A power and time resource allocation method in cellular Internet of Things based on NOMA

Abstract

The present invention relates to a NOMA-based cellular Internet of Things power and time resource allocation method, which can theoretically derive the transmission time t of a given CU under the condition that the selection of the Internet of Things gateway #imgabs0# and the resource sharing relationship between the IoT device and the cellular user #imgabs1# are known. k The optimal transmission power of each IoT device #imgabs2# and the optimal transmission power of the CU that shares resources with the IoT device #imgabs3# are then constructed. An objective function with penalties for all CU transmission times and constraint #imgabs4# violations is constructed, and the differential mutation algorithm is used to calculate the transmission time of each CU #imgabs5#. At the same time, a tail iterative algorithm is designed to adjust the transmission time of each CU #imgabs6# to satisfy #imgabs7#. The present invention can minimize the total energy consumption of all users of the cellular Internet of Things when the IoT gateway selection and resource sharing are known, and has low complexity.

Description

Technical Field

[0001] The present invention relates to the technical field of NOMA-based cellular Internet of Things resource allocation, and in particular to a method for allocating power and time resources in a NOMA-based cellular Internet of Things. Background Art

[0002] With the development of intelligent applications such as home automation, smart transportation, industrial automation, and smart healthcare, the Internet of Things (IoT) has become an integral part of our daily lives, leading to a dramatic increase in mobile data transmission. It is reported that global mobile machine-to-machine (M2M) connections within the IoT paradigm will grow at an average annual rate of 30% starting in 2018, reaching 4.4 billion by 2023. To meet the enormous communication demands of IoT devices, non-orthogonal multiple access (NOMA) has been introduced to improve spectral efficiency within the IoT paradigm. Furthermore, green wireless network design is becoming increasingly important in the information technology (IT) industry, not only to reduce energy costs but also to extend the battery life of mobile devices. Therefore, studying the problem of energy minimization under transmission rate constraints in the IoT paradigm with NOMA is of great importance.

[0003] To minimize energy consumption while maintaining transmission rate constraints in NOMA-based cellular IoT, it is crucial to design a rational power and time resource allocation method. The average time allocation (ATA) method fails to fully account for the differences in transmission rates among users in different NOMA groups, resulting in increased system energy consumption. While the proportional fair allocation method (PFTA) takes into account the transmission rate differences among users in NOMA groups, its time allocation is not rational. Summary of the Invention

[0004] In response to the problems existing in the prior art, the purpose of the present invention is to provide a power and time resource allocation method in a NOMA-based cellular Internet of Things, which can save more system energy consumption and support higher rate requirements.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] A NOMA-based cellular Internet of Things power and time resource allocation method, the energy consumption optimization method comprising the following steps:

[0007] Step 1: IoT Gateway Selection for a Given IoT Device and resource sharing between IoT devices connected to each IoT gateway and CU in, It is the set of IoT gateways in the network. Represents the set of IoT devices in the network, given the penalty factor φ, the maximum algebra M DE, population size Q, mutation factor λ, crossover probability η CR , the normalized time δ of the time slot;

[0008] Step 2: Generate constraints Q individuals Among them, in each individual X q In, x q,k is assigned to CUc k time, i.e.

[0009] Step 3: Calculate each individual X q The power of the corresponding IoT device and CU power Calculate the energy consumption corresponding to all individuals LP DE =1;

[0010] Step 4: For each individual X q (1≤q≤Q), and then recalculate the total energy consumption of the system E(X q );

[0011] Step 5: LP DE =LP DE +1, if the current iteration number LP DE >M DE , go to step 6; otherwise, go to step 4;

[0012] Step 6: Select the individual q corresponding to the minimum energy consumption * ,Right now

[0013] Step 7: If individual q * The elements satisfy Then go to step 8; otherwise go to step 13;

[0014] Step 8 in is an upward value function, let LP TA =1;

[0015] Step 9: For each CU Calculate CUc k Time resources from becomes Energy consumption increase In CUc k When the reduction of time resources makes the transmission power infeasible, Set to +∞;

[0016] Step 10: If the energy consumption increment of all CUs is +∞, that is, If there is no solution to the current time resource optimization problem, exit; otherwise, go to step 11;

[0017] Step 11: Find the CU with the smallest energy consumption increment and reduce the transmission time of the CU.

[0018] Step 12: The number of iterations increases, i.e. LP TA =LP TA +1; if LP TA >N, go to step 13; otherwise, go to step 9;

[0019] Step 13: Output the optimal individual and the optimal individual The power of the corresponding IoT device CU power Energy consumption corresponding to the optimal individual

[0020] The step 4 is specifically as follows: for each individual X q (1≤q≤Q), perform the following substeps:

[0021] Step 4-1, mutation: Randomly extract three different individuals numbered q1, q2 and q3 from the set {1,2,...,Q}\{q}, and pass Generate a new individual

[0022] Step 4-2, cross: Randomly from Take out a serial number k1 and generate a random number η in the range of [0,1], and generate an individual by the following formula

[0023]

[0024] Calculate U q The energy consumption E(U q ), if U q If there are elements infeasible, then E(U q )=+∞;

[0025] Step 4-3, Natural Selection: If E(U q )<E(X q ), X q Replaced by U q Otherwise, X q remain unchanged;

[0026] Step 4-4, recalculate each individual X q The power of the current corresponding IoT device CU power Calculate the total energy consumption of the system E(X q ).

[0027] After adopting the above solution, the present invention can select the gateway of the known Internet of Things and resource sharing relationship between IoT devices and cellular users In the case of a given CU, the transmission time t k The optimal transmit power for each IoT device and the optimal transmit power of the CU that shares resources with IoT devices Then, construct a The objective function of the penalty for violation is to calculate the transmission time of each CU using the differential mutation algorithm At the same time, a tail iterative algorithm is designed to calculate the transmission time of each CU Adjust to meet This invention minimizes the total energy consumption of all cellular IoT users with low complexity, given known IoT gateway selection and resource sharing. In summary, this invention optimizes the transmit power of IoT devices and CUs, and rationally allocates time resources, saving system energy. Compared with common average time allocation algorithms and proportional time allocation, this invention can significantly reduce system energy consumption and support higher rate requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 For the present invention (PTADE) with the maximum algebra M DE Energy consumption convergence curve;

[0029] Figure 2 is the impact of the penalty factor φ on energy consumption performance in the present invention (PTADE);

[0030] Figure 3 The performance of the present invention (PTADE) is compared with that of the ATA method and the PFTA method under different IoT device rate requirements.

[0031] Figure 4 Performance comparison of the invention (PTADE), the ATA method, and the PFTA method under different CU rate requirements. DETAILED DESCRIPTION

[0032] The present invention discloses a NOMA-based cellular Internet of Things system energy consumption optimization method, which uses an uplink cellular Internet of Things, which includes a group of IoT devices A set of IoT gateways A group of CU The IoT gateway acts as a data aggregator for IoT devices, which are connected to the base station via wired transmission. Each IoT device selects an IoT gateway to access. IoT devices are divided into a series of IoT device groups based on the IoT gateway they are connected to. Each IoT device group transmits data to its IoT gateway connected to NOMA by sharing the uplink resources of a CU. The IoT gateway decodes the data from the IoT devices by applying SIC technology.

[0033] Considering the case where each IoT device has a rate requirement, to reduce the impact of IoT device transmission on CU transmission and simplify theoretical analysis, we assume that each CU's uplink resources can be shared by at most a group of IoT devices connected to a single IoT gateway. Using centralized resource allocation, the base station (BS) is responsible for allocating resources and directing transmissions for all devices. In LTE cellular systems, time is divided into fixed-length frames, and the resources within each frame are further divided into resource blocks in the time and frequency domains. This allows CUs to set different transmit powers for different frequency subchannels, thereby counteracting frequency-selective fading. A frequency-flat fading channel is considered. Therefore, it is assumed that each CU occupies all frequency subchannels with the same transmit power when transmitting. In other words, the BS only needs to allocate time resources to the CU and its resource-sharing IoT devices based on the rate requirements.

[0034] The time of each time frame is normalized to 1. Let t k Indicates allocation to CUc k For ease of analysis, the continuous variable t k To approximate the amount of time allocated to c in each frame k For example, if there are 10 time slots per frame, “t k =0.2" means CUc k occupies two time slots in each frame. Obviously, this approximation is more accurate when the number of time slots in each frame is larger. Therefore, we have the following transmission time and constraints for all CUs:

[0035]

[0036] Total energy consumption E of users in the system total can be calculated as

[0037]

[0038] in, Indicates connection to IoT gateway h j A collection of IoT devices, here s' j,l yes The lth decoded IoT device, y j,k Respectively represent resource sharing, yj,k =1 means connecting to IoT gateway h j IoT devices share CUc k The resources are transferred, otherwise y j,k =0; and IoT devices' j,l and CUc k The transmission power, and θ respectively IoT devices s' j,l and CUc k The power consumption of the circuit blocks other than the power amplifier, IoT devices' j,l and CUc k The idle energy consumption and the drain efficiency of the power amplifier.

[0039] When choosing an IoT gateway for a given IoT device and resource sharing between IoT devices connected to each IoT gateway and CU j,k (h j Laos ), it can be proved that when each group is decoded in descending order of channel gain When the signal of IoT device is in the middle, the power of IoT device is the minimum, that is, Minimum power consumption for IoT devices It can be calculated by the following formula:

[0040]

[0041]

[0042] Among them, R j,l , g(c k ,h j ), W and n0 represent IoT devices s' j,l The rate requirement of CUc k To IoT Gateway j Channel gain, IoT devices' j,l To IoT Gateway j The channel gain, channel bandwidth, and Gaussian white noise power spectral density.

[0043] When choosing an IoT gateway for a given IoT device and resource sharing between IoT devices connected to each IoT gateway and CU At a given transmission time In the case of, based on formula (2) (3), each CUc k The optimal power can be calculated optimally through the following two cases.

[0044] Case 1: When CUc k Share resources with a group of IoT devices. Then, CUc k The optimal power It can be calculated theoretically as

[0045]

[0046] Among them, j(c k ) indicates that the k The serial number of the CR router connected to the resource-sharing IoT group, R k CUc k The rate requirement, g(c k ,BS) represents CUc k Channel gain to base station BS.

[0047] Case 2: When CUc k There is no sharing of resources with any group of IoT devices. k The optimal power It can be calculated theoretically as

[0048]

[0049] Obviously, when choosing an IoT gateway for a given IoT device and resource sharing between IoT devices connected to each IoT gateway and CU At a given transmission time In the case of , we can use formula (2) to calculate the system energy consumption according to (3)(4)(5)(6). However, the allocation of time resources has not been involved. Considering that the allocation of different time resources has different effects on the energy consumption of the entire system, it is necessary to allocate time resources. Make reasonable allocation. Next, DE algorithm is used to solve

[0050] Considering that the DE algorithm is difficult to meet the CU transmission time and constraints Therefore, in the DE algorithm, we relax this constraint to Then a constraint The optimization objective function of violation penalty, that is, the time resource allocation problem is transformed into:

[0051]

[0052] Since the DE algorithm's running results may still lead to We designed a tail iterative processing algorithm to adjust Make it satisfy the constraints Finally achieve optimization goal.

[0053] The power and time resource allocation method of the present invention specifically includes the following steps:

[0054] Step 1: IoT Gateway Selection for a Given IoT Device and resource sharing between IoT devices connected to each IoT gateway and CU Given a penalty factor φ, the maximum number of generations M DE , population size Q, mutation factor λ, crossover probability η CR , the normalized time δ of the time slot;

[0055] Step 2: Generate constraints Q individuals Among them, in each individual X q In, x q,k is assigned to CUc k time, i.e.

[0056] Step 3: Calculate each individual X according to formula (3)(4)(5)(6) q The power of the corresponding IoT device and CU power The energy consumption corresponding to all individuals is calculated using the objective function of formula (7): LP DE =1;

[0057] Step 4: For each individual X q (1≤q≤Q), perform the following substeps:

[0058] Step 4-1, mutation: Randomly extract three different individuals numbered q1, q2 and q3 from the set {1,2,...,Q}\{q}, and pass Generate a new individual

[0059] Step 4-2, cross: Randomly from Take out a serial number k1 and generate a random number η in the range of [0,1], and generate an individual by the following formula (8):

[0060]

[0061] Calculate U using formula (3)(4)(5)(6) q The energy consumption E(U q ), if U q If there are elements infeasible, then E(U q )=+∞;

[0062] Step 4-3, Natural Selection: If E(U q )<E(X q ), X q Replaced by U q Otherwise, X q Remain unchanged.

[0063] Step 4-4: Reuse equations (3)(4)(5)(6) to calculate each individual X q The power of the current corresponding IoT device CU power And use the objective function of (7) to calculate the total energy consumption of the system E(X q );

[0064] Step 5: LP DE =LP DE +1, if the current iteration number LP DE >M DE , go to step 6; otherwise, go to step 4;

[0065] Step 6: Select the individual q corresponding to the minimum energy consumption * ,Right now

[0066] Step 7: If individual q * The elements satisfy Then go to step 8; otherwise go to step 13;

[0067] Step 8 in is an upward value function, let LP TA =1;

[0068] Step 9: For each CU Calculate CUc k Time resources from becomes Energy consumption increase In CUc k When the reduction of time resources makes the transmission power infeasible, Set to +∞;

[0069] Step 10: If the energy consumption increment of all CUs is +∞, that is, If there is no solution to the current time resource optimization problem, exit; otherwise, go to step 11;

[0070] Step 11: Find the CU with the smallest energy consumption increment and reduce the transmission time of the CU.

[0071] Step 12: The number of iterations increases, i.e. LPTA =LP TA +1; if LP TA >N, go to step 13; otherwise, go to step 9;

[0072] Step 13: Output the optimal individual and the optimal individual The power of the corresponding IoT device CU power Energy consumption corresponding to the optimal individual

[0073] In order to evaluate the performance of the present invention, the following simulations were performed, and the simulation parameters were set as follows: the cell radius was set to 500m. The BS was located in the center, and the IoT devices, CUs, and IoT gateways were evenly distributed. The path loss index was set to 4. The system bandwidth and noise power density were set to 1MHz and -174dBm / Hz. The maximum transmit power of the CU and IoT devices was set to 33dBm and 30dBm, respectively. The drain efficiency of the power amplifier, the circuit power consumption, and the idle power consumption of the CU were set to 0.2, 106.4mW, and 25mW, respectively. The population size Q, the variation factor λ, and the crossover probability η in the PTADE algorithm are CR and the normalized time δ of one time slot are set to 20, 0.6, 0.3, and 0.01, respectively. The maximum allowed number L of IoT devices connected to one IoT gateway is set to 10.

[0074] Figure 1 and Figure 2 Shows the maximum number of algebras M DE The performance impact of the penalty factor φ on the PTADE algorithm. For each point of the curve, we generate 50 random cases and calculate the average energy consumption. In each case, the number of IoT devices, the number of CUs, the number of IoT gateways, the rate requirements of each CU and each IoT device are randomly selected in the range of [20, 34], [8, 17], [2, 8], [20knats / s, 240knats / s] and [10knats / s, 160knats / s], and random IoT gateways and resource sharing selections are generated. Figure 1 and Figure 2 We have two observations. One is that the system energy consumption increases with M DE The value of M decreases rapidly, but when DE When M is greater than 80, the energy consumption does not change much. Another reason is that the penalty factor φ has little effect on the system energy consumption. Therefore, in the subsequent simulation, M DE and φ are set to 100 and 50 respectively.

[0075] Figure 3 and Figure 4The performance of the proposed PTADE algorithm is compared with the average time allocation (ATA) scheme and the proportional fair time allocation (PFTA) scheme under randomly generated IoT gateways and resource sharing selection. The simulation considers a network with 30 IoT devices, 12 CUs and 5 IoT gateways. In the ATA scheme, each CU occupies the same time resources; while in the PFTA scheme, time resources are allocated to CUs proportionally according to the sum of the rate requirements of each CU and its resource-sharing IoT devices. Figure 3 It can be seen that compared with the ATA and PFTA schemes, the proposed PTADE algorithm can greatly save system energy consumption and support higher rate requirements. Figure 3 The speed requirement of IoT devices is greater than 74knat / s, and Figure 4 When the rate requirements of the CU exceed 70 kbps, the ATA solution becomes unfeasible. The reasons are as follows. The time resources required by each CU and its shared IoT devices are determined by multiple factors, such as the rate requirements of each device, the channel gain of each transmission, and interference between devices. With the ATA solution, time resource allocation is fixed, making it easily infeasible when the device rate requirements are large. The PFTA solution has limited performance because it adjusts time resources only based on the rate requirements.

[0076] The above description is merely an embodiment of the present invention and does not limit the technical scope of the present invention. Therefore, any minor modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A NOMA-based cellular Internet of Things power and time resource allocation method, characterized by: The energy consumption optimization method includes the following steps: Step 1: IoT Gateway Selection for a Given IoT Device and resource sharing between IoT devices connected to each IoT gateway and CU ,in, , It is the set of IoT gateways in the network. Represents the set of IoT devices in the network, given a penalty factor , the maximum algebra , population size , the variation factor , crossover probability , the normalized time of the time slot ; Step 2: Generate constraints of individual , , where in each individual middle, is assigned to CU time, i.e. ; Step 3: Calculate each individual The power of the corresponding IoT device and CU power , calculate the energy consumption corresponding to all individuals ,make ; Step 4: For each individual , recalculate the total energy consumption of the system after mutation, crossover, and natural selection ,in, ; Step 5 , if the current iteration number , go to step 6; otherwise, go to step 4; Step 6: Select the individual with the minimum energy consumption ,Right now ; Step 7: If the individual The elements satisfy , then go to step 8; otherwise go to step 13; Step 8 , ,in Let be the upward value function, ; Step 9: For each CU , calculate CU Time resources from becomes Energy consumption increase , in CU When the reduction of time resources makes the transmission power infeasible, Set to ; Step 10: If the energy consumption increment of all CUs is ,Right now , then the current time resource optimization problem has no solution and exit; otherwise, go to step 11; Step 11: Find the CU with the smallest energy consumption increment and reduce the transmission time of the CU. , ; Step 12: The number of iterations increases, i.e. ;like , then go to step 13; otherwise, go to step 9; Step 13: Output the optimal individual , and the optimal individual The power of the corresponding IoT device , CU power Energy consumption corresponding to the optimal individual ,in, .

2. The NOMA-based cellular Internet of Things power and time resource allocation method according to claim 1, characterized in that: The step 4 is specifically as follows: for each individual , perform the following substeps: Step 4-1, mutation: Randomly from the set Three different individual numbers are drawn from , and ,pass Generate a new individual ; Step 4-2, cross: Randomly from Take out a serial number , and generates a Random number in range , an individual is generated by the following formula calculate Energy consumption corresponding to the individual ,if If an element in is not feasible, then ; Step 4-3, Natural Selection: If ,Will Replace with ; otherwise, remain unchanged; Step 4-4, recalculate each individual The power of the current corresponding IoT device , CU power , calculate the total energy consumption of the system .

Images