Resource allocation method and system for drone-assisted 6G-supported intelligent transportation system

Abstract

The present invention relates to the technical field of Internet of Things, and specifically discloses a resource allocation method and system for a drone-assisted 6G-supported intelligent transportation system. Aiming at the overlapping interference and energy consumption problems in the drone-assisted 6G-ITS network, the types and relationships of interference in the network are analyzed, an interference hypergraph model is established, and a method for reducing overlapping interference is designed based on the interference hypergraph model to improve spectrum utilization efficiency. Then, based on power constraints, interference constraints, EH constraints, and signal-to-interference ratio constraints, an EH optimization model under imperfect CSI is established. Finally, the IT-EHRA algorithm is used to obtain the power allocation under the maximum EE of the network. The simulation results show that the method and system can effectively improve energy utilization, extend the working time of drones, and improve network throughput, providing a feasible solution for resource allocation of drone-assisted 6G-ITS networks.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and in particular to a resource allocation method and system for a drone-assisted 6G-supported intelligent transportation system. Background Art

[0002] UAV-Assisted 6G-supported Intelligent Transportation System (UAV-Assisted 6G-ITS) has become an important research topic in intelligent transportation. It can provide huge connectivity and interaction for various connected and autonomous vehicles for different applications, and realize ubiquitous data processing and sharing. However, situations such as low signal coverage or damage to ground communication equipment may prevent networked autonomous vehicles from achieving reliable communication, which poses a challenge to large-scale ITS (Intelligent Transportation System). UAV-Assisted 6G-ITS (6G-supported Intelligent Transportation System) is a feasible solution. Due to its easy deployment, high mobility and low operating cost, drones can be deployed as repeaters to collect data from vehicles and provide network connectivity for moving vehicles.

[0003] However, despite its great potential, drone-assisted 6G-ITS also faces some notable challenges and limitations. First, the battery life of drones limits the flight time, which affects the usage time of drones, the sustained combat capability of the network, and the energy efficiency of the network. Second, drone flights are limited by weather and environmental conditions such as strong winds, bad weather, and complex terrain. This may cause signal interference, multipath fading, and other effects in urban environments, thus affecting the resource allocation efficiency of drone-assisted 6G-ITS systems. Therefore, achieving efficient resource allocation for drone-assisted 6GITS systems is a new challenge.

[0004] Some researchers have achieved high energy efficiency (EE) of the network by extending the life of drones. Energy harvesting (EH) technology is considered to be a promising technology to reduce energy consumption by converting energy in the environment into usable energy. Therefore, some scholars have applied EH technology to drones. Some scholars have studied the application of nonlinear EH technology in drone-assisted full-duplex networks, deeply considering key factors such as shadow scenarios, energy harvesting receiver architecture, device mobility, interference and self-interference elimination, and improving the reliability of the system. Some scholars have studied the performance of wireless energy transmission networks for multiple ground sensor nodes of drones, and maximized the performance of the network by jointly designing the trajectory and directional antenna direction of the drone. Some scholars have studied the energy efficiency of IoT distributed antenna systems using synchronous wireless information and power transmission technology under fading channels, and proposed an optimal allocation scheme to achieve maximum energy utilization by jointly adjusting power allocation (PA) and power splitting (PS). In response to two practical problems in cellular IoT, namely incomplete continuous interference cancellation and nonlinear EH, some scholars have studied a large-scale non-orthogonal multiple access (NOMA) scheme (SWIPT) for simultaneous transmission of wireless information and power transmission to alleviate the impact of incomplete continuous interference cancellation and nonlinear EH. Some scholars have studied the energy minimization problem of multiple energy harvesting devices in the UAV mobile edge computing system, reducing the energy consumption of the UAV and improving the system throughput. The above research uses EH technology to reduce or eliminate the dependence of UAVs on traditional batteries, enabling UAVs to extend their flight time and improve their aerial work efficiency, thereby improving the energy efficiency of the network.

[0005] Other scholars have also optimized the network structure to improve the EE of the network and achieve sustainable development of the network. Some scholars have proposed a drone-based intelligent offloading solution for computing and communication to solve the resource limitations of interconnected devices in intelligent transportation systems under 6G networks, achieve efficient and energy-saving offloading tasks, and minimize the overall energy consumption of the interaction between IoT devices and connected vehicles. Some scholars have considered the performance analysis and energy-saving resource allocation optimization of large-scale IoT networks based on large-scale multi-input multi-output decoding and forwarding relays, effectively improving the EE of the network. Some scholars have explored the age of network information problems. They use the information age as a configuration technology for network service quality, aiming to minimize the average peak age of information and improve the EE of the network.

[0006] The aforementioned research has made considerable progress in the EH of drones. However, with the surge in connected and autonomous vehicles in 6G-ITS, the number of data interactions will inevitably grow exponentially, which may cause serious overlapping interference, affect communications between vehicles, and challenge the reliability and communication of the network. On the one hand, when a large number of vehicles communicate on the same frequency, the signals between vehicles may overlap, resulting in communication errors and increased energy consumption. On the other hand, in the 6G-ITS environment, structures such as buildings may cause signal reflections and multipath propagation, resulting in increased data transmission delays and reduced communication quality, thereby reducing the energy efficiency of the network. At present, there is a lack of an effective solution to reduce interference in drone-assisted 6G-ITS while improving the EE of the network. Summary of the invention

[0007] The present invention provides a resource allocation method and system for a drone-assisted 6G-supported intelligent transportation system, and solves the technical problem of how to improve the EE of the network while reducing interference in the drone-assisted 6G-ITS.

[0008] In order to solve the above technical problems, the present invention provides a resource allocation method for a drone-assisted 6G-supported intelligent transportation system, comprising the steps of:

[0009] S1. Determine the network structure of drone-assisted 6G supporting intelligent transportation system;

[0010] S2. Determine the channel model of the drone-assisted 6G-supported intelligent transportation system based on the network structure of the drone-assisted 6G-supported intelligent transportation system;

[0011] S3, determining the interference type of the drone-assisted 6G-supported intelligent transportation system based on the network structure of the drone-assisted 6G-supported intelligent transportation system;

[0012] S4. Based on the hypergraph theory, a hypergraph interference model of network structure and interference type based on UAV-assisted 6G-supported intelligent transportation system is constructed;

[0013] S5. Determine the energy efficiency calculation method of the drone-assisted 6G-supported intelligent transportation system based on the network structure and channel model of the drone-assisted 6G-supported intelligent transportation system;

[0014] S6. With the goal of maximizing the energy efficiency of the drone-assisted 6G-supported intelligent transportation system, with the constraint that the interference value calculated based on the hypergraph interference model does not exceed the interference threshold, and with the constraint that the signal interference plus noise ratio requirement, energy efficiency requirement, and power requirement of the drone-assisted 6G-supported intelligent transportation system are met, an optimization problem for optimizing the transmitter transmission power in the drone-assisted 6G-supported intelligent transportation system is constructed;

[0015] S7, determining channel gain data of the UAV-assisted 6G-supported intelligent transportation system based on channel learning;

[0016] S8. Solve the optimization problem based on the channel gain data to obtain the transmission power of the transmitter in the UAV-assisted 6G-supported intelligent transportation system.

[0017] Further, in the step S1, the drone-assisted 6G supported intelligent transportation system includes U intelligent vehicles, namely CAV, and B base stations, namely BS, the U intelligent vehicles are divided into M authorized vehicles, namely AV, and N unauthorized vehicles, namely UV; the drone-assisted 6G supported intelligent transportation system is also equipped with N drones, namely UAV, a UV and a drone form a D2D, namely a device-to-device pair, and the drone transmits the uplink spectrum to the UV by multiplexing the AV; the bandwidth of the drone-assisted 6G supported intelligent transportation system is W Hz, which is evenly divided into L resource blocks, namely RBs, and each resource block has W / L Hz; the drone-assisted 6G supported intelligent transportation system has a total of J sub-channels.

[0018] Furthermore, in step S2, the channel model of the drone-assisted 6G-supported intelligent transportation system is constructed as follows:

[0019] The BS uses NOMA to provide services to the CAV, and the UAV uses SWIPT technology to divide the received signal into two parts: information decoding and energy collection;

[0020] The received signal x of the mth AV, AV m, and the nth D2D pair, D2D pair n m and x n Respectively expressed as:

[0021]

[0022] Among them, p m,j and p n,j denote the transmission power of AV m and D2D pair n on the jth subchannel, i.e., subchannel j, respectively. and denote the channel gains of AV m and D2D pair n on subchannel j, n m 、n n are the additive white Gaussian noise on AV m and D2D pair n, respectively, m=1,2,…,M, n=1,2,…,N, j=1,2,…,J;

[0023] The signal to interference plus noise ratio or SINR of AV m is given by:

[0024]

[0025] in, is the noise power of AV m, i corresponds to AV i, i ≠ m;

[0026] Each drone will receive data x n Divided into two parts:

[0027]

[0028]

[0029] Where ω is the power partition (PS) factor, σ I represents the noise generated by the PS method;

[0030] The SINR of D2D pair n is expressed as:

[0031]

[0032] x corresponds to the D2D pair x, x≠n.

[0033] Further, in the step S3, the D2D pair and the AV are uniformly represented as DV, and DV is used to represent any one of the D2D pair and the AV, then the interference types of the drone-assisted 6G supporting intelligent transportation system include cumulative interference and independent interference;

[0034] Cumulative interference means that when multiple DVs are running simultaneously, the interference generated by other DVs exceeds the interference threshold of a single DV, and cumulative interference is generated between the multiple DVs running simultaneously;

[0035] Independent interference means that in a scenario where two DVs are operating, when the interference generated by one DV exceeds the interference threshold of the other DV, independent interference will occur.

[0036] Further, in the step S4, a hypergraph interference model is constructed with the DVs as vertices and the interference relationships between the DVs as hyperedges;

[0037] The relationship between vertices in the hypergraph interference model is represented by the adjacency matrix A, and the values ​​of the elements in the kth row and lth column are as follows:

[0038]

[0039] Among them, v k represents the kth vertex, v l represents the lth hyperedge;

[0040] The product of the adjacency matrix A and the power matrix P must be less than or equal to the interference threshold matrix I, and the interference threshold matrix is ​​set to the bandwidth W / L of each RB.

[0041] Furthermore, in step S5, the energy efficiency calculation method of the drone-assisted 6G-supported intelligent transportation system is:

[0042]

[0043] R SUM represents the total rate of users, P SUM Indicates the total power consumption of the system;

[0044] R SUM Calculated by the following formula:

[0045]

[0046] Among them, γ n,j represents the signal-to-noise ratio of UV n on the jth subchannel;

[0047] P SUM Calculated by the following formula:

[0048]

[0049] Where α represents the power amplifier efficiency factor, P x Indicates the constant power consumption of the circuit module.

[0050] Furthermore, in step S6, the optimization problem is constructed as:

[0051] max:η EE

[0052] st

[0053]

[0054]

[0055] C3:E≥E min

[0056] C4:AP≤I

[0057]

[0058]

[0059] Among them, max means maximization, st means need to be satisfied, C1 to C6 represent different constraints. Indicates the minimum AV SINR threshold, represents the minimum SINR threshold of D2D, E represents the collection energy on the drone, and E min Indicates the minimum energy collection threshold of the drone. and Respectively represent the maximum threshold of the transmission power of AV and UV;

[0060] E is calculated by the following formula:

[0061]

[0062] Where a and b correspond to parameters related to the specification of the energy harvesting circuit, P max It indicates the maximum energy harvesting value of the drone when the energy harvesting circuit is saturated. exp() refers to the exponential function with the natural constant e as the base. P in Indicates the input power of the D2D pair.

[0063] Furthermore, the step S7 specifically includes the steps of:

[0064] S71. Collect N UV channel data sets: The channel data set of M AVs is

[0065] S72. Model the channel gain as a polyhedron model:

[0066]

[0067]

[0068] in, represents the channel gain modeling of AV, Represents the channel gain modeling of UV, represents the channel gain set of the mth AV on the jth subchannel, represents the channel gain set of the nth UV on the jth subchannel, and Represents auxiliary variables;

[0069] S73, using the sample mean of the channel data set to respectively determine and The value of

[0070] S74. Set the reliability to a calibration function of 1-θ to adjust the sample set:

[0071]

[0072] express The prior distribution of express Prior distribution of , Pr{} represents the probability of the event in brackets;

[0073] S75. Calculate Z and X:

[0074]

[0075] f(D * ) indicates that D * The solution of the calibration function obtained, f(H * ) indicates that H * The solution to the calibration function is obtained, and are the upper bounds of the 1-θ quantiles of f(D) and f(H), respectively, where f(D) represents the sample set The calibration function, f(H) represents the sample set The calibration function of

[0076] S76, the result calculated in step S73 is and Substitute Z and X calculated in step S75 into the polyhedron model constructed in step S72 to obtain the learned channel gain data.

[0077] Furthermore, the step S8 specifically includes the steps of:

[0078] S81, converting the optimization model into:

[0079] max:

[0080] stC1-C6

[0081] in, Represents auxiliary variables;

[0082] S82, transform the model shown in step S81 into a Lagrangian dual problem:

[0083]

[0084] st

[0085]

[0086] in, is the Lagrangian function, and is a non-negative Lagrange multiplier, Intermediate variables

[0087] S83. Based on the KKT condition of convex optimization theory, find F(x) at p n,j and p m,j The partial derivative at ;

[0088] S84, let F(x) be n,j and p m,jThe partial derivative at is 0, and the transmission power p is obtained by solving n,j and p m,j The optimal solution and

[0089] The present invention also provides a resource allocation system for a drone-assisted 6G-supported intelligent transportation system, the key of which is that an intelligent agent is provided, and the intelligent agent is used to implement the above-mentioned resource allocation method.

[0090] The resource allocation method and system for the drone-assisted 6G-supported intelligent transportation system provided by the present invention analyzes the types and relationships of interference in the drone-assisted 6G-ITS network, establishes an interference hypergraph model, and designs a method for reducing overlapping interference based on the interference hypergraph model to improve spectrum utilization efficiency. Then, based on power constraints, interference constraints, EH constraints, and signal-to-interference ratio constraints, an EH optimization model under imperfect CSI is established. Finally, the IT-EHRA algorithm is used to obtain the power allocation under the maximum EE of the network. The simulation results show that the method and system can effectively improve energy utilization, extend the working time of the drone, and improve network throughput, providing a feasible solution for resource allocation of the drone-assisted 6G-ITS network. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 is a network structure diagram of a drone-assisted 6G-supported intelligent transportation system provided by an embodiment of the present invention;

[0092] Figure 2 is an example diagram of an interference hypergraph provided by an embodiment of the present invention;

[0093] Figure 3 is an EH comparison diagram of a linear EH model and a nonlinear EH model provided by an embodiment of the present invention;

[0094] Figure 4 is the minimum harvest energy threshold E of the four algorithms provided in the embodiment of the present invention. min Relationship diagram with the total EE of the network;

[0095] Figure 5 is the minimum harvest energy threshold E of the four algorithms provided in the embodiment of the present invention. min Relationship graph with total network throughput;

[0096] Figure 6 The power thresholds of the four algorithms provided in the embodiments of the present invention are Relationship diagram with drone-assisted 6G-ITS network performance;

[0097] Figure 7is the minimum SINR threshold of the D2D pair of the four algorithms provided in the embodiment of the present invention Relationship diagram with the total EE of the network;

[0098] Figure 8 is the minimum SINR threshold of the D2D pair of the four algorithms provided in the embodiment of the present invention Graphed against total network throughput. DETAILED DESCRIPTION

[0099] The following specifically illustrates the implementation mode of the present invention in conjunction with the accompanying drawings. The embodiments are provided for illustrative purposes only and cannot be understood as limiting the present invention. The accompanying drawings are provided for reference and illustration only and do not constitute a limitation on the scope of patent protection of the present invention, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0100] The method for allocating resources of a drone-assisted 6G-supported intelligent transportation system provided by an embodiment of the present invention is as follows: Figure 1 As shown, the steps include:

[0101] S1. Determine the network structure of drone-assisted 6G supporting intelligent transportation system;

[0102] S2. Determine the channel model of the drone-assisted 6G-supported intelligent transportation system based on the network structure of the drone-assisted 6G-supported intelligent transportation system;

[0103] S3, determining the interference type of the drone-assisted 6G-supported intelligent transportation system based on the network structure of the drone-assisted 6G-supported intelligent transportation system;

[0104] S4. Based on the hypergraph theory, a hypergraph interference model of network structure and interference type based on UAV-assisted 6G-supported intelligent transportation system is constructed;

[0105] S5. Determine the energy efficiency calculation method of the drone-assisted 6G-supported intelligent transportation system based on the network structure and channel model of the drone-assisted 6G-supported intelligent transportation system;

[0106] S6. With the goal of maximizing the energy efficiency of the drone-assisted 6G-supported intelligent transportation system, with the constraint that the interference value calculated based on the hypergraph interference model does not exceed the interference threshold, and with the constraint that the signal interference plus noise ratio requirement, energy efficiency requirement, and power requirement of the drone-assisted 6G-supported intelligent transportation system are met, an optimization problem for optimizing the transmitter transmission power in the drone-assisted 6G-supported intelligent transportation system is constructed;

[0107] S7, determining channel gain data of the UAV-assisted 6G-supported intelligent transportation system based on channel learning;

[0108] S8. Solve the optimization problem based on the channel gain data to obtain the transmission power of the transmitter in the UAV-assisted 6G-supported intelligent transportation system.

[0109] For ease of description, the drone-assisted 6G-supported intelligent transportation system is referred to as the system.

[0110] (1) Step S1: System network structure

[0111] The network structure of the drone-assisted 6G-supported intelligent transportation system provided in this example is as follows Figure 1 As shown, it includes U intelligent vehicles (CAV) and B base stations (BS). U intelligent vehicles (CAV) are divided into M authorized vehicles (AV) and N unauthorized vehicles (UV). The system also has N drones UAV. A UV and a drone form a D2D (device to device) pair. The drone transmits the uplink spectrum to the UV by multiplexing the AV. During the transmission process, the drone collects and stores RF energy to achieve sustainable development. In the drone-assisted 6G-ITS network, this example assumes that the entire system bandwidth is W Hz, which is evenly divided into L resource blocks (RBs), and each resource block has (W / L) Hz. The system has a total of J subchannels.

[0112] (2) Step S1: System Channel Model

[0113] The specific communication scenario is as follows: the BS uses non-orthogonal multiple access (NOMA) to provide services for CAVs, and the UAV uses wireless communication with energy (SWIPT) technology to divide the received signal into two parts: information decoding (ID) and EH (energy harvesting). The received signal x of the mth AV (AV m) and the nth D2D pair (D2D pair n) m and x n Respectively expressed as:

[0114]

[0115] Among them, p m,j and p n,j denote the transmit power of AV m and D2D pair n on the jth subchannel (subchannel j), respectively. and denote the channel gains of AV m and D2D pair n on subchannel j, respectively, and are the additive white Gaussian noise on AV m and D2D pair n, respectively, m=1,2,…,M, n=1,2,…,N, j=1,2,…,J; represents additive white Gaussian noise, denote the covariance matrices of the noise of AV m and D2D pair n respectively.

[0116] Therefore, the signal-to-interference-plus-noise ratio (SINR) of AV m is given by:

[0117]

[0118] in, is the background noise power of AV m, i corresponds to AV i, i≠m.

[0119] Since each drone is equipped with SWIPT technology, the received data n It is divided into two parts, as follows:

[0120]

[0121]

[0122] Where ω is the PS factor, represents the noise generated by the power partition (PS) method, represents additive white Gaussian noise. Therefore, the SINR of D2D pair n is expressed as:

[0123]

[0124] x corresponds to the D2D pair x, x≠n.

[0125] (3) Step S3: System interference type

[0126] In drone-assisted 6G-ITS, this example considers three scenarios of mutual interference, including interference between D2D pairs, interference between D2D pairs and AV, and interference between AV and AV. In order to simplify the representation of interference between D2D pairs and AV, this example uniformly represents D2D pairs and AV as DV, and uses DV to represent any one of the D2D pairs and AV, thereby converting the three interferences into interference between DVs. In order to alleviate the interference problem in drone-assisted 6G-ITS, it is essential to understand the interference sources. Therefore, this example divides interference into two types: cumulative interference and independent interference.

[0127] Cumulative interference: When multiple DVs are running simultaneously, the interference generated by other DVs exceeds the interference threshold of a single DV, and cumulative interference is generated between the multiple DVs running simultaneously.

[0128] Independent interference: In a scenario where two DVs are operating, independent interference occurs when the interference generated by one DV exceeds the interference threshold of the other DV.

[0129] (4) Step S4: Hypergraph Interference Model

[0130] In drone-assisted 6G-ITS networks, there are numerous communication devices and interference sources, resulting in a large amount of cumulative interference. In traditional graph theory methods, one-to-one connection relationships cannot well describe such interference between DVs. In contrast, hyperedges in hypergraphs can establish connections between two or more vertices, providing a more suitable framework for comprehensively addressing the complex interference challenges inherent in drone-assisted 6G-ITS networks.

[0131] The hypergraph is represented by G = {V, E}, V = {ν1, v2, ...} represents the vertex set, and E = {e1, e2, ...} represents the hyperedge set. Based on the definition of the hypergraph, an interference hypergraph is constructed, where the vertices in the hypergraph represent DVs, and the hyperedges represent the interference relationships between DVs.

[0132] Figure 2 An example of an interference hypergraph is shown in FIG, where five different colors correspond to five hyperedges (E1, E2, ..., E5), including a total of eight DVs represented as (DV1, DV2, ..., DV8). When a DV is linked to multiple hyperedges at the same time, it indicates that it exists in the overlapping area of ​​these hyperedges, causing interference with the corresponding DVs therein.

[0133] Hyperedge construction rules: Hyperedges are formed based on independent and cumulative interference. Figure 2 For example, when the interference caused by DV1 and DV8 exceeds the interference threshold of DV2, cumulative interference will occur, resulting in the formation of a hyperedge E1 between them. Similarly, if the interference caused by DV7 exceeds the interference threshold of DV5, independent interference will occur, and a hyperedge will be formed between DV7 and DV5. DV3, DV4, DV5, and DV7 all have independent interference with each other, and three of the four will produce three independent interferences to another one. The sum of the three independent interferences is the cumulative interference.

[0134] The relationship between vertices in the hypergraph can be represented by the adjacency matrix A, where the elements in the kth row and lth column of A have the following values:

[0135]

[0136] Among them, a k,l is the value of the kth row and lth column of the matrix, v k represents the kth vertex, v l represents the lth hyperedge. Based on this rule, Figure 2 The adjacency matrix can be expressed as:

[0137]

[0138] In order to improve the spectrum utilization efficiency, this example introduces the concept of interference threshold, that is, the product of the adjacency matrix and the power matrix must be less than or equal to the interference threshold matrix. This example chooses to set this interference threshold matrix to the bandwidth W / L of each RB, representing the available spectrum range of each RB. The mathematical model is given by the following formula:

[0139] AP≤I (8)

[0140] Wherein, I is an interference threshold matrix, and the elements of I are composed of the interference threshold W / L of each RB. Represents the power matrix. When the element P in the uth row and jth column of P u,j = 0 means that the uth intelligent vehicle (intelligent vehicle u) is not associated with subchannel j. If it is associated, then P u,j =1,u=1,2,…,U.

[0141] In the above model, this example analyzes the interference types and relationships between users and obtains the interference tolerance matrix AP between devices. Through this matrix, this example can observe the resource allocation of all nodes. Setting the interference threshold to the bandwidth of each RB helps ensure that the interference received by the device during communication is within an acceptable range. This method helps optimize spectrum utilization, reduce interference, and improve communication quality, thereby improving the performance and stability of drone-assisted 6G-ITS.

[0142] (5) Step S5: Energy efficiency (EH) calculation

[0143] Based on equations (3) and (4), the input power of the D2D pair is given by:

[0144]

[0145] Therefore, the energy obtained by the drone can be expressed as:

[0146]

[0147] in is a constant representing the energy conversion efficiency.

[0148] However, in practice, the EH circuit exhibits nonlinear characteristics. That is, in the low power area, the energy conversion efficiency increases with the increase of input power, but when the power is high enough, the collected energy will reach saturation. Therefore, when the energy receiver is at a high receiving power, the linear EH model may not be accurate enough, which may cause the system to fail to effectively utilize the available energy, thus affecting the stability and reliability of the network. In order to ensure accurate analysis of the EH model, this example uses a nonlinear EH model to consider the actual EH circuit situation. Therefore, the EH on the drone can be given by the following formula:

[0149]

[0150] Where a and b correspond to parameters related to the specification of the EH circuit. max It indicates the maximum EH value of the drone when the EH circuit is saturated, and exp() refers to the exponential function with the natural constant e as the base.

[0151] From equation (11), we can observe that as the power P in As , the EH efficiency will first increase and then decrease.

[0152] Figure 3 Comparison of EH between linear EH model and nonlinear EH model, where a=150, b=0.014, P max =24mW, Note that these are not the data used in the simulation analysis section below. In this example, it can be observed that the nonlinear EH model gradually approaches saturation as the input power increases. In contrast, the EH of the linear EH model continues to increase linearly.

[0153] In addition, the total power consumption of drone-assisted 6G-ITS is expressed as follows:

[0154]

[0155] Where α is the power amplifier efficiency factor. x Indicates the constant power consumption of the circuit module.

[0156] According to formula (1), the user's total rate expression is obtained in the following way:

[0157]

[0158] γ n,j represents the signal-to-noise ratio of UV n on the jth subchannel.

[0159] Therefore, the EE of drone-assisted 6G-ITS can be expressed as:

[0160]

[0161] (6) Step S6: Optimization Problem

[0162] Energy consumption is an important challenge faced by drone-assisted 6G-ITS. Based on equations (2), (5), (8), (11), and (14), this example constructs an EH optimization model to maximize the EE of the network and achieve more efficient energy utilization. The mathematical model is shown in the following equation:

[0163]

[0164] Among them, max means maximization, st means that it must be satisfied, and constraint C1 means that the SINR of each AV is not less than the AV minimum SINR threshold Constraint C2 indicates that the SINR of each D2D pair is not less than the minimum SINR threshold of the D2D pair. Constraint C3 means that the collected energy of each drone is not less than the minimum energy threshold E min Constraint C4 means that the product of the adjacency matrix and the power matrix must be less than or equal to the interference threshold matrix. Constraints C5 and C6 respectively mean that the transmission power of AV and UV is not less than their maximum power thresholds. and

[0165] In the robust optimization model, the uncertainty of channel gain makes it difficult to solve the global optimal solution of the robust optimization model using traditional optimization methods, which poses a challenge to the solution of the EH optimization model shown in equation (15).

[0166] (7) Step S7: Channel Gain Learning

[0167] In this section, the goal of this example is to use the channel samples collected by the 3GPP protocol for learning to reduce the impact of imperfect CSI. In order to reduce the uncertainty of the channel gain, this example can model the channel gain as a polyhedron model to describe the uncertainty of the channel.

[0168] Assume that the collected N UV channel data sets are The channel data set of M AVs is Next, the channel gain model is expressed as follows:

[0169]

[0170]

[0171] in, represents the channel gain modeling of AV, Represents the channel gain modeling of UV, represents the channel gain set of the mth AV on the jth subchannel, represents the channel gain set of the nth UV on the jth subchannel, and Represents auxiliary variables.

[0172] represents the channel gain of the mth AV on the jth subchannel, represents the channel gain of the nth UV on the jth subchannel, the superscript T represents the matrix transpose, and || represents the absolute value. In equations (16) and (17), the learned parameters include Z and X.

[0173] This example uses the sample mean to determine and Therefore, in this case:

[0174]

[0175] H n Right now The hth sample of m Right now The mth sample of .

[0176] Next, by calculating the parameters Z and X, this example can obtain the learned channel gain parameters. However, the uncertainty of incomplete CSI brings difficulties to the learning of the data set. Therefore, this example uses the confidence method to construct the prior distribution of incomplete CSI. It can be expressed as:

[0177]

[0178] Among them, 1-θ is the specified confidence level. Pr{} represents the probability of the event in the brackets. express The prior distribution of express The prior distribution of .

[0179] Next, this example needs to confirm the size of the polyhedron model. The calibration function is as follows:

[0180]

[0181] f(D) represents the sample set The calibration function, f(H) represents the sample set Calibration function, D(0) and H(0) represent the first sample in the sample set, and D(1) and H(1) represent the second sample in the sample set.

[0182] The task of the calibration function is to adjust the uncertainty set to align with equation (19) with a confidence level of 1-θ.

[0183] Then, this example defines the (1-θ) quantile q of the underlying distribution of f(D) and f(H) based on the sample 1-θ as follows:

[0184]

[0185] By calculating the function values ​​of f(D) and f(H) on each sample, we can get the corresponding function values ​​and then sort the function values. and are the upper bounds of the 1-θ quantiles of f(D) and f(H), respectively. Indicates rounding up. The sizes of the uncertain sets D and H can be obtained as follows:

[0186]

[0187] f(D * ) indicates that D * The solution of the calibration function obtained, f(H * ) indicates that H * The solution to the calibration function is obtained.

[0188] Finally, substitute (22) and (18) into (16) and (17) respectively to obtain the learned

[0189] The specific steps of channel learning are shown in Algorithm 1 in Table 1.

[0190] Table 1

[0191]

[0192] (8) Step S8: Model solution

[0193] Before solving, since the EH problem is not NP-hard, this example needs to transform the EH model into a convex optimization problem through the duality theorem. On this basis, the IT-EHRA strategy is used to solve the model. The solution process is as follows.

[0194] First, this example transforms the objective function of the model based on the Dinkelbach method, and its expression is as follows:

[0195]

[0196] in, Is a auxiliary parameter convex function. When the optimal solution of the model is p* (representing and )hour, for:

[0197]

[0198] According to the above analysis, the original optimization problem (15) is transformed into:

[0199]

[0200] Then, the Lagrangian function of problem (25) is:

[0201]

[0202] in, and is a non-negative Lagrange multiplier, subscript n corresponds to UV, and subscript m corresponds to AV. Then, in order to facilitate the solution, this example converts equation (26) into equation (27):

[0203]

[0204] in,

[0205] Finally, the optimization model (25) becomes:

[0206]

[0207] That is the function F.

[0208] Based on the KKT condition of convex optimization theory, the following results can be obtained in this example:

[0209]

[0210] By solving (29) we can obtain the transmission power p n,j and p m,j The optimal solution and Then, according to the definition of signal-to-interference-noise ratio, the optimal solution is obtained, thereby obtaining the comprehensive efficiency of drone-assisted 6G-ITS.

[0211] Algorithm 2 shown in Table 2 gives the pseudo code of the IT-EHRA resource allocation method. The specific steps are as follows:

[0212] First, obtain relevant channel data based on Algorithm 1;

[0213] Then, the optimization model is transformed into a Lagrangian dual problem;

[0214] Finally, this example uses the KKT condition to solve the problem.

[0215] Table 2

[0216]

[0217] Algorithm 2 specifically includes the following steps:

[0218] 1) Obtain a sample data set through Algorithm 1, and use the sample data set for feature extraction and parameter learning, including data set separation, shape learning, and size calibration steps, to learn channel gain parameters and obtain communication link gain;

[0219] 2) Initialization Set the step size k t =k, t = 1, 2, ..., T, convergence accuracy Ψ, initial point and Gradient Function

[0220] 3) Calculate the function by formula (29) and and gradient In each iteration, the parameters are updated as follows:

[0221] p k+1 =p k +k t η k

[0222] When the absolute value of the difference between two points ||F(p k+1 )-F(p k )||>Ψ, the iteration ends and the transmission power is output and

[0223] 4) Finally, the two different EH models are reconstructed into corresponding Lagrangian dual problems, using the transmission power and And the corresponding solution is obtained through Lagrangian KKT conditions.

[0224] Based on the above method, an embodiment of the present invention also provides a resource allocation system for a drone-assisted 6G-supported intelligent transportation system, which is provided with an intelligent agent, and the intelligent agent is used to implement the above resource allocation method.

[0225] (9) Simulation

[0226] During the simulation, the performance of the proposed algorithm, namely IT-EHRA algorithm, is evaluated. For comparison, the robust allocation algorithm (RAA) is introduced, which reduces the impact of imperfect CSI through another channel modeling to obtain the optimal power allocation. At the same time, this example also compares this method with the joint robust algorithm (JRA) that uses channel modeling samples as ellipse sets, and its channel training process is similar to that of the proposed algorithm. In addition, this example also uses the non-robust algorithm (NRA) for comparison, which does not consider the case of imperfect CSI of channel gain.

[0227] 1) Simulation settings

[0228] In the simulation setting, the base station coverage is a circle with a radius of 500m, the D2D link distance is 20m, the BS bandwidth W is 10MHz, the D2D pair number is 30, and the AV number is 30. In the drone-assisted 6G-ITS, this example refers to the channel model defined in 3GPP and uses the channel model of the 5G toolbox in Python software to generate the required imperfect CSI channel samples. The data set contains 1000 imperfect CSI samples. For the nonlinear EH model, a=1500, b=0.0022, P max =3.9. The other parameters are shown in Table 3.

[0229] Table 3

[0230]

[0231] 2) Performance evaluation

[0232] Figure 4 The minimum harvest energy threshold E of the four algorithms is shown min Relationship with the total EE of the network. Figure 4 It can be observed that in the case of imperfect CSI, the EE of the IT-EHRA algorithm is better than other baseline algorithms in UAV-assisted 6G-ITS. This is because there is imperfect CSI in the network, which may cause errors in the transmission and communication process. The IT-EHRA algorithm can handle these errors and incomplete information more effectively by building a channel model, thereby improving the performance and efficiency of the system. At the same time, in this example, it can be seen that with the minimum energy harvesting threshold E min As the minimum energy harvesting threshold E min As the minimum energy harvesting threshold E increases, the system needs to collect more energy to meet the requirements, resulting in increased user energy consumption and thus increased power consumption. min When increases, although the system may harvest more energy, it will also consume more energy to meet the increasing requirements, resulting in a lower energy efficiency of the algorithm.

[0233] Figure 5 The minimum energy harvesting threshold E of the four algorithms is shown. min Relationship with total network throughput. Figure 5 It shows that the IT-EHRA algorithm has a higher throughput than other baseline algorithms. At the same time, it is also observed that the throughput increases with the minimum energy harvesting threshold E min This is because when E min As the minimum energy harvesting threshold E increases, the energy consumed by the drone will also increase, which leads to an increase in power consumption. minWith the increase of EE, the UAV-assisted 6G-ITS network decreases and the throughput increases.

[0234] Figure 6 Shows the power thresholds of the four algorithms Relationship with drone-assisted 6G-ITS network performance, where Figure 6 (a) is the power threshold Relationship with the total EE of the network, Figure 6 (b) is the power threshold Relationship with the total network throughput. Figure 6 (a) and Figure 6 (b) shows that when the power threshold When increases, the throughput of the network also increases, but the EE of the network decreases. The proposed IT-EHRA strategy has higher efficiency and throughput than the comparison algorithms, proving that the performance of the IT-EHRA algorithm is superior to the baseline algorithm. This indicates that in the drone-assisted 6G-ITS with imperfect CSI, the IT-EHRA algorithm can obtain better QoS than the comparison algorithms, thus improving the system performance and enhancing reliability.

[0235] Figure 7 The minimum SINR thresholds of D2D pairs for the four algorithms are shown Relationship with the total EE of the network. Figure 7 It can be seen that when the minimum SINR threshold When the minimum SINR threshold increases, the EE of the drone-assisted 6G-ITS network decreases. Under the condition of , the IT-EHRA algorithm has better EE performance than the baseline algorithm. This is because the proposed IT-EHRA algorithm has good robustness and can effectively reduce the impact of imperfect CSI on UAV-assisted 6G-ITS.

[0236] Figure 8 The minimum SINR thresholds of D2D pairs for the four algorithms are shown Relationship with the total network throughput. Figure 8 It can be seen that when the minimum SINR threshold When increases, the throughput of the UAV-assisted 6G-ITS network increases. This improvement is attributed to the ability of the IT-EHRA algorithm to reduce the channel uncertainty, thereby improving the channel availability and increasing the overall throughput. In addition, it can be observed in this example that the proposed IT-EHRA algorithm has a better throughput than the baseline algorithm.

[0237] Therefore, the above simulation results show that for the drone-assisted 6G-ITS network, the IT-EHRA algorithm outperforms the baseline algorithm in terms of EE and total throughput. This shows that the IT-EHRA algorithm proposed in this example can effectively improve the network EE and throughput, thereby improving the communication quality of the drone-assisted 6G-ITS network.

[0238] In summary, the resource allocation method and system for the drone-assisted 6G-supported intelligent transportation system provided in the embodiment of the present invention analyzes the types and relationships of interference in the drone-assisted 6G-ITS network, establishes an interference hypergraph model, and designs a method for reducing overlapping interference based on the interference hypergraph model to improve spectrum utilization efficiency. Then, based on power constraints, interference constraints, EH constraints, and signal-to-interference ratio constraints, an EH optimization model under imperfect CSI is established. Finally, the IT-EHRA algorithm is used to obtain the power allocation under the maximum EE of the network. The simulation results show that the method and system can effectively improve energy utilization, extend the working time of the drone, and improve network throughput, providing a feasible solution for resource allocation of the drone-assisted 6G-ITS network.

[0239] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A resource allocation method for a drone-assisted 6G-supported intelligent transportation system, characterized in that: Includes steps: S1. Determine the network structure of drone-assisted 6G supporting intelligent transportation system; In the step S1, the drone-assisted 6G-supported intelligent transportation system includes U intelligent vehicles, namely CAVs, and B base stations, namely BSs, wherein the U intelligent vehicles are divided into M authorized vehicles, namely AVs, and N unauthorized vehicles, namely UVs; the drone-assisted 6G-supported intelligent transportation system is also provided with N drones, namely UAVs, wherein a UV and a drone form a D2D, namely a device-to-device pair, and the drone transmits an uplink spectrum to the UV by multiplexing the AV; the bandwidth of the drone-assisted 6G-supported intelligent transportation system is W Hz, which is evenly divided into L resource blocks, namely RBs, and each resource block has W / L Hz; the drone-assisted 6G-supported intelligent transportation system has a total of J subchannels; S2. Determine the channel model of the drone-assisted 6G-supported intelligent transportation system based on the network structure of the drone-assisted 6G-supported intelligent transportation system; S3, determining the interference type of the drone-assisted 6G-supported intelligent transportation system based on the network structure of the drone-assisted 6G-supported intelligent transportation system; In the step S3, the D2D pair and the AV are uniformly represented as DV, and DV is used to represent any one of the D2D pair and the AV, and the interference types of the drone-assisted 6G supporting intelligent transportation system include cumulative interference and independent interference; Cumulative interference means that when multiple DVs are running simultaneously, the interference generated by other DVs exceeds the interference threshold of a single DV, and cumulative interference is generated between the multiple DVs running simultaneously; Independent interference means that in a scenario where two DVs are operating, when the interference generated by one DV exceeds the interference threshold of the other DV, independent interference will occur; S4. Based on the hypergraph theory, a hypergraph interference model of network structure and interference type based on UAV-assisted 6G-supported intelligent transportation system is constructed; S5. Determine the energy efficiency calculation method of the drone-assisted 6G-supported intelligent transportation system based on the network structure and channel model of the drone-assisted 6G-supported intelligent transportation system; In step S5, the energy efficiency calculation method of the drone-assisted 6G-supported intelligent transportation system is: R SUM represents the total rate of users, P SUM Indicates the total power consumption of the system; R SUM Calculated by the following formula: Among them, γ n,j represents the signal-to-noise ratio of UV n on the jth subchannel, γ m represents the signal-to-interference-plus-noise ratio of AV m; P SUM Calculated by the following formula: Where α is the power amplifier efficiency factor, P x Represents the constant power consumption of the circuit module, p m,j and p n,j denote the transmission power of AV m and D2D pair n on the jth subchannel, i.e., subchannel j; S6. With the goal of maximizing the energy efficiency of the drone-assisted 6G-supported intelligent transportation system, with the constraint that the interference value calculated based on the hypergraph interference model does not exceed the interference threshold, and with the constraint that the signal interference plus noise ratio requirement, energy efficiency requirement, and power requirement of the drone-assisted 6G-supported intelligent transportation system are met, an optimization problem for optimizing the transmitter transmission power in the drone-assisted 6G-supported intelligent transportation system is constructed; S7, determining channel gain data of the UAV-assisted 6G-supported intelligent transportation system based on channel learning; The step S7 specifically includes the following steps: S71. Collect N UV channel data sets: The channel data set of M AVs is S72. Model the channel gain as a polyhedron model: in, represents the channel gain modeling of AV, Represents the channel gain modeling of UV, represents the channel gain set of the mth AV on the jth subchannel, represents the channel gain set of the nth UV on the jth subchannel, and Represents auxiliary variables; S73, using the sample mean of the channel data set to respectively determine and The value of S74. Set the reliability to a calibration function of 1-θ to adjust the sample set: express The prior distribution of express Prior distribution of , Pr{} represents the probability of the event in brackets; S75. Calculate Z and X: f(D * ) indicates that D * The solution of the calibration function obtained, f(H * ) indicates that H * The solution to the calibration function is obtained, and are the upper bounds of the 1-θ quantiles of f(D) and f(H), respectively, where f(D) represents the sample set The calibration function, f(H) represents the sample set The calibration function of S76, the result calculated in step S73 is and Substitute the Z and X calculated in step S75 into the polyhedron model constructed in step S72 to obtain the learned channel gain data; S8. Solve the optimization problem based on the channel gain data to obtain the transmission power of the transmitter in the UAV-assisted 6G-supported intelligent transportation system.

2. The resource allocation method for the UAV-assisted 6G-supported intelligent transportation system according to claim 1, characterized in that: In step S2, the channel model of the drone-assisted 6G-supported intelligent transportation system is constructed as follows: The BS uses non-orthogonal multiple access to provide services for CAVs, and the UAV uses wireless communication and energy carrying technology to divide the received signal into two parts: information decoding and energy collection; The received signal x of the mth AV, AV m, and the nth D2D pair, D2D pair n m and x n Respectively expressed as: in, and denote the channel gains of AV m and D2D pair n on subchannel j, n m 、n n are the additive white Gaussian noise on AV m and D2D pair n, respectively, m=1,2,…,M, n=1,2,…,N, j=1,2,…,J; The signal to interference plus noise ratio or SINR of AV m is given by: in, is the noise power of AV m, i corresponds to AV i, i ≠ m; Each drone will receive data x n Divided into two parts: Where ω is the energy distribution factor, σ I represents the noise generated by the energy allocation method; The SINR of D2D pair n is expressed as: x corresponds to the D2D pair x, x≠n, represents the covariance matrix of the noise of D2D pair n, represents additive white Gaussian noise.

3. The resource allocation method for the UAV-assisted 6G-supported intelligent transportation system according to claim 1, characterized in that: In the step S4, a hypergraph interference model is constructed with DVs as vertices and interference relationships between DVs as hyperedges; The relationship between vertices in the hypergraph interference model is represented by the adjacency matrix A, and the values ​​of the elements in the kth row and lth column are as follows: Among them, v k represents the kth vertex, v l represents the lth hyperedge; The product of the adjacency matrix A and the power matrix P must be less than or equal to the interference threshold matrix I, and the interference threshold matrix is ​​set to the bandwidth W / L of each RB.

4. The resource allocation method for the UAV-assisted 6G-supported intelligent transportation system according to claim 1, characterized in that: In step S6, the optimization problem is constructed as: Among them, max means maximization, st means need to be satisfied, C1 to C6 represent different constraints. Indicates the minimum AV SINR threshold, represents the minimum SINR threshold of D2D, E represents the collection energy on the drone, and E min Indicates the minimum energy collection threshold of the drone. and Respectively represent the maximum threshold of transmission power of AV and UV; E is calculated by the following formula: Where a and b correspond to parameters related to the specification of the energy harvesting circuit, P max represents the maximum energy harvesting value of the drone when the energy harvesting circuit is saturated, exp( ) refers to the exponential function with the natural constant e as the base, P in Indicates the input power of the D2D pair.

5. The resource allocation method for the UAV-assisted 6G-supported intelligent transportation system according to claim 4, characterized in that: The step S8 specifically includes the following steps: S81, converting the optimization problem into: max: stC1-C6 in, Represents auxiliary variables; S82, transform the model shown in step S81 into a Lagrangian dual problem: in, is the Lagrangian function, and is a non-negative Lagrange multiplier, Intermediate variables S83. Based on the KKT condition of convex optimization theory, find F(x) at p n,j and p m,j The partial derivative at ; S84, let F(x) be n,j and p m,j The partial derivative at is 0, and the transmission power p is obtained by solving n,j and p m,j The optimal solution and 6. UAV-assisted 6G-supported resource allocation system for intelligent transportation systems, characterized by: An intelligent agent is provided, and the intelligent agent is used to implement the resource allocation method described in any one of claims 1 to 5.

Images