error rate of user U i is defined as:

[0048]

[0049] where represents the data transmission rate, with the unit of bit / symbol, and Q(·) represents the Gaussian Q function, that is The expression C(γ i ) is defined as C(γ i ) = log(1 + γ i ), representing the ideal Shannon channel capacity. The expression V(γ i ) is defined as represents the channel scattering. According to this formula, the block error rate ε i of the short packet transmission of user U i is affected jointly by the transmission power P of all source nodes and the code length n i of the short packet transmission of user U i , and is expressed as:

[0050]

[0051] Considering the reliability transmission of all users fairly, the maximum transmission bit error rate among all users is used as a measure of the system reliability, and its definition is:

[0052]

[0053] Furthermore, an optimization problem of minimizing the maximum bit error rate of users is constructed, and the objective of the optimization problem is defined as:

[0054]

[0055] The resource optimization problem is defined as follows:

[0056]

[0057]

[0058]

[0059]

[0060] 0 ≤ P j ≤ P max

[0061]

[0062] where represents the signal-to-noise ratio of the short packet transmission from user U i to the terminal, k i represents the packet size of user U i , and n i represents the user U iThe code length of short packet transmission. P total Represents the upper limit of the total transmission power of multiple source nodes in the wireless energy transmission stage; P max Represents the maximum transmission power of a single source node in the wireless energy transmission stage; n total Represents the upper limit of the total code length of short packet transmission when multiple users perform short packet transmission in a time-division multiplexing manner in the short packet transmission stage.

[0063] Furthermore, the power optimization problem in step 5 is defined as follows:

[0064]

[0065]

[0066]

[0067]

[0068] 0 ≤ P j ≤ P max

[0069] Decouple the direct correlation between the transmission power of the source node and the system reliability, and introduce a slack variable That is, take the transmission power of the user as the optimization variable, perform variable substitution on the transmission power of the source node, that is, introduce the variable Where

[0070] Based on the above steps, for user U i The short packet transmission error probability ε i Is defined as a function of the user's transmission power:

[0071]

[0072] The original equation relationship between the user's transmission power and multiple source nodes, that is

[0073] Is transformed into an inequality constraint, defined as:

[0074]

[0075] Where Represents the DC current power obtained by user U i Through energy harvesting. The physical meaning of this constraint is that the total energy consumption of user U i For short packet transmission is not greater than the total energy obtained through energy harvesting in the first stage. Based on this relaxation, sub-problem (SP1) is transformed into (SP2), and (SP2) is defined as follows:

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082] In problem (SP2), for the first constraint, continuous convex approximation is performed as follows:

[0083] Based on the property Regarding Joint convexity [2], the following inequality can be obtained:

[0084]

[0085] where the inequality holds when and the constants are defined as follows respectively:

[0086]

[0087]

[0088] Based on the above operations, the original non-convex constraint is transformed into a convex constraint, that is Meanwhile, problem (SP2) is transformed into multiple convex local problems; the local problem in the τ-th iteration is defined as:

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095] The optimal solution of the local convex problem can be obtained by the interior point method. By iterating the local problem multiple times, a sub-optimal solution of sub-problem (SP1) is obtained.

[0096] Furthermore, in step 6, in the code length allocation problem, the relationship between the transmission error probability and the code length is defined as:

[0097]

[0098] where the constant is defined as When the transmission rate r i ≥0.0683, γ i ≥1, that is, ε i is convex with respect to n i According to the properties of convex functions, it can be obtained that is jointly convex with respect to {n i}, i ∈ {1,..., N};

[0099] Perform continuous convex approximation on the non-convex constraint 2, which is defined as follows:

[0100]

[0101] The equation holds when n = n (χ) , where the constants and are respectively defined as:

[0102]

[0103]

[0104] Transform the sub-problem (SP1) into multiple convex local problems (LP2), and the local problem in the χ-th iteration is defined as:

[0105]

[0106]

[0107]

[0108]

[0109] Obtain the optimal solution of the local sub-problem (LP2) through the interior point method, and obtain the sub-optimal solution of the code length allocation problem by iterating the local problem.

[0110] Furthermore, step 7 includes the following sub-steps:

[0111] Step 7.1: Determine the number of source nodes M, the number of user nodes N, and determine the channel state information {z i,j}, i ∈ {1,..., N}, j ∈ {1,..., M} between the source nodes and the multi-user channels, and the channel state information between the multi-user and the data processing terminal Determine the user service requirements, that is, the packet size k = {k i}, i ∈ {1,..., N};

[0112] Step 7.2: Determine the initial feasible solution, that is, initialize the transmission power of multi-source nodes and the code lengths of multi-user short packets ($P$ (0) , $n$ (0) ), and the number of iterations $\delta = 0$;

[0113] Step 7.3: Perform power allocation and initialize the power allocation iteration number $\tau = 0$;

[0114] Step 7.4: Based on ($P$ (τ) , $n$ (τ) ), determine at the local point calculate the constant and perform convex approximation on the expression ; obtain the local optimal solution by solving (LP1)

[0115] Step 7.5: If the minimum maximum transmission error rate obtained in this power optimization iteration is less than $\varepsilon$ compared with the result of the previous iteration converge , stop the iteration and output the power allocation result Otherwise, $\tau=\tau + 1$, go back to Step 7.4;

[0116] Step 7.6: Initialize the transmission power of multi-source nodes with the power allocation result obtained in Step 7.5; initialize the code length allocation iteration number $\chi = 0$;

[0117] Step 7.7: At the local point $n$ (χ) calculate the constant and and perform convex approximation on the expression ; obtain the local optimal solution $n$ * ;

[0118] Step 7.8: If the minimum maximum transmission error rate obtained in this code length optimization iteration is less than $\varepsilon$ compared with the result of the previous iteration converge , stop the iteration and output the power allocation result $n$ * ; Otherwise, $\chi=\chi + 1$, $n$ (χ) =$n$ * , go back to Step 7.7;

[0119] Step 7.9: In the $\delta$-th iteration of power allocation, if the minimum maximum transmission error rate output in Step 7.8 is less than $\varepsilon$ compared with the result of the previous iteration $\delta - 1$ converge , stop the power and code length iteration and output the power and code length allocation result ($P$ * , $n$ * ); Otherwise, $\delta=\delta + 1$, go back to Step 7.3 to start a new round of power and code length optimization.

[0120] Compared with the prior art, the present invention has the following advantages:

[0121] 1. By utilizing the feature that the energy of radio frequency signals can be collected, the present invention constructs a multi-passive sensor short-packet communication network enabled by multi-radio frequency signal energy collection technology, which can be used in a large number of special scenarios where traditional power supplies are not supported.

[0122] 2. The resource scheduling method between multiple sources and multiple sensors can flexibly respond to the different service requirements of sensors and perform reasonable and fair power and code length resource scheduling. While improving the resource utilization rate, this algorithm also improves the fairness and reliability of the system.

[0123] 3. The algorithm has strong scalability. It can not only be used in the joint optimization scenario of power and code length, but also be applicable to scenarios with conditional limitations, such as the scenario where the transmission power of the source node is constant and the scenario of fixed frame structure transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0124] Figure 1 is a schematic diagram of a multi-sensor communication network based on multi-radio frequency signal energy collection in the present invention;

[0125] Figure 2 is a schematic diagram of the process of the present invention; DETAILED DESCRIPTION OF THE INVENTION

[0126] For the convenience of those of ordinary skill in the art to understand and implement the present invention, the following is a further detailed description of the present invention with reference to the accompanying drawings and examples. It should be understood that the embodiments described herein are only for the purpose of illustrating and explaining the present invention, and are not intended to limit the present invention.

[0127] Figure 1 is a schematic diagram of a multi-sensor communication network based on multi-radio frequency signal energy collection in the present invention. The present invention proposes a multi-sensor resource scheduling method based on multi-radio frequency signal energy collection. First, multiple source nodes broadcast radio frequency signals with different transmission powers, and multiple passive sensors convert the radio frequency signals into their own available electrical energy through energy collection technology. Then, under the time division multiplexing protocol, the multiple sensors use the collected energy to perform short-packet transmission to the same information processing terminal in sequence with a certain transmission code length. To ensure the fairness and short-packet transmission reliability among multiple sensors, it is necessary to jointly optimize the transmission power of multiple source nodes and the code length of short-packet transmission of multiple sensors, so as to minimize the maximum block error rate in the short-packet transmission of all sensors. It not only solves the problem of difficult access to safe and stable power supply in special scenarios, ensures the fairness of resource allocation among multiple sensors, but also improves the resource utilization rate and system reliability at the same time. It has the characteristics of low cost, low latency, and high reliability.

[0128] The specific process of the present invention is as follows:

[0129] Step 1: Network topology structure and service requirement analysis: Determine the number of nodes in the multi-sensor wireless energy transmission system, including the location information of multi-source nodes, multi-sensors, and information processing terminals; Based on the existing channel estimation method, determine the channel quality, including the channel state information between multi-source nodes and multi-sensors and the channel state information between multi-sensors and data processing terminals; According to the actual service requirements, determine the traffic volume (packet size) of multi-sensor information transmission.

[0130] Step 2: Energy transmission model construction: In the energy transmission stage, multiple source nodes send RF signals to all sensors in a broadcast form with different transmission powers. The multi-sensors convert the received RF signals from multiple source nodes into DC signals through the energy harvesting process and further convert them into electrical energy. Considering the influence of non-linear elements in the energy harvesting circuit, a practical non-linear energy harvesting model based on the circuit is adopted to integrate and process the received multi-RF signals for energy harvesting. This step constructs a mathematical model between the energy collected by users and the transmission power of multi-source nodes.

[0131] Step 3: Information (short packet) transmission model construction: In the short packet transmission stage, considering the finiteness of the code length in the actual sensing communication scenario, a reliability metric for multi-sensor short packet transmission in the finite code length domain is proposed. Specifically, the multi-sensors consume the energy collected in the first stage and sequentially transmit their data short packets to the same terminal in a time-division multiplexing manner. Fairly considering the reliable transmission of all sensors, the maximum block error rate of all sensor short packet transmissions is used as a measure of system reliability. This step constructs a mathematical model between the block error rate of sensor short packet transmission, the short packet transmission code length, and the source node transmission power.

[0132] Step 4: Construction of a resource joint optimization problem for fairness and reliability: Based on this network, in the case of limited resources, that is, the total power of multi-source nodes and the total code length of multi-sensor short packet transmission are limited, by jointly optimizing the transmission power of multi-source nodes and the short packet transmission code length of multi-sensors, an optimization problem of minimizing the maximum block error rate of sensor short packet transmission is constructed. The constructed resource scheduling problem is a non-convex problem. To solve this problem, the original problem is decomposed into two sub-problems: power allocation problem and code length allocation problem.

[0133] Step 5: Power resource scheduling strategy: In the power allocation problem, fix the short packet transmission code length of the sensors. Decouple the direct association of source nodes with system reliability and introduce a relaxation variable, that is, the transmission power of the sensors. Then, the continuous convex approximation algorithm is used to shape the original problem into multiple convex local problems, and an iterative algorithm is used to obtain a sub-optimal solution to the power allocation problem.

[0134] Step 6: Code length resource scheduling strategy: In the problem of code length allocation, fix the transmission power of multiple source nodes, and perform continuous convex approximation on the non-convex constraints in the constructed problem. In the case where the total code length of sensor short-packet transmission is limited, optimize the code length of each sensor through an iterative algorithm to obtain a sub-optimal solution to the code length allocation problem.

[0135] Step 7: Alternately execute Step 5 (power allocation) and Step 6 (code length allocation), and update the scheduling strategies of power and code length based on each round of iteration. If the difference between the minimized maximum transmission block error rate obtained by iterative optimization and the result of the previous round of iteration is less than the convergence condition, the iteration ends, and the power and code length resource allocation results of the wireless powered communication network are obtained. Otherwise, return to Step 5 and continue the alternating optimization of power and code length in the next round.

[0136] In the above embodiments, the source node set and sensor set described in Step 1 are respectively defined as:

[0137] S = {S j}, j ∈ {1,..., M}

[0138] U = {U i}, i ∈ {1,..., N}

[0139] where j is the source node number, S j represents the source node numbered j, M represents the number of source nodes, i is the sensor node number, and U i represents the sensor numbered i, and N represents the number of sensor nodes. The terminal node is represented as D.

[0140] Among the channel gains described in Step 1, the channel gain between the j-th source node S j and the i-th sensor U i is defined as z i,j , and the channel gain between the i-th sensor U i and the terminal D is defined as

[0141] The multi-sensor short-packet data packet set described in Step 1 is defined as

[0142] k = {k i}, i ∈ {1,..., N}

[0143] where k i represents the data volume of the short data packet transmitted from sensor i, with the unit of bit (bit).

[0144] In the energy transfer phase described in Step 2, M source nodes simultaneously transmit RF signals to N users in a broadcast manner, and the transmission power set of M source nodes is defined as:

[0145] P = {Pj}, j ∈ {1, ..., M}

[0146] Where P j represents the transmission power of the source node S j . Correspondingly, each user U i will receive a set of RF signals, and the power is defined as:

[0147]

[0148] Where Q i,j (P j ) represents the power obtained from the j-th source node, i.e., S j , and it is defined as:

[0149] Q i,j (P j ) = P j z i,j

[0150] The non-linear energy harvesting model described in Step 2 is defined as

[0151]

[0152] Where P dc,i represents the DC signal power obtained by energy harvesting. The constant a is defined as Where represents the load impedance, represents the ideality factor, v t represents the thermal voltage. W0(·) represents the branch of the Lambert W function, which is the inverse function of f(x) = xe x , and I s represents the reverse bias saturation current. is the sum of monomials about Q i , that is, the positive polynomial of Q, and is defined as:

[0153]

[0154] Where n0 represents the truncation constant, and M represents the number of received RF signals. The constant Where is the constant coefficient, and it is defined as I s represents the reverse bias saturation current, is a constant value, represents the ideality factor, v t represents the thermal voltage, and R ant represents the matched antenna impedance. The expression represents the sum of any non-negative sequence is equal to Constant is defined as:

[0155]

[0156] where represents the constant waveform factor of unit power. The expression reflects the interference between multiple received RF signals, represents the m-th value in the non-negative sequence , and Q m represents the power of the m-th received RF signal.

[0157] Based on this non-linear energy harvesting model, the energy harvested by user U in step 2 i is defined as:

[0158] E i (P) = P dc,i n0T s = F eh (Q i )n0T s

[0159] where E i (P) represents the energy harvested by U i in the first stage, P dc,i represents the non-linear energy harvesting model, n0 represents the code length of energy harvesting in the first stage (in symbols), and T s represents the time length of each symbol (in seconds).

[0160] In the short packet transmission stage described in step 3, multiple users sequentially transmit their data short packets to the same terminal in a time-division multiplexing manner by consuming the harvested energy. All users have a stable short packet transmission ability, that is, the transmit power of user U i remains unchanged in its transmission frame, and its value is jointly affected by the energy harvested by user U i and the short packet transmission code length n i .

[0161] The set of short packet transmission code lengths for multiple users described in step 3 is defined as:

[0162] n = {n i}, i ∈ {1,..., N}

[0163] where n i represents the short packet transmission code length of user U i . The transmit power of user U i is defined as:

[0164]

[0165] Among which, P dc,i represents the energy harvesting power of user U i , and n i represents the code length of the short packet transmission of user U i . Based on this, the transmission signal-to-noise ratio (SNR) of user U i is defined as:

[0166]

[0167] Among which, σ 2 represents the noise power, and represents the channel gain between user U i and terminal D. To ensure reliable communication, the constraint on the short packet transmission signal-to-noise ratio of the user is defined as follows:

[0168] γ i ≥1, i ∈ {1, …, N}

[0169] That is, the signal-to-noise ratio of all short packet transmissions is not less than 1.

[0170] The transmission bit error rate of user U described in step 3 i is defined as:

[0171]

[0172] Among which represents the data transmission rate, with the unit of bit / symbol, and Q(·) represents the Gaussian Q function, that is The expression C(γ i ) is defined as C(γ i ) = log(1 + γ i ), which represents the ideal Shannon channel capacity. The expression V(γ i ) is defined as represents the channel scattering. According to this formula, the block error rate ε i of the short packet transmission of user U i is jointly affected by the transmission power P of all source nodes and the short packet transmission code length n i of user U i , and is expressed as:

[0173]

[0174] In step 3, considering the reliability of all users fairly in the transmission, the maximum transmission bit error rate among all users is used as a measure of the system reliability, and its definition is:

[0175] In step 4, when the total power of multi-source nodes and the total code length of multi-user short-packet transmission are limited, an optimization problem of minimizing the maximum bit error rate of users is constructed by jointly optimizing the transmission power of multi-source nodes and the short-packet transmission code length of multi-users. The objective of the optimization problem is defined as:

[0176]

[0177] The optimization objective aims to minimize the maximum transmission error rate of all users by jointly optimizing the transmission power P = {P j}, j ∈ {1,..., M} of multiple source nodes and the short-packet transmission code length n = {n i}, i ∈ {1,..., N}.

[0178] The resource optimization problem in step 4 is defined as follows:

[0179]

[0180]

[0181]

[0182]

[0183] 0 ≤ P j ≤ P max

[0184]

[0185] where represents the signal-to-noise ratio of short-packet transmission from user U i to the terminal. To meet the basic requirements of reliable transmission, the signal-to-noise ratio should be no less than 1; the second constraint represents the Shannon channel capacity constraint, that is, the transmission rate is not greater than the ideal Shannon channel capacity; the third constraint represents the total transmission power constraint of multiple source nodes in the wireless energy transmission stage, and the total power upper limit is defined as P total ; the fourth constraint represents the transmission power constraint of a single source node in the wireless energy transmission stage, and the maximum transmission power is defined as P max ; the fifth constraint represents the total code length (delay) constraint of multi-users for short-packet transmission in a time-division multiplexing manner in the short-packet transmission stage, and the total code length upper limit is defined as n total .

[0186] The original problem (OP) is a non-convex and non-linear problem. On the one hand, the non-linear energy harvesting model makes the block error rate non-jointly convex with respect to the transmission powers of multiple source nodes. On the other hand, under the finite blocklength transmission model, the Q-function makes the problem difficult to solve. To address the above problems, the original problem is decomposed into two sub-problems, namely the power allocation problem and the blocklength allocation problem, aiming to obtain a sub-optimal solution to the original problem by alternately solving the allocation problems and the blocklength allocation problem.

[0187] The power allocation sub-problem described in step 5 is defined as allocating the transmission powers of multiple source nodes while fixing the blocklengths of multi-user short packet transmissions. Among them, the transmission power P of the source node = {P j}, j ∈ {1,..., M} is the optimization variable, and the blocklengths n of multi-user short packet transmissions = {n i}, i ∈ {1,..., N} are constants.

[0188] The power optimization problem described in step 5 is defined as follows:

[0189]

[0190]

[0191]

[0192]

[0193] 0 ≤ P j ≤ P max

[0194] Among them, the objective function is non-convex and the first constraint is non-convex, so this problem is still a non-convex problem.

[0195] The decoupling of the complex relationship and the introduction of relaxation variables described in step 5 are specifically explained as follows: Decouple the direct correlation between the transmission power of the source node and the system reliability, and introduce the relaxation variable That is, take the transmission power of the user as the optimization variable. Perform variable substitution on the transmission power of the source node, that is, introduce the variable Among them

[0196] Based on the above steps, the short packet transmission error probability ε i of user U i is defined as a function of the transmission power of the user:

[0197]

[0198] The equation relationship between the original user transmission power and multiple source nodes, that is is transformed into an inequality constraint, defined as:

[0199]

[0200] wherein represents the DC current power obtained by user U i The physical meaning of the DC current power obtained by energy harvesting represents user U i The total energy consumption of short - packet transmission is not greater than the total energy obtained by energy harvesting in the first stage. Based on this relaxation, sub - problem (SP1) is transformed into (SP2). (SP2) is defined as follows:

[0201]

[0202]

[0203]

[0204]

[0205]

[0206]

[0207] In problem (SP2), the transmission block error rate (ε i ) is only affected by the transmit power of user U i , and ε is convex with respect to i with respect to [1], so the objective function in problem (SP2) is convex. The second, third, fourth, and fifth constraints are all convex constraints. The first constraint is a non - convex constraint. For this, a successive convex approximation is performed on the first constraint. The specific steps are as follows:

[0208] Based on the property with respect to jointly convex [2], the following inequality can be obtained:

[0209]

[0210] where the inequality holds when . The constants are defined as follows respectively:

[0211]

[0212]

[0213] Based on the above operations, the original non - convex constraint is transformed into a convex constraint, that is At the same time, problem (SP2) is transformed into multiple convex local problems. The local problem in the τ - th iteration is defined as:

[0214]

[0215]

[0216]

[0217]

[0218]

[0219]

[0220] The locally convex problem can obtain its optimal solution through the interior point method. By iterating the local problem multiple times, a sub-optimal solution of the sub-problem (SP1) is obtained.

[0221] Step 6 The code length allocation problem is defined as minimizing the maximum transmission error probability among all users by optimizing the code length of multi-user short packet transmission under the premise of given transmission powers of multi-source nodes. In the code length allocation problem, the relationship between the transmission error probability and the code length is defined as:

[0222]

[0223] where the constant is defined as When the transmission rate r i ≥0.0683, γ i ≥1, i.e., ε i is convex with respect to n i According to the properties of convex functions, it can be obtained that is jointly convex with respect to {n i}, i ∈ {1,..., N}.

[0224] Step 6 performs a successive convex approximation on the non-convex constraint 2, which is defined as follows:

[0225]

[0226] The equation holds when n = n (χ) where the constants and are defined as:

[0227]

[0228]

[0229] Step 6 transforms the sub-problem (SP1) into multiple convex local problems (LP2), and the local problem in the χ-th iteration is defined as:

[0230]

[0231]

[0232]

[0233]

[0234] The optimal solution of the local sub - problem (LP2) is obtained by the interior - point method. By iterating the local problems, a sub - optimal solution of the code - length allocation problem is obtained.

[0235] Step 7 alternately executes Step 6 (power allocation) and Step 7 (code - length allocation) until the iteration result converges, and the power and code - length resource allocation results of the wireless power - supply communication network are obtained.

[0236] The resource allocation optimization algorithm described in Step 7 is as follows:

[0237] Step 7.1: Determine the number of source nodes M, the number of user nodes N, and determine the channel state information between the source nodes and multi - user {z i,j}, i ∈ {1,..., N}, j ∈ {1,..., M}, and the channel state information between the multi - user and the data - processing terminal Determine the user service requirements, that is, the data - packet size k = {k i}, i ∈ {1,..., N}.

[0238] Step 7.2: Determine the initial feasible solution, that is, initialize the transmission power of multi - source nodes and the code - length of multi - user short - packet transmission (P (0) , n (0) ) and the number of iterations δ = 0.

[0239] Step 7.3: Perform power allocation and initialize the number of power - allocation iterations τ = 0.

[0240] Step 7.4: Based on (P (τ) , n (τ) ), determine At the local point Calculate the constant And perform a convex approximation on the expression Obtain the local optimal solution by solving (LP1)

[0241] Step 7.5: If the minimum maximum transmission error rate obtained in this power optimization iteration is less than ε converge compared with the result of the previous iteration, stop the iteration and output the power - allocation result Otherwise, τ = τ + 1, Go back to Step 7.4.

[0242] Step 7.6: Initialize the transmission powers of the multi-source nodes with the power allocation results obtained in Step 7.5; initialize the iteration count χ of code length allocation to 0.

[0243] Step 7.7: At the local point n (χ) Calculate the constants and and perform a convex approximation on the expression Obtain the local optimal solution n by solving (LP2) * .

[0244] Step 7.8: If the minimum maximum transmission error rate obtained in this code length optimization iteration is less than ε compared to the result of the previous iteration converge , stop the iteration and output the power allocation result n * ; otherwise, χ = χ + 1, n (χ) = n * , and go back to Step 7.7.

[0245] Step 7.9: In the δ-th iteration of power allocation, if the minimum maximum transmission error rate output in Step 7.8 is less than ε compared to the result of the previous iteration δ - 1 converge , stop the power and code length iteration and output the power and code length allocation results (P * , n * ); otherwise, δ = δ + 1, go back to Step 7.3, and start a new round of power and code length optimization.

[0246] It should be understood that the above description of the preferred embodiment is relatively detailed, and it should not be considered as a limitation to the protection scope of the invention patent. Under the inspiration of the present invention, those of ordinary skill in the art can also make substitutions or deformations without departing from the protection scope defined by the claims of the present invention, and all fall within the protection scope of the present invention. The scope of protection claimed by the present invention shall be subject to the appended claims.

Claims

1. A multi-sensor network resource scheduling method based on multi-radio signal energy harvesting, characterized in that It includes the following steps: Step 1: Determine the number of nodes in the multi-sensor wireless energy transmission system, including the location information of multi-source nodes, multi-sensors, and information processing terminals; determine the channel quality, including the channel state information between multi-source nodes and multi-sensors and the channel state information between multi-sensors and data processing terminals; determine the packet size of multi-sensor information transmission according to the actual service requirements; Step 2: Integrate and process the received multi-radio frequency signals using an actual non-linear energy harvesting model based on circuits for energy harvesting, and construct a mathematical model between the energy harvested by users and the transmission power of multi-source nodes based on this non-linear energy harvesting model to obtain the energy harvested by users; Step 3: The multi-sensors sequentially transmit their data short packets to the same terminal in a time-division multiplexing manner by consuming the energy harvested in the first stage. Fairly considering the reliable transmission of all sensors, take the maximum block error rate of all sensor short packet transmissions as a measure of system reliability, and construct a mathematical model between the block error rate of sensor short packet transmission, the code length of short packet transmission, and the transmission power of source nodes based on this; Step 4: Under the condition that the total power of multi-source nodes and the total code length of multi-sensor short packet transmission are limited, by jointly optimizing the transmission power of multi-source nodes and the code length of multi-sensor short packets, construct an optimization problem to minimize the maximum block error rate of sensor short packet transmission, and decompose the original problem into two sub-problems: power allocation problem and code length allocation problem; Step 5: In the power allocation problem, fix the code length of sensor short packet transmission, decouple the direct association of source nodes with system reliability, introduce a relaxation variable, that is, the transmission power of sensors, and then use the successive convex approximation algorithm to shape the original problem into multiple convex local problems, and use the iterative algorithm to obtain a sub-optimal solution to the power allocation problem; Step 6: In the code length allocation problem, fix the transmission power of multi-source nodes, perform successive convex approximation on the non-convex constraints in the constructed problem, and optimize the code length of each sensor through the iterative algorithm under the condition that the total code length of sensor short packet transmission is limited to obtain a sub-optimal solution to the code length allocation problem; Step 7: Alternately execute the power allocation in Step 5 and the code length allocation in Step 6, and based on the scheduling strategy of updating power and code length in each round of iteration. If the difference between the minimized maximum transmission block error rate obtained by iterative optimization and the result of the previous round of iteration is less than the convergence condition, the iteration ends, and the power and code length resource allocation results of this wireless power supply communication network are obtained; otherwise, return to Step 5 and continue the alternating optimization of power and code length in the next round.

2. A multi-sensor network resource scheduling method based on multi-radio signal energy harvesting according to claim 1, characterized in that, In the above Step 1, the source node set and the sensor set are respectively defined as: S = {S j}, j ∈ {1,..., M} U = {U i}, i ∈ {1,..., N} Among them, j is the source node number, and S j represents the source node numbered j, M represents the number of source nodes, i is the sensor node number, and U i represents the sensor numbered i, N represents the number of sensor nodes, and the terminal node is represented as D.

3. A multi-sensor network resource scheduling method based on multi-radio signal energy harvesting according to claim 2, characterized in that, Among the channel gains described in Step 1, the channel gain between the j-th source node S j and the i-th sensor U i is defined as z i,j , and the channel gain between the i-th sensor U i and the terminal D is defined as z 2,i ; The multi-sensor short packet data packet set described in Step 1 is defined as k = {k i}, i ∈ {1,..., N} where k i represents the amount of data of the short data packet transmitted from sensor i, in bits.

4. A multi-sensor network resource scheduling method based on multi-radio signal energy harvesting according to claim 3, characterized in that In the above Step 2, in the energy transmission stage, M source nodes simultaneously transmit radio frequency signals to N users in a broadcast manner, and the transmission power set of M source nodes is defined as: P = {P j}, j ∈ {1,..., M} where P j represents the transmission power of the source node S j Accordingly, each user U i will receive a set of RF signals, and the power is defined as: where Q i,j (P j ) represents the power obtained from the j-th source node, i.e., S j and is defined as: Q i,j (P j ) = P j z i,j 。 5. A multi-sensor network resource scheduling method based on multi-radio signal energy harvesting according to claim 4, characterized in that, The non-linear energy harvesting model constructed in the above Step 2 is defined as: where P dc,i represents the DC signal power obtained by user U i through energy harvesting, and F eh (Q i ) represents the DC signal power P output by user U i during the energy harvesting process, and the function relationship between the DC signal power P dc,i and the RF signal power Q i is defined as the constant a where represents the load impedance, represents the ideality factor, and v t represents the thermal voltage, W0(·) represents the branch of the Lambert W function, which is the inverse function of f(x) = xe x , and I s represents the reverse bias saturation current, is the sum of monomials with respect to Q i , that is, the positive polynomial of Q, and is defined as: where n0 represents a truncation constant, M represents the number of received RF signals, and the constant where is a constant coefficient, which is defined as I s represents the reverse bias saturation current, is a constant value, represents the ideality factor, and v t represents the thermal voltage; R ant represents the matched antenna impedance, and the expression represents any non - negative sequence the sum of which is equal to the constant is defined as: wherein represents the constant waveform factor of unit power, and the expression reflects the interference between multiple received RF signals represents the m-th value in the non-negative sequence Q m represents the power of the m-th received RF signal; Based on the non-linear energy harvesting model, user U i The harvested energy is defined as: E i (P) = P dc,i n0T s = F eh (Q i )n0T s Among which E i (P) represents U i The energy collected in the first stage, n0 represents the code length of the energy collection in the first stage, in symbols, T s represents the time length of each symbol, in seconds.

6. A multi-sensor network resource scheduling method based on multi-radio signal energy harvesting according to claim 3, characterized in that In Step 3, The multi-user short packet transmission code length set is defined as: n = {n i}, i ∈ {1,..., N} User U i The transmission power is defined as: Among which P dc,i represents the energy harvesting power of user U i , n i represents the code length of the short packet transmission of user U i . Based on this, the transmission signal-to-noise ratio of user U i is defined as: where σ 2 represents the noise power. To ensure reliable communication, the constraint on the signal-to-noise ratio for user short-packet transmission is defined as follows: γ i ≥ 1, i ∈ {1, …, N} That is, the signal-to-noise ratio of all short packet transmissions is not less than 1; User U i The transmission bit error rate is defined as: Among them represents the data transmission rate, with the unit of bit / symbol. Q(·) represents the Gaussian Q function, that is The expression C(γ i ) is defined as C(γ i ) = log(1 + γ i ), which represents the ideal Shannon channel capacity. The expression V(γ i ) is defined as represents the channel scattering. According to this formula, the block error rate ε i of the short packet transmission of user U i is jointly affected by the transmission power P of all source nodes and the code length n i of the short packet transmission of user U i , and is expressed as: Fairly considering the reliable transmission of all users, take the maximum transmission error rate among all users as a measure of system reliability, and its definition is:

7. A multi-sensor network resource scheduling method based on multi-radio signal energy harvesting according to claim 3, characterized in that Construct an optimization problem to minimize the maximum bit error rate of users, and the objective of the optimization problem is defined as: The resource optimization problem is defined as follows: (OP): 0 ≤ P j ≤ P max wherein represents the short - packet transmission signal - to - noise ratio from user U i to the terminal, k i represents the data packet size of user U i and n i represents the short - packet transmission code length of user U i ; P total represents the upper limit of the total transmission power of multiple source nodes in the wireless energy transmission phase; P max represents the maximum transmission power of a single source node during the wireless energy transfer phase; n total Indicates the upper limit of the total code length for short packet transmission by multiple users in a time-division multiplexing manner during the short packet transmission phase.

8. A multi-sensor network resource scheduling method based on multi-radio signal energy harvesting according to claim 7, characterized in that The power optimization problem in step 5 is defined as follows: (SP1): 0 ≤ P j ≤ P max Decouple the direct correlation between the source node transmission power and system reliability, and introduce slack variables That is, take the user's transmission power as the optimization variable, perform variable substitution on the source node transmission power, that is, introduce the variable where Based on the above steps, user U i 's short packet transmission error probability ε i is defined as a function of the user's transmit power: The equation relationship between the original user transmission power and the multi-source nodes, i.e., is transformed into an inequality constraint, defined as: Among them represents user U i The DC current power obtained by energy harvesting. The physical meaning of this constraint is that user U i The total energy consumption of short-packet transmission is not greater than the total energy obtained by energy harvesting in the first stage. Based on this relaxation, sub-problem (SP1) is transformed into (SP2), and (SP2) is defined as follows: (SP2): In problem (SP2), for the first constraint, continuous convex approximation is performed, and the specific steps are as follows: Based on the property Regarding For jointly convex [2], the following inequality can be obtained: where the inequality holds for and the constants are defined as follows, respectively: Based on the above operations, the original non-convex constraints are transformed into convex constraints, that is Meanwhile, problem (SP2) is transformed into multiple convex local problems; the local problem in the τ-th iteration is defined as: (LP1): The optimal solution of the local convex problem can be obtained by the interior point method. By iterating the local problem multiple times, a sub-optimal solution of sub-problem (SP1) is obtained.

9. A multi-sensor network resource scheduling method based on multi-radio signal energy harvesting according to claim 8, characterized in that, In step 6, in the code length allocation problem, the relationship between the transmission error probability and the code length is defined as: where the constant is defined as When the transmission rate r i ≥0.0683, γ i ≥1, i.e., ε i is convex with respect to n i According to the properties of convex functions, we can obtain is jointly convex with respect to {n i}, i ∈ {1,..., N}; Perform continuous convex approximation on the non-convex constraint 2, and the definition is as follows: The equation holds when n = n (χ) where the constants and are defined respectively as: Convert sub-problem (SP1) into multiple convex local problems (LP2), and the local problem in the χ-th iteration is defined as: (LP2): Obtain the optimal solution of the local problem (LP2) by the interior point method. By iterating the local problem, a sub-optimal solution of the code length allocation problem is obtained.

10. A multi-sensor network resource scheduling method based on multi-radio signal energy harvesting according to claim 9, characterized in that, Step 7 includes the following sub-steps: Step 7.1: Determine the number of source nodes M, the number of user nodes N, and determine the channel state information of the source nodes and multiple users {z i,j}, i ∈ {1,..., N}, j ∈ {1,..., M}, and the channel state information between multiple users and the data processing terminal Determine the user service requirements, that is, the packet size k = {k i}, i ∈ {1,..., N}; Step 7.2: Determine the initial feasible solution, that is, initialize the transmission powers of multi-source nodes and the code lengths of multi-user short packets (P (0) , n (0) ), and the number of iterations δ = 0; Step 7.3: Perform power allocation, and initialize the power allocation iteration number τ = 0; Step 7.4: Based on (P (τ) , n (τ) ), determine at the local point compute the constant and perform a convex approximation on the expression ; obtain the local optimal solution by solving (LP1) Step 7.5: If the minimum maximum transmission error rate obtained in this power optimization iteration is less than ε compared to the result of the previous iteration converge , stop the iteration and output the power allocation result Otherwise, τ = τ + 1, go back to Step 7.4; Step 7.6: Initialize the transmission power of multiple source nodes with the power allocation result obtained in step 7.5; initialize the code length allocation iteration number χ = 0; Step 7.7: At the local point n (χ) Calculate the constant and and perform a convex approximation on the expression to obtain the local optimal solution n by solving (LP2) * ; Step 7.8: If the minimum maximum transmission error rate obtained in the current code length optimization iteration is less than ε compared to the result of the previous iteration converge , stop the iteration and output the power allocation result n * ; Otherwise, χ = χ + 1, n (χ) = n * , go back to step 7.7; Step 7.9: In the δ-th iteration of power allocation, if the minimum maximum transmission error rate output in Step 7.8 is less than ε compared to the result of the previous iteration δ-1 converge , stop the power and code length iteration, and output the power and code length allocation results (P * , n * ); otherwise, δ = δ + 1, go back to Step 7.3, and start a new round of power and code length optimization.

Images