Multi-Dimensional Resource Allocation Method and System for Mobile Antenna-Assisted Hybrid NOMA

Abstract

The present invention discloses a multi-dimensional resource allocation method and system for mobile antenna-assisted hybrid NOMA. The method includes: the base station constructs an optimization problem for minimizing energy consumption with the total offloading energy consumption of all users as the objective function, and minimizes the total offloading energy consumption of all users by optimizing the user offloading power, offloading time slots, and the positions of the users' mobile antennas; sets the initial values of the offloading time slots, user offloading power, and the positions of the users' mobile antennas, and enters an iterative loop: given the positions of the users' mobile antennas, jointly optimizes the offloading time slots and user offloading power; given the offloading time slots and user offloading power, optimizes the positions of the users' mobile antennas; performs iterations until convergence to obtain the optimal values of the offloading time slots, user offloading power, and the positions of the users' mobile antennas. The present invention jointly optimizes the offloading time slots, user offloading power, and the positions of the users' mobile antennas, improving the offloading flexibility of users and reducing the offloading energy consumption of users.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication, and particularly to a method and system for mobile antenna-assisted hybrid NOMA multi-dimensional resource allocation. Background Art

[0002] With the explosive growth of computationally intensive applications such as virtual reality (VR), augmented reality (AR), and autonomous driving in the wireless Internet of Things, mobile edge computing (MEC) technology has emerged as an important enabling technology for the next-generation mobile communication network. The core goal of MEC is to offload computing tasks from resource-constrained mobile devices (such as smartphones, sensors, etc.) to the infrastructure at the network edge (such as the MEC server deployed by the base station), and use the powerful computing power of the edge server to achieve efficient task processing. Compared with local computing, MEC technology can not only shorten the task processing delay through the high-performance computing unit of the edge server, but also avoid the continuous high-load operation of the local processor through computing offloading, thereby effectively extending the device battery life and reducing energy consumption.

[0003] However, the large-scale application of MEC faces the following challenges: on the one hand, the limited computing / storage resources of the edge server are difficult to meet the dynamic needs of multi-user and multi-task, and an efficient resource allocation mechanism is required; on the other hand, the scarcity of wireless spectrum resources and the terminal energy limitation make the computing offloading and resource allocation under limited resources a complex optimization problem. In this context, non-orthogonal multiple access (NOMA), also an important emerging technology for the next-generation wireless network, has become the key to breaking through the performance bottleneck of MEC by virtue of its spectral efficiency advantage. NOMA allows multiple users to share the same time-frequency resource block for signal transmission, and uses successive interference cancellation (SIC) technology to separate signals. In the MEC scenario, NOMA supports multi-users to concurrently transmit offloading requests through power domain multiplexing, effectively improving the system throughput.

[0004] Although NOMA-assisted MEC has shown significant advantages in improving the offloading transmission efficiency and computing offloading capacity, it still faces challenges such as insufficient dynamic resource adaptation and single-dimensional resource allocation in practical applications. Specifically, the performance of NOMA transmission depends on the channel conditions of mobile devices. The heterogeneous task loads of mobile terminals (such as the coexistence of computationally intensive tasks and low-latency sensitive tasks), the different degrees of urgency of terminal offloading, and the fixed channel conditions will lead to low offloading transmission efficiency and limitations in device fairness, thus reducing the performance of the overall system. Therefore, it is urgent to develop a new type of transmission technology to enhance the dynamic adaptability and robustness of the system to meet the requirements of future ultra-dense, low-latency, and high-reliability mobile edge computing scenarios. Summary of the Invention

[0005] To at least solve one of the problems existing in the prior art, the present invention provides a method and system for multi-dimensional resource allocation of a movable antenna-assisted hybrid NOMA (non-orthogonal multiple access), which can actively regulate the channel gain to reduce the offloading transmission energy consumption of mobile devices in a mobile edge computing network.

[0006] A method and system for multi-dimensional resource allocation of a movable antenna-assisted hybrid non-orthogonal multiple access are provided for a mobile edge computing (MEC) network. In this application, consider an MEC network including a single-antenna base station deploying an MEC server and multiple users equipped with single movable antennas, where each user needs to compute a computationally intensive and latency-critical task and offload it to the base station. During the offloading process, the offloading volume and the maximum offloading time of each user are different. In this application, the architecture of the movable antenna can be to adjust the position of the antenna through a mechanical structure or to achieve dynamic adjustment of the antenna by using the morphological changes of fluid materials such as liquid metal and electrolyte. This application adopts an MEC strategy of hybrid NOMA transmission, and the offloading of users is arranged according to the urgency of their tasks. Different from the MEC strategy of orthogonal transmission, when a user with a more urgent task offloads, other users with less urgent tasks can still offload. Different from the MEC strategy of traditional NOMA transmission that forces all users to start and complete offloading simultaneously, the MEC strategy of hybrid NOMA transmission divides the transmission cycle into multiple offloading time slots, and users can offload task data segmentally in different offloading time slots. The task data in the same offloading time slot is transmitted using NOMA. This strategy not only improves the offloading efficiency but also provides users with more flexible opportunities to offload tasks.

[0007] In the method proposed in this application, the base station estimates the channel state information (CSI) of each user's communication link through channel estimation and compressive sensing methods, including the response of each transmission path, the angle of departure, and the angle of arrival. The base station divides the entire transmission offloading period into offloading time slots equal to the number of users. Users perform offloading in different time slots according to the urgency of their own tasks and the amount of offloaded data. Different users use NOMA transmission to offload task data in the same offloading time slot. The offloading completion time of users will restrict users to complete offloading in specific offloading time slots, and at the same time, the offloading time slots will also be limited by the offloading completion time of each user. For example, the user with the shortest offloading completion time, that is, the user considered to have the highest task urgency, must complete offloading in the first offloading time slot, and the duration of the first offloading time slot shall not be greater than the offloading completion time of the user with the shortest offloading completion time. The user with the second shortest offloading completion time can perform task data offloading in the first two time slots, but must also complete offloading within the first two time slots, and the total duration of the first two offloading time slots shall not be greater than the offloading completion time of the user with the second shortest offloading completion time. And so on, the user with the longest offloading completion time can perform task data offloading in all offloading time slots, and the total duration of all offloading time slots shall not be greater than the offloading completion time of the user with the longest offloading completion time.

[0008] In the method proposed in this application, the offloading time slots, user offloading power, and position of the movable antenna are jointly optimized and designed as follows:

[0009] (1) The base station will construct an optimization problem for minimizing energy consumption with the goal of minimizing the total offloading energy consumption of users based on the obtained channel response information, the urgency of user offloading, and the amount of user task offloading, and design the offloading time slots, user offloading power, and the position of the user's movable antenna. In each offloading time slot, the movable antennas equipped by all users will adjust their positions to actively cause the time-varying characteristics of the channel on the user link to increase the additional time diversity gain and further reduce the offloading energy consumption of network users.

[0010] (2) The offloading time slots, user offloading power, and the position of the movable antenna are obtained through alternating iterative calculations to get the optimal values. At the initial stage of the iteration, the initial offloading time slots, user offloading power, and the position of the movable antenna are set. In each iteration, given the position of the movable antenna in the previous iteration, the offloading time slots and user offloading power are jointly designed; given the offloading time slots and user offloading power in this iteration, the positions of the movable antennas of all users in all offloading time slots are jointly designed until convergence or the maximum number of iterations is reached. The optimal offloading time slots, optimal user offloading power, and optimal positions of the users' movable antennas are obtained. After the base station calculates the above parameters, they are sent to each user through the downlink control channel.

[0011] To implement the proposed method, a multi-dimensional resource allocation system for a mobile antenna-assisted hybrid NOMA is constructed. The system includes a channel reconstruction module, a multi-dimensional resource collaborative computing module, a multi-dimensional resource execution module, and a wireless transmit / receive transmission module. The channel reconstruction module is used to obtain the multi-path response parameters of the wireless channel from the base station to the mobile user, including the angle of departure, the angle of arrival of each transmission path, and the complex response value of each path, and generate a channel state information matrix. The multi-dimensional resource collaborative computing module is mainly used to jointly calculate the offloading time slot, the user offloading power, and the position of the user's mobile antenna. The function of the multi-dimensional resource execution module is to write the offloading time slot into the scheduler of the base station, and configure the user radio frequency front-end and the position of the mobile antenna according to the offloading time slot, the user offloading power, and the mobile antenna. The wireless transmit / receive transmission module is used for receiving and transmitting signals between the base station and the mobile user. Before the mobile user offloads the task, the channel reconstruction module extracts the multi-path response parameters through a sparse signal recovery algorithm, and this parameter serves as the input environmental state of the multi-dimensional resource collaborative computing module.

[0012] Compared with the prior art, the present invention can at least achieve the following beneficial effects:

[0013] The present invention can be applied in the mobile edge computing network scenario. By introducing a mobile antenna in the offloading transmission scheme of hybrid non-orthogonal multiple access, the time diversity of the computing network is increased. The offloading time slot, the user offloading power, and the position of the user's mobile antenna in the network are jointly designed to obtain the optimal values, which can improve the offloading flexibility of users in the network and reduce the user offloading energy consumption. Description of the Drawings

[0014] Figure 1 It is a flowchart of the multi-dimensional resource allocation method for a mobile antenna-assisted hybrid NOMA provided by an embodiment of the present application.

[0015] Figure 2 It is a schematic diagram of an implementation environment provided by an embodiment of the present application.

[0016] Figure 3 It is a diagram of a hybrid NOMA-assisted MEC transmission strategy provided by an embodiment of the present application.

[0017] Figure 4 It is a flowchart of the joint design and optimization of the offloading time slot, the user offloading power, and the position of the user's mobile antenna provided by an embodiment of the present application. Detailed Embodiments

[0018] To make the objectives, technical solutions, and advantages of this application more clear, the following will further describe the embodiments of this application in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0019] The core of the embodiments of the present invention is to provide a multi-dimensional resource allocation method and system for mobile antenna-assisted hybrid non-orthogonal multiple access (NOMA) to reduce the offloading energy consumption of users in the mobile edge computing (MEC) scenario.

[0020] A multi-dimensional resource allocation method for mobile antenna-assisted hybrid NOMA provided by the embodiments of the present invention is applied to a mobile edge computing network, such as Figure 2 , the mobile edge computing network includes a base station with a single antenna and deployed with an MEC server and K users equipped with single movable antennas, where the user index and the user index set are respectively represented as k and . Each user needs to calculate a computationally intensive and latency-critical task and offload it to the base station. During the offloading process, the offloading amount and the maximum offloading time of each user are different. The offloading data volume and the maximum offloading time of user k are respectively represented as and . In each offloading cycle, the user uses a hybrid NOMA transmission method to offload task data. First, the base station divides the entire transmission offloading cycle into offloading time slots equal to the number of users, where the i th offloading time slot is represented as . The user offloads tasks with different offloading powers in different offloading time slots according to the task offloading urgency of their own tasks and the data volume required for offloading tasks. Different users use the NOMA transmission method to offload task data in the same offloading time slot. The offloading completion time of the user will constrain the user to complete the offloading in a specific offloading time slot, and at the same time, the offloading time slot is also limited by the offloading completion time of each user. The task offloading urgency of the user is judged according to the offloading completion time of the user. In one of the embodiments of the present invention, it is set that the task offloading urgency of the user is strongly correlated with the user offloading completion time, that is, the shortest user offloading completion time corresponds to the highest offloading urgency, and vice versa, the lowest urgency. To better describe the method proposed by the embodiments of the present invention, assume that the relationship of the user offloading completion time is as follows , that is, the offloading completion time of user 1 The shortest, regarded as the user with the highest offloading urgency of the task, the user K 's offloading completion time The longest, regarded as the user with the lowest offloading urgency of the task. According to the hybrid NOMA criterion, user 1 must complete offloading in the first offloading time slot and shall not be greater than the offloading completion time of the user with the shortest offloading completion time, that is shall satisfy ; User 2 can perform task data offloading in the first two offloading time slots, and the total duration of the first two offloading time slots shall not be greater than the maximum offloading time of user 2, that is, the following constraint is satisfied ; And so on, the k th user can perform task data offloading before the k th offloading time slot, and its offloading time slot satisfies , represents the offloading completion time of the k th user; and the total duration of all offloading time slots shall not be greater than the offloading completion time of the user with the longest offloading completion time. The hybrid NOMA criterion is as Figure 3 shown, where in the th offloading time slot, only users i to user K can perform task offloading, because users 1 to user i -1 have completed offloading. In the th offloading time slot, only user K performs task offloading, because other users have completed offloading. In this embodiment, considering the far-field response channel with slow fading, where the moving area size of the movable antenna of each user is significantly smaller than the signal propagation distance, and the movable antenna of user k moves in the one-dimensional feasible region . Therefore, it can be reasonably assumed that the angle of departure (AoD), the angle of arrival (AoA), and the amplitude of the complex path coefficient remain unchanged on each transmission path, and the change in the movable antenna position only affects the change in the complex path coefficient. In one embodiment of the present invention, the received field response vector k from the base station to user and the transmitted field response vector k from user to the base station are expressed as:

[0021] ;

[0022] where represents the imaginary unit; Represents the radiation wavelength of the signal; Represents the position of the base station's fixed antenna; Represents the distance from the base station to the user k At the arrival angle of the -th received path, Represents the number of received paths; Represents the user k to the base station at the transmission angle of the -th transmitted path, Represents the number of transmitted paths, Represents the user k 's movable antenna at the i -th offloading time slot.

[0023] The method provided by the embodiments of the present invention is as follows:

[0024] Step 1: The base station divides an offloading transmission period into multiple variable offloading time slots, and constructs an optimization problem of minimizing energy consumption with the total offloading energy consumption of all users as the objective function according to the channel response information, the urgency of the user's offloading task, and the amount of data required for the user's offloading task , by optimizing the user offloading power , offloading time slots and the position of the user's movable antenna to minimize the total offloading energy consumption of all users, jointly designing the offloading time slots, user offloading power, and the position of the user's movable antenna, where represents the set of user k offloading power, represents the offloading power of user k at the 1st offloading time slot, represents the offloading power of user k at the i -th offloading time slot, represents the offloading power of user k at the k -th offloading time slot; represents the set of the positions of the movable antennas of user k , represents the position of the movable antenna of user 1 at the i -th offloading time slot, represents the position of the movable antenna of user k at the i -th offloading time slot, represents the position of the movable antenna of user k at the k -th offloading time slot. Set the initial iteration value , set the initial offloading time slot , the initial user offloading power and the initial user movable antenna position .

[0025] Step 2: Given the movable antenna position of the user in the current iteration , define the optimization variables according to the coupling relationship between the offloading time slot and the user offloading power and , denotes the optimization variable of the i -th offloading time slot, denotes the energy consumption optimization variable of user k in the i -th offloading time slot, denotes the offloading power of user k in the i -th offloading time slot. Substitute and into the optimization problem of minimizing energy consumption , so as to convert the optimization problem of minimizing energy consumption into a convex difference programming problem . Solve the convex difference programming problem by using the iterative concave-convex programming method to obtain the optimal offloading time slot and the optimal user offloading power under the given movable antenna position of the user.

[0026] The expression of the convex difference programming problem is as follows:

[0027]

[0028] where denotes the number of users; the corresponding set of user indices is denoted as ; denotes the energy consumption optimization variable of user m in the i -th offloading time slot; denotes the power of Gaussian white noise; denotes the channel response coefficient of the movable antenna of user m in the i -th offloading time slot; Constraint C1 ensures that the offloading throughput of the user should meet the nats required for the offloading task, where denotes the nats required for the offloading task of user k ; Constraint C2 represents the limitation on the offloading time slot optimization variable under the hybrid NOMA criterion, Constraint C3 ensures that the offloading time slot optimization variable is non-negative, and Constraint C4 ensures that the energy consumption optimization variable is non-negative; in the formula , representing the sub-item of the user offloading throughput expression.

[0029] Solve the convex difference programming problem through an iterative concave-convex programming method to obtain the optimal and , and then calculate the optimal offloading time slot and the user offloading power .

[0030] In one embodiment of the present invention, the steps for solving the convex difference programming problem are as follows:

[0031] (1) Set the initial iteration value , initialize the optimization variables , , representing the value at the n -th iteration, representing the value at the n -th iteration, and respectively representing and the values at the q -th iteration, representing the offloading power of user m at the i -th offloading time slot.

[0032] (2) Perform a first-order Taylor expansion of Equation at the point to obtain the convex upper bound approximation formula . Substitute Equation into Equation in the convex difference programming problem , to obtain the problem .

[0033] (3) Use the interior point method to solve the problem , to obtain the optimal solution 、 、 and of the problem 、 and respectively represent the optimal values of the optimization variables 、 and in the problem

[0034] (4) Let , , , .

[0035] (5) (2), (3) and (4), until convergence, output the optimal value of the offloading time slot of the q th iteration for a given user's movable antenna position and the i th and the user k in the i th time slot of the optimal offloading power . According to and get the optimal offloading time slot q and the optimal user offloading power for a given user's movable antenna position in the th iteration.

[0036] Step 3: Given the offloading time slot and the user offloading power , introduce auxiliary variables and , and the base station constructs an optimization problem for the design of the movable antenna position with respect to , and with the criterion of improving the throughput of the user in each offloading time slot ; then introduce an auxiliary variable , and use the quadratic transformation method to equivalently transform the problem into the problem , that is, the convex approximation expression. Update the auxiliary variable and the user's movable antenna position in turn according to the criterion of alternating iteration, and finally converge to obtain the optimal user movable antenna position for the given offloading time slot and the user offloading power .

[0037] The update rules of the auxiliary variable and the user's movable antenna position are as follows:

[0038] (1) Given the user's movable antenna position , the update rule of the auxiliary variable is as follows:

[0039] ;

[0040] where Denote the optimal value of the auxiliary variable under the given condition; , denote the received field response vector from the base station to the user k ; denote the path response matrix of the user k , and its element denote the response of the transmission link between the user k and the base station from the p -th transmit path to the l -th receive path; , denote the received field response vector from the base station to the user m ; denote the path response matrix of the user m , and its element denote the response of the transmission link between the user m and the base station from the p -th transmit path to the l -th receive path; denote the transmit field response vector from the user k to the base station at the position of the movable antenna ; denote the transmit field response vector from the user m to the base station at the position of the movable antenna ; denote the offloading power of the user k in the i -th offloading time slot.

[0041] (2) Under the condition of the given auxiliary variable , solve the problem by the iterative successive convex approximation method to update the position of the movable antenna of the user, and the specific steps are as follows:

[0042] a. Define the position optimization variable of the movable antenna of the user k in the i -th offloading time slot as , define the function , , where denotes the real part energy gain function of the k -th offloading time slot when the base station receives the movable antenna of the user at the position i ; denote that under the given value of the auxiliary variable , the movable antenna of the user m is at the position Channel gain function of the i th offloading time slot; Denotes taking the real part of the complex value inside the parentheses; Denotes performing conjugate transpose operation on the scalar / vector inside the parentheses; Denotes from user k 's movable antenna position at the i th offloading time slot for the transmit field response vector to the base station; Denotes from user m 's movable antenna position at the i th offloading time slot for the transmit field response vector to the base station. Substitute and into problem , to obtain problem .

[0043] b. Initialize the iterative values , , where denotes the value of the position optimization variable at the u th iteration, denotes at the q th iteration.

[0044] c. Calculate the convex lower and upper bounds of the functions and at the points and respectively by Taylor's principle. Their expressions are as follows:

[0045]

[0046] where, and denote the convex lower and upper bounds of the functions and at the points and respectively, and denote the gradients of the functions and at the points and respectively, denotes the value of at the u th iteration under the iterative successive convex approximation method, denotes the value of at the u th iteration under the iterative successive convex approximation method, and represent the regularization parameters, which are obtained by calculating the upper bounds of the Hessian values at each iteration of and respectively.

[0047] d. By substituting and for in problem and , the convex approximation problem at the u -th iteration, is obtained as follows:

[0048]

[0049] where represents the set of offloading time slot indices corresponding to the user k 's offloading data. The interior point method is used to solve problem , and the optimal solution of the optimization variable in problem is obtained. represents the one-dimensional feasible region of the movable antenna of user k .

[0050] e. Let , , and repeat steps c, d, and e until convergence. Output the optimal position of the movable antenna of the user given the auxiliary variable .

[0051] (3) Repeat steps (1) and (2) until convergence. Output the optimal position q of the movable antenna of the user given the offloading time slot and the user offloading power at the -th iteration.

[0052] Step 4: Calculate the total energy consumption of the user for each iteration based on the optimal offloading time slot and user offloading power obtained from the iteration. Determine whether the convergence condition is satisfied based on the total energy consumption value of the user. Update the variables for the iteration. That is, calculate the total energy consumption value of the user at the -th iteration according to and , calculate the total energy consumption value of the user at the -th iteration according to and , and respectively represent the movable antenna positions of a given user in the th iteration, the optimal offloading time slots and the optimal user offloading power in the case of; determine whether the is satisfied, , where represents the convergence error parameter.

[0053] Step 5: If the condition in Step 4 is not satisfied, repeat Step 2 and Step 3.

[0054] Step 6: If the condition in Step 4 is satisfied, , , . Output the optimal offloading time slots , the optimal user offloading power and the optimal user movable antenna position .

[0055] In one embodiment of the present invention, to implement the method proposed in the foregoing embodiment, a transmission system based on movable antenna-assisted hybrid NOMA is proposed, including a channel reconstruction module, a multi-dimensional resource collaborative computing module, a multi-dimensional resource execution module, a wireless transmission module, and a wireless reception module. The channel reconstruction module is built based on a heterogeneous computing platform (CPU + GPU), and performs sparse recovery on the received signal through a sparse signal recovery algorithm based on compressive sensing, extracts the parameters of several transmission paths with higher energy, calculates the complex response values of each extracted path, and generates a channel state information matrix. Transmit the parameters of the transmission path and the channel state information matrix to the multi-dimensional resource collaborative computing module through an optical fiber link; the multi-dimensional resource collaborative computing module is also built based on a heterogeneous computing platform, and by constructing an optimization model (an optimization problem of minimizing energy consumption ), jointly calculate the offloading time slots, the user offloading power, and the position of the user's movable antenna; the output optimal offloading time slots are written into the scheduler of the base station through the multi-dimensional resource execution module respectively, and configure the user radio frequency front end and the movable antenna position according to the offloading time slots, the user offloading power, and the user's movable antenna; the optimal user offloading power and the position of the user's movable antenna are sent to all mobile users through the wireless transmission module. The mobile users receive and parse to obtain the offloading power and the movable antenna position through the wireless reception module.

[0056] Based on the multi-dimensional resource allocation of the cooperation of time, power, and space domains assisted by a movable antenna in the foregoing embodiments of the present invention, the multi-dimensional resources of the three dimensions of time, power, and space domains are cooperatively optimized to reduce the energy consumption of mobile users in a mobile edge computing network system. The base station realizes time-domain resource scheduling by allocating the offloading time slots of user tasks according to the urgency of the offloading tasks of mobile users, the amount of offloading tasks, and the obtained channel response information; realizes power-domain resource allocation by regulating the offloading power of mobile users according to the hybrid NOMA criterion; mobile users actively regulate the channel gain during the offloading period by changing the spatial position of the equipped movable antenna in each offloading time slot, bringing additional time diversity gain, and realizing spatial resource allocation; according to the relationship between the user offloading completion time and the offloading time slot, the superimposed interference limit of hybrid NOMA, and the spatial movement range of the movable antenna, a joint optimization problem of multi-dimensional resources is constructed to jointly optimize the offloading time slot, the offloading power, and the position of the user's movable antenna. Applied to a movable antenna-assisted mobile edge computing network, the network has a base station deploying a mobile edge computing server and mobile users that need to offload tasks. In the method, the movable antenna is equipped on the mobile user, and the mobile user offloads its own tasks to the mobile edge computing server through the transmission mode of hybrid non-orthogonal multiple access. By constructing a multi-dimensional resource allocation problem in the time, power, and space domains with the goal of minimizing user energy consumption, the initial offloading time slot, the initial user offloading power, and the initial position of the user's movable antenna are set; then enter the iterative loop. Given the position of the user's movable antenna in the current iteration, the problem is transformed into a difference-of-convex programming form by defining new optimization variables, and then the offloading time slot and the user offloading power are obtained by solving through the iterative convex-concave programming method; in this iteration, by given the offloading time slot and the user offloading power, the base station constructs an optimization problem for the position design of the user's movable antenna with the criterion of enhancing the user offloading throughput, then calculates the convex approximation expression of the relevant throughput expression and obtains the position of the user's movable antenna by solving through the iterative successive convex approximation method; repeat the above two steps until convergence to obtain the optimal offloading time slot, the optimal user offloading power, and the optimal position of the user's movable antenna.

[0057] In summary, this embodiment discloses a multi-dimensional resource allocation method and system assisted by a movable antenna in hybrid NOMA. By jointly designing the offloading time slot, the user offloading power, and the position of the movable antenna in a mobile edge computing network, it realizes the efficient utilization of multi-dimensional resources in the time, power, and space domains, introduces a new time diversity gain, and reduces the user offloading energy consumption in the mobile edge computing network.

[0058] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principle of the present invention shall be equivalent replacement methods and are all included in the protection scope of the present invention.

Claims

1. A multi-dimensional resource allocation method for hybrid NOMA assisted by movable antennas, characterized in that: The method is applied to a mobile edge computing network, where users of the mobile edge computing network are equipped with movable antennas, and tasks are offloaded to a base station where a MEC server is deployed through a hybrid NOMA criterion, including the following steps: The base station divides an offloading transmission cycle into multiple variable offloading time slots, and the base station constructs an optimization problem of minimizing energy consumption based on the channel response information, the urgency of the user's offloading task and the amount of data required for the user's offloading task, taking the total offloading energy consumption of all users as the objective function; Set the initial unloading time slot, the initial user unloading power and the initial position of the user's movable antenna, and enter an iterative loop: given the user's movable antenna position, jointly optimize the unloading time slot and the user's unloading power; given the unloading time slot and the user's unloading power, optimize the user's movable antenna position; repeat the iteration until convergence, and obtain the optimal unloading time slot, the optimal user unloading power and the optimal user's movable antenna position; The specific steps of converting the optimization problem of minimizing energy consumption into a convex programming problem and solving it to obtain the optimal unloading time slot and the optimal user unloading power under the movable antenna position of a given user include: Given the user's movable antenna position, define the optimization variable and , Indicates i The optimization variables of the unloading time slots are Indicates user k In the i The energy consumption optimization variable for unloading time slots is Indicates user k In the i The unloading power of the unloading time slot is and Substitute the energy consumption minimization optimization problem into the convex difference programming problem, solve the convex difference programming problem through the iterative concave-convex programming method, and obtain the optimal unloading time slot and the optimal user unloading power under the given user's movable antenna position; The method of optimizing the user's movable antenna position for a given unloading time slot and user unloading power comprises the steps of: Given unloading time slot and user unloading power , introduce auxiliary variables and The base station is based on enhancing the user offload throughput and building , and The design optimization problem of the movable antenna position; then introduce auxiliary variables , the problem is solved by using the quadratic transformation method Equivalently converted into a convex approximate expression, the auxiliary variables are updated in sequence through the alternating iterative criterion and the user's removable antenna position , and finally converge to obtain the optimal user movable antenna position under given unloading time slot and user unloading power; Auxiliary variables and the user's removable antenna position The updated rules are: The movable antenna position of a given user , for the auxiliary variable The update criteria are: ； In the formula, Indicates that in a given Auxiliary variables in the case The optimal value of , From base station to user k The received field response vector, Indicates user k The path response matrix, whose elements Indicates user k The transmission link between the base station and p Root emission diameter to l The root receiving path response, Indicates the number of receiving paths. Indicates the diameter number of the launch diameter; , From base station to user m The received field response vector, Indicates the conjugate transpose operation of the scalar / vector in the brackets. Indicates user m The path response matrix, whose elements Indicates user m The transmission link between the base station and p Root emission diameter to l Response of the root receiving path; Indicates that from the user k To the base station at the movable antenna position The emission field response vector of Indicates that from the user m To the base station at the movable antenna position The emission field response vector of Indicates user k In the i The unloading power of unloading time slots, Indicates user m In the i Unloading power of unloading time slot; Given auxiliary variables In the case of , the convex approximation expression is solved by an iterative continuous convex approximation method to update the user's movable antenna position ; Repeat for auxiliary variables and the user's removable antenna position The update is performed until convergence, and the optimal user's movable antenna position is output under a given unloading time slot and user unloading power.

2. The multi-dimensional resource allocation method for movable antenna assisted hybrid NOMA according to claim 1, characterized in that: In each unloading time slot, the movable antennas equipped by all users will adjust their positions, actively causing the channel time-varying characteristics on the user link to increase additional time diversity gain.

3. The multi-dimensional resource allocation method for movable antenna assisted hybrid NOMA according to claim 1, characterized in that: According to the hybrid NOMA principle, the base station divides an offloading cycle into offloading time slots equal to the number of users. Users use different offloading powers to offload tasks in different offloading time slots according to the urgency of the offloading task and the amount of data required for the offloading task. In the same unloading time slot, users use NOMA transmission to unload task data. The base station sorts the urgency of users' task unloading. The unloading time slot in which users are allowed to unload tasks depends on the urgency of their task unloading.

4. The multi-dimensional resource allocation method for movable antenna assisted hybrid NOMA according to claim 3 is characterized in that: The urgency of the user's task uninstallation is determined according to the user's uninstallation completion time; the user's uninstallation completion time constrains the user to complete the uninstallation within a specific uninstallation time slot, and the uninstallation time slot is also limited by each user's uninstallation completion time.

5. The multi-dimensional resource allocation method for movable antenna assisted hybrid NOMA according to claim 4, characterized in that: The user with the shortest uninstallation completion time is regarded as the user with the highest task urgency. This user must complete the uninstallation in the first uninstallation time slot, and the duration of the first uninstallation time slot must not be greater than the uninstallation completion time of this user. The user with the second shortest unloading completion time can unload task data in the first two time slots, and must also complete the unloading in the first two unloading time slots, and the total duration of the first two unloading time slots shall not be greater than the unloading completion time of the user with the second shortest unloading completion time; the user with the longest unloading completion time can unload task data in all unloading time slots, and the total duration of all unloading time slots shall not be greater than the unloading completion time of the user with the longest unloading completion time.

6. The multi-dimensional resource allocation method for movable antenna assisted hybrid NOMA according to claim 1, characterized in that: Given the user's movable antenna position, when jointly optimizing the unloading time slot and the user's unloading power, the optimization problem of minimizing energy consumption is converted into a convex difference programming problem and solved, so as to obtain the optimal unloading time slot and the optimal user unloading power under the given user's movable antenna position.

7. A multi-dimensional resource allocation system for implementing the multi-dimensional resource allocation method of movable antenna assisted hybrid NOMA according to claim 1, characterized in that: The system includes the following modules: The channel reconstruction module is used to obtain the multipath response parameters of the wireless channel from the base station to the mobile user, including the transmission angle, arrival angle and complex response value of each transmission path, and generate a channel state information matrix; Multi-dimensional resource collaborative computing module, used to jointly optimize the offloading time slot, user offloading power, and the position of the user's movable antenna; A multi-dimensional resource execution module, used for writing the unloading time slot into the scheduler of the base station, and configuring the user radio frequency front end and the movable antenna position according to the unloading time slot, the user unloading power and the user's movable antenna; Wireless receiving and transmitting module, used for receiving base station and mobile user signals; Wireless transmission module, used for receiving and transmitting base station and mobile user signals.

Images