Optimization method and system for mobile antenna vehicle network including antenna location cost

Abstract

A method and system for optimizing the cost of antenna position in a mobile antenna-based vehicle network, including a mobile antenna, includes: constructing a mobile vehicle network model based on a privacy-preserving federated learning architecture assisted by intelligent mobile antennas. The base station provides corresponding compensation based on the total cost of vehicle users within the base station's signal coverage area, and establishes an objective function that minimizes the base station's compensation cost. The objective function is further decomposed into two sub-problems, and the proposed greedy vehicle user selection algorithm and adaptive particle swarm optimization algorithm are used to solve and optimize the parameters of the objective function. In each round of federated communication, the base station obtains the selected set of vehicle users and schedules the position of the intelligent mobile antenna to minimize the base station's compensation cost. This method optimizes the base station compensation cost in the mobile vehicle network by modeling the base station's compensation cost. Furthermore, by scheduling the intelligent mobile antenna position, the method improves the communication rate between vehicle users and the base station and reduces communication latency.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle networking, and in particular relates to a method and system for optimizing the cost of antenna position in a vehicle networking system based on movable antennas. Background Art

[0002] The optimization of base station compensation costs in the Internet of Vehicles (IoV) is a hot research topic. The heterogeneous resources of vehicle users, such as computing and communication resources, their mobility, their sensitivity to local data, the energy consumed by vehicle users in updating local models, the energy consumed by vehicle users in adding noise to the updated local models, and the energy consumed by vehicle users in uploading the noisy local models all have a certain degree of influence on the base station's decision to select vehicle users. Furthermore, intelligent mobile antennas deployed at base stations can improve the communication environment and reduce communication latency by optimizing their positions. Due to the numerous influencing factors, it is difficult to derive the optimal method through theoretical derivation. Therefore, it is necessary to propose a base station's optimal vehicle user selection strategy and a method for scheduling intelligent mobile antennas.

[0003] Compared with the technology of application document CN118333191A "A vehicle selection and resource optimization method and system for federated learning in Internet of Vehicles", there are the following differences:

[0004] 1. Model construction and objective function;

[0005] Application document CN118333191A establishes the optimization objectives of the federated learning system based solely on vehicle computing power, transmission capacity, and fairness constraints, and transforms the time-coupled offline problem into an online problem through Lyapunov optimization. The system also uses linear regression to predict vehicle inherent parameters, and jointly optimizes resource allocation and vehicle selection based on reinforcement learning to minimize overall system costs and improve model accuracy and training efficiency.

[0006] This application, within the framework of privacy-preserving federated learning, introduces four components: communication cost, computational cost, privacy cost, and vehicle mobility model. Based on the compensation requirements of base stations for vehicle users within signal coverage, it constructs an objective function that minimizes base station compensation costs. Furthermore, we introduce intelligent mobile antenna scheduling and incorporate communication rate and latency into the modeling, achieving dual optimization of base station compensation and user experience.

[0007] 2. Algorithm design and solution method;

[0008] In application document CN118333191A, linear regression and reinforcement learning are used for parameter prediction and online resource scheduling, respectively. The algorithm has a high degree of coupling but requires the training of a large number of strategy models, and does not consider the physical layer gain of the time-varying communication link.

[0009] This application splits the objective function into two sub-problems: vehicle user selection and antenna position scheduling. For the former, a greedy vehicle user selection algorithm is designed to quickly screen high-value participants. For the latter, an adaptive particle swarm optimization (PSO) algorithm is used to dynamically adjust antenna positions to maximize communication quality and minimize compensation costs. This combined approach balances computational efficiency and convergence performance, and synchronizes the user set and antenna layout in real time through each round of federated communication.

[0010] Compared with the technology of application document CN114051222B "A wireless resource allocation and communication optimization method based on federated learning in a connected vehicle environment", there are the following differences:

[0011] 1. System model and optimization objectives;

[0012] Application document CN114051222B constructs a wireless resource allocation and communication optimization model based on the interaction between base stations and on-board equipment. It first uses the Jonker–Volgenant algorithm to optimally match channel status and data volume, then selects high-performance vehicles through a traversal algorithm, and finally uses stochastic gradient quantization to compress uplink model parameters to save bandwidth and improve convergence accuracy.

[0013] In addition to the above-mentioned communication and resource framework, this application introduces the modeling dimensions of privacy cost and compensation cost, and incorporates intelligent movable antennas into the overall architecture: in each round of federated communication, the base station not only selects participating vehicles, but also schedules the antenna position in real time, minimizing compensation overhead and communication delay in a multi-objective manner, thereby improving system flexibility and user experience.

[0014] 2. Algorithm process and performance advantages;

[0015] The method in application document CN114051222B sequentially performs matching → user screening → quantization compression. Although it optimizes resource allocation and bandwidth utilization, it does not jointly consider user incentives and dynamic changes in the communication environment.

[0016] This application adopts a "divide and conquer" approach to collaboratively optimize the minimization of compensation costs and the maximization of communication rates: the greedy algorithm ensures low complexity in vehicle selection, while the adaptive PSO efficiently converges to the optimal antenna layout in continuous space; the two operate in a coupled manner to achieve unified control of user incentives, communication quality, and latency. Summary of the Invention

[0017] To address the above issues, the present invention provides a method and system for optimizing the cost of antenna position in a vehicle network based on a movable antenna. By selecting vehicle users and optimizing the position of intelligent movable antennas, the base station compensation cost in the vehicle network is optimized, the transmission rate of vehicle users is increased, and the transmission delay is reduced.

[0018] In order to achieve the above-mentioned object of the invention, a method for optimizing the cost of antenna location in a mobile antenna vehicle network is provided, the method comprising the following steps:

[0019] S1: Build a privacy-preserving federated learning architecture based on intelligent mobile antennas. The total cost of vehicle users is composed of three parts: communication cost, computing cost, and privacy cost. The base station provides corresponding compensation and determines the number of vehicle users within the base station signal coverage range based on the vehicle mobility model. The objective function of minimizing the total cost of the base station is established.

[0020] S2: Decompose the target problem into two sub-problems and solve them using the greedy vehicle user selection algorithm and the adaptive particle swarm optimization algorithm respectively, to obtain the vehicle user set selected by the base station and the position of the intelligent movable antenna;

[0021] S3: Obtain the set of vehicle users selected by the federal communication base station in each round and the position of the antenna after scheduling to calculate the compensation cost of the base station in the current round, and output the optimized base station compensation cost.

[0022] As a further improvement of the present invention, the objective function of minimizing the total cost of the base station established in step S1 includes:

[0023] The objective function for minimizing the total cost of a base station consists of three parts: communication cost, computational cost, and privacy cost. The optimization problem is as follows:

[0024] ;

[0025] ;

[0026] in, represents the vehicle user selection set, represents a collection of intelligent movable antennas; Shows the total set of vehicle users; Indicates the vehicle user, Indicates whether the vehicle user is selected; represents the total compensation of the base station to the vehicle user; Indicates the number of smart movable antennas; Respectively represent Hedi antennas, A represents the antenna system, which consists of a two-dimensional coordinate system. and Respectively represent Antenna and The position of the antenna in the two-dimensional coordinate system A, Indicates the minimum distance between two adjacent antennas; represents the total delay of vehicle users, is the maximum delay that the base station can tolerate; represents the distance traveled by the vehicle user, Indicates the base station signal coverage radius, represents the cosine angle between the base station coverage area and the road, Indicates the distance between the vehicle user and the entrance of the base station coverage area; The variance of the noise added by the vehicle user to the updated local model, Represents the total noise variance threshold of the base station for the selected vehicle user set.

[0027] As a further improvement of the present invention, the communication cost of establishing the objective function of minimizing the total cost of the base station is expressed as:

[0028] ;

[0029] is the transmission rate of vehicle users, is the transmission power of the vehicle user, is the mixed vector received by the base station for the vehicle user, is the Gaussian white noise during channel transmission; is the transmission delay for vehicle users, The amount of data uploaded for vehicle users; Transmission energy consumption for vehicle users; is the transmission cost for vehicle users, is the unit cost of transmission energy consumption.

[0030] As a further improvement of the present invention, the computational cost of establishing the objective function for minimizing the total cost of the base station is expressed as:

[0031] ;

[0032] The computation delay for vehicle users, The number of CPU cycles required to process a single sample for a vehicle user, is the size of the vehicle user's local dataset, CPU frequency calculated locally for vehicle users; Calculate energy consumption for vehicle users, CPU energy consumption for local computations for vehicle users; is the computational cost for vehicle users, Calculate the unit cost of energy consumption.

[0033] As a further improvement of the present invention, the privacy cost of establishing the objective function of minimizing the total cost of the base station is expressed as:

[0034] ;

[0035] represents the privacy cost of the vehicle user, is the sensitivity of two adjacent datasets in differential privacy, represents the unit cost of noise energy consumption added by vehicle users to the local model, represents the privacy budget of the vehicle user.

[0036] As a further improvement of the present invention, the vehicle mobility model in step S1 is expressed as:

[0037] ;

[0038] represents the distance traveled by the vehicle user, is the driving speed of the vehicle user, represents the transmission delay of the vehicle user, represents the computational delay of the vehicle user, Indicates the distance traveled by the vehicle user after uploading the updated local model. R Indicates the base station signal coverage radius, represents the cosine angle between the base station coverage area and the road, Indicates the distance between the vehicle user and the entrance of the base station coverage area, Indicates the distance between the vehicle user and the base station.

[0039] As a further improvement of the present invention, the greedy vehicle user selection algorithm in step S2, which decomposes the target problem into two sub-problems, is expressed as:

[0040] ;

[0041] ;

[0042] in, represents the vehicle user selection set; V represents the total vehicle user set; Indicates the vehicle user, Indicates whether the vehicle user is selected; represents the total compensation of the base station to the vehicle user; Z Indicates the number of smart movable antennas; represents the total delay of vehicle users, is the maximum delay that the base station can tolerate; represents the distance traveled by the vehicle user, RIndicates the base station signal coverage radius, represents the cosine angle between the base station coverage area and the road, Indicates the distance between the vehicle user and the entrance of the base station coverage area; The variance of the noise added by the vehicle user to the updated local model, Represents the total noise variance threshold of the base station for the selected vehicle user set;

[0043] The solution is solved by greedy vehicle user selection algorithm, and the specific steps include:

[0044] First, the base station determines the set of vehicle users that meet the delay constraints and are within the base station signal coverage range. Then, all vehicle users in the set calculate their respective communication costs, computing costs, and privacy costs, and calculate the total cost before uploading it to the base station. The base station sorts the total costs uploaded by the vehicle users in the set from low to high. The base station starts selecting the vehicle users with the lowest total cost until the noise sum threshold is met. Finally, the base station obtains the vehicle user subset selected in the current round.

[0045] As a further improvement of the present invention, the adaptive particle swarm optimization algorithm is expressed as follows in step S2 where the target problem is decomposed into two sub-problems:

[0046] ;

[0047] ;

[0048] in, represents a collection of intelligent movable antennas; C represents the vehicle-user subset obtained by the greedy vehicle-user selection algorithm; Indicates the vehicle user, represents the total compensation of the base station to the vehicle user; Indicates the number of smart movable antennas; Respectively represent Hedi antennas, A represents the antenna system, which consists of a two-dimensional coordinate system. and Respectively represent Antenna and The position of the antenna in the two-dimensional coordinate system A, Indicates the minimum distance between two adjacent antennas;

[0049] The subproblem is solved using the adaptive particle swarm optimization algorithm. The specific steps are as follows:

[0050] The first step is to initialize the particle's velocity and position:

[0051] ;

[0052] in represents the initial position set of the particle swarm, Represents particles The initial position of represents the initial velocity set of the particle swarm, Represents particles The initial velocity is I particles, of which Further expressed as:

[0053] ;

[0054] in 、 Represents particles The stored The initial position of each antenna in the coordinate system;

[0055] The feasible area of ​​the smart movable antenna is represented as A, which satisfies , so the positions of particles and antennas need to satisfy the following:

[0056] ;

[0057] in Represents particles The stored Antennas in the coordinate system The position of the axis, Represents particles The stored Antennas in the coordinate system The position of the axis, H represents the radius of the antenna system;

[0058] The second step is to update the particle velocity and position:

[0059] During the adaptive particle swarm optimization algorithm, the particle swarm will update its position and velocity in each iteration, which are expressed as follows:

[0060] ;

[0061] in Indicates the current iteration number of the particle swarm algorithm, represents the global optimal position of the individual particle swarm, represents the global optimal position of the particle swarm, Indicates the Wheel particles speed, Indicates the Wheel-time particles location, Indicates the Wheel-time particles location, is the inertia factor of the particle, and It is The learning factor of the wheel particle, and Is a random number that satisfies ;

[0062] The third step is to adaptively update the inertia factor:

[0063] Currently, the more popular way to update the inertia factor is to adaptively update the inertia factor of each iteration, as shown below:

[0064] ;

[0065] in is the inertia factor of the particle, Indicates the maximum inertia weight of the particle, represents the minimum inertia weight of the particle, Indicates the current iteration number of the particle swarm algorithm, M Represents the total number of iterations of the adaptive particle swarm algorithm;

[0066] The fourth step is to adaptively update the learning factor:

[0067] The learning factor is updated based on the distance between the current particle and the global optimal position of its individual particle group. And the global optimal position of the particle swarm The particles farther away from the two optimal positions should obtain a larger learning factor so that they can approach their optimal positions faster. The adaptive weighted update rule is expressed as follows:

[0068] ;

[0069] in and It is The learning factor of the wheel particle, function F represents the adaptive weighted update function, and Represents the particle The distance of the wheel from its individual optimal position and the distance from the group optimal position, specifically, and It is expressed as follows:

[0070] ;

[0071] A piecewise linear function is used as the adaptive weighted update function, which is specifically expressed as follows:

[0072] ;

[0073] in and Represents the particle The distance of the wheel from its individual optimal position and the distance from the group optimal position, represents the global optimal position of the individual particle swarm, represents the global optimal position of the particle swarm, Indicates the Wheel-time particles location, and is a parameter that controls the learning factor and satisfies ; represents the distance between the particle and the optimal position, Indicates the threshold of the current particle distance from the optimal position;

[0074] The fifth step is to set the penalty function and fitness function:

[0075] Considering the constraints of this subproblem, a penalty function is set to punish particles that violate the constraints. The penalty function The specific expressions are as follows:

[0076] ;

[0077] in is the penalty factor, represents the penalty function of the particle, is the number of particles violating the constraint;

[0078] Combined with the penalty function, the particle's fitness function It is expressed as follows:

[0079] ;

[0080] in represents the fitness function of the particle, C represents the vehicle-user subset obtained by the greedy vehicle-user selection algorithm; Indicates the vehicle user; represents the total compensation of the base station to the vehicle user; represents the penalty function.

[0081] The present invention provides an optimization system for a mobile antenna vehicle network including antenna location costs, the system comprising:

[0082] A construction module for constructing a connected vehicle model based on a privacy-preserving federated learning architecture assisted by intelligent movable antennas, and establishing an objective function for minimizing base station compensation costs based on historical status data of vehicle users, including the vehicle user's communication cost, local computing cost, privacy cost, and the distance between the vehicle user and the base station;

[0083] A vehicle user selection module is used to obtain vehicle users selected by the base station, solve the objective function and optimize the parameters using a greedy vehicle user selection algorithm to obtain a vehicle user selection model, and use the result in the antenna position scheduling module;

[0084] The antenna position optimization module is used to obtain the current optimized antenna position, and adopts the adaptive particle swarm optimization algorithm and the results of the vehicle user selection module to solve the objective function and optimize the parameters to obtain the optimized antenna position.

[0085] Beneficial effects: The present invention constructs a vehicle network model based on a privacy-preserving federated learning architecture with intelligent movable antennas. According to the vehicle user's communication cost, local computing cost, privacy cost and mobility, an objective function of minimizing the base station compensation cost is established. By selecting vehicle users and optimizing the position of intelligent movable antennas, the base station compensation cost in the vehicle network is optimized, and the transmission rate of vehicle users is improved and the transmission delay is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 This is a flow chart of the cost and antenna position optimization method in the Internet of Vehicles based on the privacy-preserving federated learning architecture with smart movable antennas;

[0087] Figure 2 This is a schematic diagram of the cost and antenna position optimization system in the Internet of Vehicles in a privacy-preserving federated learning architecture based on smart movable antennas;

[0088] Figure 3 This is a comparison chart about base station compensation costs;

[0089] Figure 4 This is a comparison chart of adaptive particle swarm optimization algorithms. DETAILED DESCRIPTION

[0090] The following is a clear and complete description of the technical solutions in the embodiments of the present invention, in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts are within the scope of protection of the present invention.

[0091] Example 1;

[0092] Figure 1 This is a flow chart of the cost and antenna position optimization method in the Internet of Vehicles based on the privacy-preserving federated learning architecture with intelligent movable antennas. Figure 1 As shown, this embodiment provides a method for optimizing cost and antenna position in an Internet of Vehicles (IoV) in a privacy-preserving federated learning architecture based on smart movable antennas, the method comprising the following steps:

[0093] S1: Build a privacy-preserving federated learning architecture based on intelligent mobile antennas. The base station provides corresponding compensation based on the total cost of vehicle users within the base station signal coverage area, and establishes an objective function to minimize the total cost of the base station.

[0094] S2: Decompose the target problem into two sub-problems and solve them using the greedy vehicle user selection algorithm and the adaptive particle swarm optimization algorithm respectively, to obtain the vehicle user set selected by the base station and the position of the intelligent movable antenna;

[0095] S3: Obtain the set of vehicle users selected by the federal communication base station in each round and the position of the antenna after scheduling to calculate the compensation cost of the base station in the current round, and output the optimized base station compensation cost.

[0096] The method of this embodiment minimizes the compensation cost of the Internet of Vehicles base station by optimizing the base station compensation cost and antenna position, and improves the transmission rate of vehicle users and reduces the transmission delay.

[0097] In step S1, establishing an objective function for minimizing base station compensation cost includes:

[0098] The objective function of base station compensation cost is established, and the optimization problem is as follows:

[0099] ;

[0100] ;

[0101] in, represents the vehicle user selection set, represents a collection of intelligent movable antennas; Shows the total set of vehicle users; Indicates the vehicle user, Indicates whether the vehicle user is selected; represents the total compensation of the base station to the vehicle user; Indicates the number of smart movable antennas; Respectively represent Hedi antennas, A represents the antenna system, which consists of a two-dimensional coordinate system. and Respectively represent Antenna and The position of the antenna in the two-dimensional coordinate system A, Indicates the minimum distance between two adjacent antennas; represents the total delay of vehicle users, is the maximum delay that the base station can tolerate; represents the distance traveled by the vehicle user, Indicates the base station signal coverage radius, represents the cosine angle between the base station coverage area and the road, Indicates the distance between the vehicle user and the entrance of the base station coverage area; The variance of the noise added by the vehicle user to the updated local model, Represents the total noise variance threshold of the base station for the selected vehicle user set.

[0102] In step S1, under the IoV framework based on the intelligent movable antenna, the communication cost of the vehicle user can be expressed as:

[0103] ;

[0104] is the transmission rate of vehicle users, is the transmission power of the vehicle user, is the mixed vector received by the base station for the vehicle user, is the Gaussian white noise during channel transmission; is the transmission delay for vehicle users, The amount of data uploaded for vehicle users; Transmission energy consumption for vehicle users; is the transmission cost for vehicle users, is the unit cost of transmission energy consumption.

[0105] Specifically, a system is constructed including a base station, an edge server deployed in the base station, and an intelligent movable antenna system, and N vehicle users.

[0106] Transmission delay of vehicle users Expressed as:

[0107] ;

[0108] in, The amount of data uploaded by vehicle users, Transmission rate for vehicle users;

[0109] Required transmission energy consumption for: ;

[0110] in, The transmission power for vehicle users.

[0111] The transmission cost caused by vehicle users is expressed as: ;

[0112] in, is the unit cost of transmission energy consumption.

[0113] In step S1, under the IoV framework based on the intelligent movable antenna, the computation cost of the vehicle user can be expressed as:

[0114] ;

[0115] The computation delay for vehicle users, The number of CPU cycles required to process a single sample for a vehicle user, is the size of the vehicle user's local dataset, CPU frequency calculated locally for vehicle users; Calculate energy consumption for vehicle users, CPU energy consumption for local computations for vehicle users; is the computational cost for vehicle users, Calculate the unit cost of energy consumption.

[0116] Specifically, the vehicle user's local computing delay for:

[0117] ;

[0118] in, The number of CPU cycles required to process a single sample for a vehicle user, is the size of the vehicle user's local dataset, CPU frequency calculated locally for vehicle users;

[0119] Required computing energy for: ;

[0120] in, Calculate power for vehicle users.

[0121] The computational cost caused by vehicle users is expressed as: ;

[0122] in, Calculate the unit cost of energy consumption.

[0123] In step S1, the privacy cost model of the vehicle user can be expressed as:

[0124] ;

[0125] represents the privacy cost of the vehicle user, is the sensitivity of two adjacent datasets in differential privacy, represents the unit cost of noise energy consumption added by vehicle users to the local model, represents the privacy budget of the vehicle user.

[0126] Specifically, the sum of the three costs for vehicle users is expressed as

[0127] ;

[0128] In step S2, the objective function is solved using a greedy vehicle user selection algorithm and an adaptive particle swarm optimization algorithm. Specifically, the greedy vehicle user selection algorithm is first used to determine the vehicle user set selected by the base station. Then, based on the determined vehicle user set, the adaptive particle swarm optimization algorithm is used to solve the antenna position. The specific process is as follows:

[0129] 1. The first sub-problem after the target problem is decomposed is expressed as:

[0130] ;

[0131] ;

[0132] in, represents the vehicle user selection set; V represents the total vehicle user set; Indicates the vehicle user, Indicates whether the vehicle user is selected; represents the total compensation of the base station to the vehicle user; Z Indicates the number of smart movable antennas; represents the total delay of vehicle users, is the maximum delay that the base station can tolerate; represents the distance traveled by the vehicle user, R Indicates the base station signal coverage radius, represents the cosine angle between the base station coverage area and the road, Indicates the distance between the vehicle user and the entrance of the base station coverage area; The variance of the noise added by the vehicle user to the updated local model, Represents the total noise variance threshold of the base station for the selected vehicle user set;

[0133] The greedy vehicle user selection algorithm is used to solve this subproblem. The specific steps are as follows:

[0134] First, the base station determines the set of vehicle users that meet the delay constraints and are within the base station signal coverage range. Then, all vehicle users in the set calculate their respective communication costs, computing costs, and privacy costs, and calculate the total cost before uploading it to the base station. The base station sorts the total costs uploaded by the vehicle users in the set from low to high. The base station starts selecting the vehicle users with the lowest total cost until the noise sum threshold is met. Finally, the base station obtains the vehicle user subset selected in the current round.

[0135] 2. The second sub-problem after the target problem is decomposed is expressed as:

[0136] The solution is solved by greedy vehicle user selection algorithm, and the specific steps include:

[0137] First, the base station determines the set of vehicle users that meet the delay constraints and are within the base station signal coverage range. Then, all vehicle users in the set calculate their respective communication costs, computing costs, and privacy costs, and calculate the total cost before uploading it to the base station. The base station sorts the total costs uploaded by the vehicle users in the set from low to high. The base station starts selecting the vehicle users with the lowest total cost until the noise sum threshold is met. Finally, the base station obtains the vehicle user subset selected in the current round.

[0138] ;

[0139] ;

[0140] in, represents a collection of intelligent movable antennas; C represents the vehicle-user subset obtained by the greedy vehicle-user selection algorithm; Indicates the vehicle user, represents the total compensation of the base station to the vehicle user; Indicates the number of smart movable antennas; Respectively represent Hedi antennas, A represents the antenna system, which consists of a two-dimensional coordinate system. and Respectively represent Antenna and The position of the antenna in the two-dimensional coordinate system A, Indicates the minimum distance between two adjacent antennas;

[0141] Adaptive particle swarm optimization algorithm is used to solve this subproblem. The specific steps are as follows:

[0142] Initialize the particle's velocity and position:

[0143] ;

[0144] in represents the initial position set of the particle swarm, Represents particles The initial position of represents the initial velocity set of the particle swarm, Represents particles The initial velocity is I particles, of which Further expressed as:

[0145] ;

[0146] in 、 Represents particles The stored The initial position of each antenna in the coordinate system;

[0147] The feasible area of ​​the smart movable antenna is represented as A, which satisfies , so the positions of particles and antennas need to satisfy the following:

[0148] ;

[0149] in Represents particles The stored Antennas in the coordinate system The position of the axis, Represents particles The stored Antennas in the coordinate system The position of the axis, H represents the radius of the antenna system;

[0150] Particle velocity and position update:

[0151] During the adaptive particle swarm optimization algorithm, the particle swarm will update its position and velocity in each iteration, which are expressed as follows:

[0152] ;

[0153] in Indicates the current iteration number of the particle swarm algorithm, represents the global optimal position of the individual particle swarm, represents the global optimal position of the particle swarm, Indicates the Wheel particles speed, Indicates the Wheel-time particles location, Indicates the Wheel-time particles location, is the inertia factor of the particle, and It is The learning factor of the wheel particle, and Is a random number that satisfies ;

[0154] Adaptively update the inertia factor:

[0155] Currently, the more popular way to update the inertia factor is to adaptively update the inertia factor of each iteration, as shown below:

[0156] ;

[0157] in is the inertia factor of the particle, Indicates the maximum inertia weight of the particle, represents the minimum inertia weight of the particle, Indicates the current iteration number of the particle swarm algorithm, M Represents the total number of iterations of the adaptive particle swarm algorithm;

[0158] Adaptively update the learning factor:

[0159] Compared with the traditional particle swarm optimization algorithm, the adaptive particle swarm optimization algorithm has a faster convergence speed and better convergence performance. The learning factor is updated according to the distance between the current particle and its individual optimal position. And the global optimal position of the particle swarm The particles farther away from the two optimal positions should obtain a larger learning factor so that they can approach their optimal positions faster. The adaptive weighted update rule is expressed as follows:

[0160] ;

[0161] in and It is The learning factor of the wheel particle, function F represents the adaptive weighted update function, and Represents the particle The distance of the wheel from its individual optimal position and the distance from the group optimal position, specifically, and It is expressed as follows:

[0162] ;

[0163] A piecewise linear function is used as the adaptive weighted update function, which is specifically expressed as follows:

[0164] ;

[0165] in and Represents the particle The distance of the wheel from its individual optimal position and the distance from the group optimal position, represents the global optimal position of the individual particle swarm, represents the global optimal position of the particle swarm, Indicates the Wheel-time particles location, and is a parameter that controls the learning factor and satisfies ; represents the distance between the particle and the optimal position, Indicates the threshold of the current particle distance from the optimal position;

[0166] Set the penalty function and fitness function:

[0167] Considering the constraints of this subproblem, we set a penalty function to punish particles that violate the constraints. The penalty function The specific expressions are as follows:

[0168] ;

[0169] in is the penalty factor, represents the penalty function of the particle, is the number of particles violating the constraint;

[0170] Combined with the penalty function, the particle's fitness function It is expressed as follows:

[0171] ;

[0172] in represents the fitness function of the particle, C represents the vehicle-user subset obtained by the greedy vehicle-user selection algorithm; Indicates the vehicle user; represents the total compensation of the base station to the vehicle user; represents the penalty function.

[0173] Figure 3 This is a comparison chart of the base station compensation cost of the algorithm proposed in this patent, the random vehicle user selection algorithm, and the all vehicle user selection algorithm. Figure 4 This is a comparison chart of the adaptive particle swarm optimization algorithm proposed in this patent and the traditional particle swarm optimization algorithm. Figure 3 It can be seen from the above that the algorithm proposed in this patent has a lower compensation cost for base stations than other algorithms. Figure 4 It can be seen that the adaptive particle swarm optimization algorithm proposed in this patent has a faster convergence speed and better convergence performance than the traditional particle swarm optimization algorithm.

[0174] This embodiment constructs a vehicle network model based on a privacy-preserving federated learning architecture with intelligent movable antennas. It establishes an objective function for minimizing base station compensation costs based on vehicle user communication costs, local computing costs, privacy costs, and vehicle user mobility. By selecting vehicle users and optimizing the positions of intelligent movable antennas, it optimizes base station compensation costs in the vehicle network, improves the transmission rate of vehicle users, and reduces transmission latency.

[0175] Example 2:

[0176] Figure 2 This is a schematic diagram of the cost and antenna position optimization system in the Internet of Vehicles based on the privacy-preserving federated learning architecture with smart movable antennas. Figure 2 As shown, this embodiment provides a cost and antenna position optimization system in an Internet of Vehicles based on a privacy-preserving federated learning architecture with intelligent movable antennas, the system comprising:

[0177] Construction module 201 is configured to construct a connected vehicle model based on a privacy-preserving federated learning architecture with intelligent movable antennas, and establish an objective function for minimizing base station compensation costs based on historical status data of vehicle users, wherein the historical status data includes the vehicle user's communication cost, local computing cost, privacy cost, and the distance between the vehicle user and the base station;

[0178] The vehicle user selection module 202 is used to obtain the vehicle users selected by the base station, solve the objective function and optimize the parameters using a greedy vehicle user selection algorithm to obtain a vehicle user selection model, and use the result in the antenna position scheduling module;

[0179] The antenna position optimization module 203 is used to obtain the current optimized antenna position, and solve the objective function and optimize the parameters using the adaptive particle swarm optimization algorithm and the result of the vehicle user selection module to obtain the optimized antenna position.

[0180] Preferably, the building module 201 establishes an objective function for minimizing the base station compensation cost, including:

[0181] The objective function of base station compensation cost is established, and the optimization problem is as follows:

[0182] ;

[0183] ;

[0184] Preferably, a vehicle user selection module is constructed, and the objective function is modified as follows:

[0185] ;

[0186] ;

[0187] The optimization problem is solved by a greedy vehicle-user selection algorithm, and a vehicle-user selection model is obtained.

[0188] Preferably, an antenna position optimization module is constructed, and the objective function is modified as follows:

[0189] ;

[0190] ;

[0191] The optimization problem is solved by an adaptive particle swarm optimization algorithm, and the optimized antenna position model is obtained.

[0192] The specific implementation process of the functions realized by each module in this embodiment 2 is the same as the implementation process in embodiment 1, and will not be repeated here.

[0193] Experimental results show that the algorithm proposed in this paper exhibits excellent performance in terms of base station compensation cost and average transmission delay indicators.

Claims

1. An optimization method for mobile antenna vehicle network including antenna location cost, characterized by: The method comprises the following steps: S1: Build a connected vehicle model based on a privacy-preserving federated learning architecture with intelligent movable antennas. Based on the total cost of vehicle users, which consists of three parts: communication cost, computation cost, and privacy cost, the base station provides corresponding compensation. Based on the vehicle mobility model, the base station determines which vehicle users are within the base station's signal coverage range. By selecting vehicle users and optimizing the position of intelligent movable antennas, the base station compensation cost in the connected vehicle is optimized, and an objective function is established to minimize the total base station cost. The objective function of minimizing the total cost of the base station established in step S1 includes: The objective function for minimizing the total cost of a base station consists of three parts: communication cost, computational cost, and privacy cost. The optimization problem is as follows: ； ； in, represents the vehicle user selection set, represents a collection of intelligent movable antennas; Shows the total set of vehicle users; Indicates the vehicle user, Indicates whether the vehicle user is selected; represents the total compensation of the base station to the vehicle user; Indicates the number of smart movable antennas; Respectively represent Hedi antennas, A represents the antenna system, which consists of a two-dimensional coordinate system. and Respectively represent Antenna and The position of the antenna in the two-dimensional coordinate system A, Indicates the minimum distance between two adjacent antennas; represents the total delay of vehicle users, is the maximum delay that the base station can tolerate; represents the distance traveled by the vehicle user, Indicates the base station signal coverage radius, represents the cosine angle between the base station coverage area and the road, Indicates the distance between the vehicle user and the entrance of the base station coverage area; The variance of the noise added by the vehicle user to the updated local model, Represents the total noise variance threshold of the base station for the selected vehicle user set; The communication cost of establishing the objective function of minimizing the total cost of the base station is expressed as: ； is the transmission rate of vehicle users, is the transmission power of the vehicle user, is the mixed vector received by the base station for the vehicle user, is the Gaussian white noise during channel transmission; is the transmission delay for vehicle users, The amount of data uploaded for vehicle users; Transmission energy consumption for vehicle users; is the transmission cost for vehicle users, is the unit cost of transmission energy consumption; The computational cost of establishing the objective function of minimizing the total cost of the base station is expressed as: ； The computation delay for vehicle users, The number of CPU cycles required to process a single sample for a vehicle user, is the size of the vehicle user's local dataset, CPU frequency calculated locally for vehicle users; Calculate energy consumption for vehicle users, CPU energy consumption for local computations for vehicle users; is the computational cost for vehicle users, To calculate the unit cost of energy consumption; S2: Decompose the target problem into two sub-problems and solve them using the greedy vehicle user selection algorithm and the adaptive particle swarm optimization algorithm respectively, to obtain the vehicle user set selected by the base station and the position of the intelligent movable antenna; In step S2, the target problem is decomposed into two sub-problems, and the greedy vehicle user selection algorithm is expressed as: ； ； in, represents the vehicle user selection set; V represents the total vehicle user set; Indicates the vehicle user, Indicates whether the vehicle user is selected; represents the total compensation of the base station to the vehicle user; Z Indicates the number of smart movable antennas; represents the total delay of vehicle users, is the maximum delay that the base station can tolerate; represents the distance traveled by the vehicle user, R Indicates the base station signal coverage radius, represents the cosine angle between the base station coverage area and the road, Indicates the distance between the vehicle user and the entrance of the base station coverage area; The variance of the noise added by the vehicle user to the updated local model, Represents the total noise variance threshold of the base station for the selected vehicle user set; The solution is solved by greedy vehicle user selection algorithm, and the specific steps include: First, the base station determines the set of vehicle users that meet the delay constraint and are within the base station's signal coverage. Then, all vehicle users in the set calculate their communication cost, computing cost, and privacy cost, and calculate the total cost before uploading it to the base station. The base station sorts the total cost uploaded by the vehicle users in the set from low to high, and starts selecting the vehicle user with the lowest total cost until the noise sum threshold is met. Finally, the base station obtains the vehicle user subset selected in the current round. In step S2, the target problem is decomposed into two sub-problems, and the adaptive particle swarm optimization algorithm is expressed as: ； ； in, represents a collection of intelligent movable antennas; C represents the vehicle-user subset obtained by the greedy vehicle-user selection algorithm; Indicates the vehicle user, represents the total compensation of the base station to the vehicle user; Indicates the number of smart movable antennas; Respectively represent Hedi antennas, A represents the antenna system, which consists of a two-dimensional coordinate system. and Respectively represent Antenna and The position of the antenna in the two-dimensional coordinate system A, Indicates the minimum distance between two adjacent antennas; The subproblem is solved using the adaptive particle swarm optimization algorithm. The specific steps are as follows: The first step is to initialize the particle's velocity and position: ； in represents the initial position set of the particle swarm, Represents particles The initial position of represents the initial velocity set of the particle swarm, Represents particles The initial velocity is I particles, of which Further expressed as: ； in 、 Represents particles The stored The initial position of each antenna in the coordinate system; The feasible area of ​​the smart movable antenna is represented as A, which satisfies , so the positions of particles and antennas need to satisfy the following: ； in Represents particles The stored Antennas in the coordinate system The position of the axis, Represents particles The stored Antennas in the coordinate system The position of the axis, H represents the radius of the antenna system; The second step is to update the particle velocity and position: During the adaptive particle swarm optimization algorithm, the particle swarm will update its position and velocity in each iteration, which are expressed as follows: ； in Indicates the current iteration number of the particle swarm algorithm, represents the global optimal position of the individual particle swarm, represents the global optimal position of the particle swarm, Indicates the Wheel particles speed, Indicates the Wheel-time particles location, Indicates the Wheel-time particles location, is the inertia factor of the particle, and It is The learning factor of the wheel particle, and Is a random number that satisfies ; The third step is to adaptively update the inertia factor: Currently, the more popular way to update the inertia factor is to adaptively update the inertia factor of each iteration, as shown below: ； in is the inertia factor of the particle, Indicates the maximum inertia weight of the particle, represents the minimum inertia weight of the particle, Indicates the current iteration number of the particle swarm algorithm, M Represents the total number of iterations of the adaptive particle swarm algorithm; The fourth step is to adaptively update the learning factor: The learning factor is updated based on the distance between the current particle and the global optimal position of its individual particle group. And the global optimal position of the particle swarm The particles farther away from the two optimal positions should obtain a larger learning factor so that they can approach their optimal positions faster. The adaptive weighted update rule is expressed as follows: ； in and It is The learning factor of the wheel particle, function F represents the adaptive weighted update function, and Represents the particle The distance of the wheel from its individual optimal position and the distance from the group optimal position, specifically, and It is expressed as follows: ； A piecewise linear function is used as the adaptive weighted update function, which is specifically expressed as follows: ； in and Represents the particle The distance of the wheel from its individual optimal position and the distance from the group optimal position, represents the global optimal position of the individual particle swarm, represents the global optimal position of the particle swarm, Indicates the Wheel-time particles location, and is a parameter that controls the learning factor and satisfies ; represents the distance between the particle and the optimal position, Indicates the threshold of the current particle distance from the optimal position; The fifth step is to set the penalty function and fitness function: Considering the constraints of this subproblem, a penalty function is set to punish particles that violate the constraints. The penalty function The specific expressions are as follows: ； in is the penalty factor, represents the penalty function of the particle, is the number of particles violating the constraint; Combined with the penalty function, the particle's fitness function It is expressed as follows: ； in represents the fitness function of the particle, C represents the vehicle-user subset obtained by the greedy vehicle-user selection algorithm; represents a vehicle user; represents the total compensation of the base station to the vehicle user; represents the penalty function; S3: Obtain the set of vehicle users selected by the federal communication base station in each round and the position of the antenna after scheduling to calculate the compensation cost of the base station in the current round, and output the optimized base station compensation cost.

2. The optimization method for the movable antenna vehicle network including antenna location cost according to claim 1, characterized in that: The privacy cost of establishing the objective function of minimizing the total cost of the base station is expressed as: ； represents the privacy cost of the vehicle user, is the sensitivity of two adjacent datasets in differential privacy, represents the unit cost of noise energy consumption added by vehicle users to the local model, represents the privacy budget of the vehicle user.

3. The optimization method for the movable antenna vehicle network including antenna location cost according to claim 1, characterized in that: The vehicle mobility model in step S1 is expressed as: ； represents the distance traveled by the vehicle user, is the driving speed of the vehicle user, represents the transmission delay of the vehicle user, represents the computational delay of the vehicle user, Indicates the distance traveled by the vehicle user after uploading the updated local model. R Indicates the base station signal coverage radius, represents the cosine angle between the base station coverage area and the road, Indicates the distance between the vehicle user and the entrance of the base station coverage area, Indicates the distance between the vehicle user and the base station.

4. A system using the method for optimizing the cost of antenna location based on a mobile antenna vehicle network according to any one of claims 1 to 3, characterized in that: The system comprises: A construction module for constructing a connected vehicle model based on a privacy-preserving federated learning architecture with intelligent movable antennas, and establishing an objective function for minimizing base station compensation costs based on historical status data of vehicle users, including the vehicle user's communication cost, local computing cost, privacy cost, and distance between the vehicle user and the base station; A vehicle user selection module is used to obtain vehicle users selected by the base station, solve the objective function and optimize the parameters using a greedy vehicle user selection algorithm to obtain a vehicle user selection model, and use the result in the antenna position scheduling module; The antenna position optimization module is used to obtain the current optimized antenna position, and adopts the adaptive particle swarm optimization algorithm and the results of the vehicle user selection module to solve the objective function and optimize the parameters to obtain the optimized antenna position.

Images