Hidden air computing method, system and equipment based on movable antenna

Abstract

The invention discloses a concealed air computing method, system and equipment based on a movable antenna, and relates to the technical field of concealed air computing. The method is applied to a hidden air computing system comprising a base station, a monitor and a plurality of sensors of Internet of Things equipment. The sensor is provided with a single movable antenna; the method comprises the following steps: firstly, respectively establishing link channel models from each sensor to a base station and from each sensor to a monitor by taking the position of a movable antenna as a variable, further determining mean square error representation of the base station and error detection probability lower bound representation of the monitor, and then, determining error detection probability lower bound representation of the monitor by taking the position of the movable antenna as an optimization variable. And constructing an antenna position optimization model and solving by taking the maximum transmitting power of each sensor, the moving range of the movable antenna and the covert communication constraint as constraints and minimizing the mean square error of the base station as an optimization target, so that each movable antenna performs data transmission at the optimal position. According to the method, the cost, energy consumption and errors of hidden air calculation are reduced.

Description

Technical Field

[0001] The present invention relates to the field of concealed aerial computing technology, and in particular to a concealed aerial computing method, system and device based on a movable antenna. Background Art

[0002] At present, in the large-scale Internet of Things (IoT) scenario, the data fusion center needs to receive and calculate the massive amount of information generated by massive IoT devices. Traditional technology separates information reception and information calculation, resulting in problems such as low spectrum utilization and increased latency. AirComp technology can overcome the above shortcomings. AirComp technology uses the superposition characteristics of wireless multiple access channels to achieve simultaneous information reception and information calculation. However, due to the broadcast characteristics of wireless channels, the security of airComp is difficult to guarantee. Traditional network layer encryption technology and physical layer security technology are limited to the security of information content, ignoring the protection of the transmitter's own privacy (such as the transmitter's location, transmission mode, etc.). For this reason, covert communication technology (also known as low-probability monitoring communication technology) was proposed. The goal of covert communication technology is to ensure that the entire information transmission process is not discovered by malicious users, thereby avoiding further destructive behavior by malicious users.

[0003] In the prior art, research on covert aerial computing systems is usually based on discrete fixed arrangements of multiple antennas on each IoT device.

[0004] However, when multiple antennas are discretely arranged at fixed positions, on the one hand, each antenna must be equipped with a radio frequency unit, resulting in high hardware costs and huge energy consumption. On the other hand, all antennas are discretely fixed on the transceiver at intervals of no less than 1 / 2 wavelength, while the wireless channel fluctuates continuously in spatial distance. Fixed antennas cannot adapt to the continuous fluctuations of the channel in space, which limits the utilization of its freedom in the continuous space of the channel, thereby affecting the system performance and causing large errors in covert aerial calculations. Summary of the invention

[0005] Based on this, it is necessary to provide a covert aerial computing method, system and device based on a movable antenna to address the above technical problems.

[0006] The present invention adopts the following technical solutions:

[0007] The present invention provides a concealed aerial computing method based on a movable antenna, which is applied to a concealed aerial computing system of sensors including a base station, a monitor and a plurality of Internet of Things devices; the base station is arranged with a plurality of fixed antennas, and the monitor is arranged with a single fixed antenna; the sensor is arranged with a single movable antenna and information is uploaded to the base station through the movable antenna; the movable antenna is movable relative to the sensor and is connected to a radio frequency unit through a flexible cable; the present invention first takes the position of the movable antenna as a variable, respectively establishes link channel models from each sensor to the base station and from each sensor to the monitor, and then determines the mean square error representation of the base station and the lower bound representation of the false detection probability of the monitor, then takes the position of the movable antenna as an optimization variable, takes the maximum transmission power of each sensor, the moving range of the movable antenna and the concealed communication constraint as constraints, takes minimizing the mean square error of the base station as an optimization goal, constructs an antenna position optimization model and solves it, so that each movable antenna can upload and fuse the information collected by each sensor to the base station through concealed aerial computing at the optimal position relative to the sensor.

[0008] The present invention provides a concealed aerial computing system based on a movable antenna, comprising:

[0009] A concealed aerial computing system of a base station, a monitor and sensors of multiple IoT devices; the base station is provided with multiple fixed antennas, the monitor is provided with a single fixed antenna; each of the sensors is provided with a single movable antenna and uploads information to the base station via the movable antenna; the movable antenna is movable relative to the sensor and is connected to a radio frequency unit via a flexible cable;

[0010] A calculation module is used to establish link channel models from each sensor to a base station and from each sensor to a monitor respectively, taking the position of a movable antenna as a variable; determining a mean square error representation of the base station according to the link channel model from each sensor to the base station; determining a lower bound representation of the false detection probability of the monitor according to the link channel model from each sensor to the monitor, and determining a covert communication constraint according to the lower bound representation of the false detection probability of the monitor; constructing an antenna position optimization model with the position of the movable antenna as an optimization variable, the maximum transmission power of each sensor, the moving range of the movable antenna and the covert communication constraint as constraints, and minimizing the mean square error of the base station as an optimization goal; solving the antenna position optimization model to determine the optimal position of each movable antenna relative to the sensor;

[0011] The antennas of the sensors of each IoT device are moved to the optimal position obtained by the calculation module, and the information collected by each sensor is uploaded and integrated to the base station through hidden aerial calculation.

[0012] The present invention provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the concealed aerial computing method based on a movable antenna is implemented.

[0013] The present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the above-mentioned concealed aerial computing method based on a movable antenna is implemented.

[0014] At least one of the above technical solutions adopted by the present invention can achieve the following beneficial effects:

[0015] First, the sensor of the IoT device is equipped with a movable antenna, which is connected to the RF unit through a flexible cable. Based on the position of the movable antenna as the optimization variable, the maximum transmission power of each sensor, the moving range of the movable antenna and the covert communication constraint as the constraints, and the minimization of the mean square error of the base station as the optimization goal, the antenna position optimization model is constructed and solved, so as to determine the optimal position of the movable antenna relative to the sensor, so as to drive the antenna to move to the optimal position in the two-dimensional plane, improve the degree of freedom of the channel in the continuous space, effectively avoid the position where the channel state is in deep fading, further improve the system performance, and reduce the system error. At the same time, only one RF unit is required, which reduces the cost and energy consumption. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0017] Figure 1 A schematic diagram of a flow chart of a concealed aerial computing method based on a movable antenna provided by the present invention;

[0018] Figure 2 A schematic diagram of a concealed aerial computing system provided by the present invention;

[0019] Figure 3 A schematic diagram of a process for constructing and solving an antenna position optimization model provided by the present invention;

[0020] Figure 4 A schematic diagram of the comparison of mean square error between the method of the present invention and the antenna position fixed algorithm and the antenna position random algorithm provided by the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0022] At present, the research on covert aerial computing systems assumes that each IoT device is equipped with multiple fixed antennas. The discrete arrangement of multiple antennas in fixed positions, on the one hand, results in high hardware costs and huge energy consumption, and on the other hand, it also makes it impossible to fully utilize the degrees of freedom in continuous space, and the error of covert aerial computing is large. The present invention proposes a joint optimization method for the transceiver beam and antenna position of a covert aerial computing system based on a movable antenna. The core idea is to use a single movable antenna to replace multiple fixed antennas, so as to reduce hardware costs, reduce energy consumption, increase spatial freedom and reduce the error of covert aerial computing.

[0023] The technical solutions provided by various embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0024] Figure 1 The present invention is a flow chart of a concealed aerial computing method based on a movable antenna, which specifically includes the following steps:

[0025] S101: Taking the position of the movable antenna as a variable, respectively establish link channel models from each sensor to the base station and from each sensor to the monitor.

[0026] S102: Determine the mean square error representation of the base station according to the link channel model from each sensor to the base station; determine the lower bound representation of the false detection probability of the monitor according to the link channel model from each sensor to the monitor, and determine the covert communication constraint according to the lower bound representation of the false detection probability of the monitor.

[0027] S103: Taking the position of the movable antenna as the optimization variable, taking the maximum transmission power of each sensor, the moving range of the movable antenna and the covert communication constraint as the constraints, and taking minimizing the mean square error of the base station as the optimization goal, an antenna position optimization model is constructed.

[0028] S104: Solve the antenna position optimization model to determine the optimal position of each movable antenna relative to the sensor, so that each movable antenna can upload and fuse the information collected by each sensor to the base station through hidden aerial calculation at the optimal position.

[0029] For the convenience of explanation, the following description is only based on the server as the execution subject. The server mentioned in the present invention can be a server set up on the business platform, or a device such as a desktop computer, a notebook computer, etc. that can execute the solution of the present invention.

[0030] Generally, when base stations and IoT devices transmit and fuse data information based on covert aerial computing, optimization problems can be constructed and solved based on the covert aerial computing system model, so as to determine and apply the working parameters of each part of the covert aerial computing system.

[0031] Based on this, in one or more embodiments of the present invention, the covert aerial computing method based on a movable antenna provided by the present invention can be applied to a covert aerial computing system including a base station, a monitor, and sensors of multiple IoT devices. Figure 2 As shown, Figure 2 This is a schematic diagram of a concealed aerial computing system in the present invention.

[0032] The base station is equipped with multiple fixed antennas, the monitor is equipped with a single fixed antenna, each sensor is equipped with a single movable antenna and uploads information to the base station through the movable antenna. The movable antenna is movable relative to the sensor and is connected to the radio frequency unit through a flexible cable. The base station uses air computing technology to collect and aggregate the information uploaded by each sensor. The monitor attempts to determine whether the sensor of each IoT device sends information based on its received signal. In order to conceal information, each sensor can send information when the movable antenna is in the optimal antenna position.

[0033] When constructing the optimization problem, the position of the movable antenna can be used as a variable to establish the link channel model from each sensor to the base station and from each sensor to the monitor. Specifically, in one or more embodiments of the present invention, the position of the movable antenna can be used as a variable to establish the link channel model from each sensor to the base station through the following formula:

[0034] Taking the position of the movable antenna as a variable, the link channel model from each sensor to the monitor is established by the following formula:

[0035] Where, t k =[x k ,y k ] T is the movable antenna position of the kth sensor, [x k ,y k ] T is the position coordinate of the movable antenna of the kth sensor in the local two-dimensional rectangular coordinate system, and the subscript T represents transposition. is the receiving field response matrix of the base station, is the location of N receiving antennas of the base station, [x n ,y n ] T is the position coordinate of the nth receiving antenna of the base station in the local two-dimensional rectangular coordinate system, is the receiving field response vector of the nth receiving antenna of the base station, j is an imaginary number, j 2 =-1, L is the number of channel paths, λ is the wavelength, ∑ k is the path response matrix from the kth sensor coordinate origin to the base station coordinate origin, h k (t k ) is the link channel model from the kth sensor to the base station, is the receiving field response vector of the monitor, r w =[x w ,y w ] T is the receiving antenna position of the monitor, [x w ,y w ] T is the position coordinate of the monitor’s receiving antenna in the local two-dimensional rectangular coordinate system, is the path response matrix from the kth sensor coordinate origin to the monitor coordinate origin, is the link channel model from the kth sensor to the monitor.

[0036] is the signal distance difference between the nth receiving antenna of the base station and the lth receiving path of the coordinate origin, is the signal path difference between the monitor receiving antenna and the lth receiving path of the coordinate origin. but is the elevation angle of the base station (monitor) receiving end arrival angle, is the azimuth of the arrival angle of the base station (monitor) receiving end, is the path response matrix from the kth sensor coordinate origin to the base station (monitor) coordinate origin, is the sending field response vector of the base station (monitor), is the signal distance difference between the movable antenna of the kth sensor and the lth transmission path of the coordinate origin, is the elevation angle of the kth sensor departure angle, is the azimuth of the departure angle of the kth sensor.

[0037] Then, based on the link channel models from sensor to monitor and sensor to base station determined above, the mean square error representation of the base station can be further determined, and the lower bound representation of the false detection probability of the monitor can be determined, so that the covert communication constraints can be determined based on the lower bound representation of the false detection probability of the monitor, and the optimization problem can be constructed with the mean square error of the base station as the target.

[0038] Specifically, in one or more embodiments of the present invention, the mean square error of the base station determined by the link channel model from each sensor to the base station is expressed as:

[0039] In the formula, is the signal misalignment error, u is the receiving beam of the base station, and v k is the transmission beam of the kth sensor, and the subscript H represents the conjugate transpose. is the error caused by base station noise, is the base station noise power, and MSE is the mean square error of the base station.

[0040] The lower bound of the probability of false detection by the monitor can be obtained using the Pinsker inequality, which can be expressed as: D(p0(y w )||p1(y w ))≤2ε 2 ,

[0041] In the formula, y w is the received signal of the monitor, Null hypothesis The following y w The likelihood function of the null hypothesis Indicates that the sensor did not send information. For the alternative hypothesis The following y w The likelihood function of Indicates that the sensor sends information, D(p0(y w )||p1(y w )) is p0(y w ) to p1(y w ), σ0 is the received signal power of the monitor under the null hypothesis, is the received noise power of the monitor, σ1 is the received signal power of the monitor under the alternative hypothesis, and ε is the preset covert communication constraint threshold.

[0042] After that, the optimization problem can be constructed based on the mean square error expression of the above base station and the lower bound of the error detection probability of the monitor. Specifically, in one or more embodiments of the present invention, the antenna position optimization model can be constructed by the following formula:

[0043]

[0044] Where C1 is the covert communication constraint, C2 is the maximum transmission power constraint of the sensor, C3 is the movable area constraint of each movable antenna, Γ k is the movable area of ​​the movable antenna of the kth sensor, P k is the maximum transmit power of the kth sensor.

[0045] When solving the antenna position optimization model constructed above, it is difficult to solve it directly. Figure 3 This is a schematic diagram of the process of constructing and solving an antenna position optimization model in the present invention. Figure 3 The process shown solves the antenna position optimization model.

[0046] In one or more embodiments of the present invention, the concealed communication constraints in the antenna position optimization model can be equivalently transformed based on the Lambert W function, and the antenna position optimization model can be equivalently transformed based on the concealed communication constraints after the equivalent transformation. Then, the antenna position optimization model after the equivalent transformation is converted into an inner and outer double-layer iterative problem that satisfies the penalty dual decomposition algorithm framework by using the augmented Lagrangian function method. Finally, the inner and outer double-layer iterative problem is solved to determine the optimal position of each movable antenna.

[0047] Among them, when solving the inner and outer double-layer iterative problems, the inner layer optimization problem can be decomposed by the block coordinate descent method according to the penalty parameters and dual variables of the current inner and outer double-layer iterative problems, and the base station receiving beam optimization subproblem, covert communication optimization subproblem, sensor transmitting beam optimization subproblem and antenna position optimization subproblem are obtained.

[0048] Then, the four optimization sub-problems obtained by decomposition are iteratively solved until the number of iterations reaches a preset number of iterations, and the current optimal inner optimization variable value of the solution containing the four optimization sub-problems is determined.

[0049] Afterwards, for the outer optimization problem, the equality constraint violation degree value and control parameter value of the current inner and outer double-layer iterative problem are determined according to the current optimal inner optimization variable value.

[0050] Finally, determine whether the degree of violation of the equality constraint of the current inner and outer double-layer iterative problem is less than the preset threshold; if so, take the current optimal inner optimization variable value as the solution to the inner and outer double-layer iterative problem; if not, update the penalty parameter and dual variable according to the current control parameter value, and iteratively solve the inner optimization problem.

[0051] Specifically, in one or more embodiments of the present invention, the definition of the Lambert W function can be used to equivalently transform the C1 hidden constraint in the optimization problem P1 into:

[0052] In the formula, is the Lambert W function.

[0053] Then, by introducing the auxiliary variable τ k , the optimization problem P1 can be equivalently expressed as:

[0054]

[0055] In problem P2, the objective function and constraints are highly coupled. P2 is a non-convex problem. The augmented Lagrangian function method can be used to put the equality constraint C4 into the objective function. Problem P2 can be converted into an inner and outer double-layer iterative problem that satisfies the penalty dual decomposition algorithm framework. Its expression is:

[0056]

[0057] Where χ={u,{t k},{v k},{τ k}} is the inner optimization variable, μ=[μ1,…,μ k ,…,μ K ] T and ξ are outer optimization variables, {μ k} is the dual variable corresponding to the equality constraint C4, and ξ is the penalty parameter.

[0058] For the inner optimization problem, the penalty parameter and the dual variable are fixed to {μ (i) ,ξ (i)}, which is the inner optimization problem. This problem is a non-convex problem. Using the block coordinate descent method, the inner optimization problem can be decomposed into four sub-problems, namely the base station receiving beam u optimization sub-problem, the auxiliary variable (hidden communication optimization) {τ k}Optimization sub-problem, the sensor's transmission beam {v k}Optimization subproblem and movable antenna position {t k}Optimize subproblems.

[0059] Among them, the base station's receiving beam u optimization subproblem can be expressed as: This problem is an unconstrained optimization problem, and its optimal solution closed expression is:

[0060] Auxiliary variables (covert communication optimization) {τ k The optimization sub-problem can be expressed as:

[0061]

[0062] Problem P5 can be decomposed into K sub-problems that can be optimized in parallel. The closed-form expression of its optimal solution is obtained by using KKT conditions and Lagrange multiplier method:

[0063] The sensor's transmit beam {v k The optimization sub-problem can be expressed as:

[0064]

[0065] Problem P6 can be decomposed into K sub-problems that can be optimized in parallel. The Lagrange multiplier method is used to obtain the closed-form expression of its optimal solution:

[0066] Antenna position {t k The optimization sub-problem can be expressed as:

[0067]

[0068] Problem P7 can be decomposed into K sub-problems that can be optimized in parallel. and Substituting the expression into the objective function and using the MM (Majorization Minimization) method to minimize the upper bound, each subproblem of problem P7 can be transformed into:

[0069]

[0070] in, ||·||1 means finding the 1-norm of the matrix. Express request The first-order derivative of , Re{x} means taking the real part of x.

[0071] Problem P8 is a convex quadratic programming problem. We use the projection algorithm and let The optimal closed-form expression can be obtained as: in, The mth coordinate component value representing the lower bound of the movable area of ​​the movable antenna of the kth sensor, The mth coordinate component value representing the upper limit of the movable area of ​​the movable antenna of the kth sensor.

[0072] Repeat the optimization of the four inner sub-problems until the maximum number of iterations is greater than the preset number of iterations, and obtain the current optimal inner optimization variable value χ (i) .

[0073] For the outer optimization problem, the outer layer first determines the constraint violation degree value in the current inner and outer double-layer iterative problem according to the current optimal inner optimization variable value. and the control parameter value β (i) =0.9ω(χ (i-1) ), if the current equality constraint violation value is less than the preset threshold, the outer loop stops and outputs the final result; otherwise, the penalty parameter and the dual variable are updated according to the current control parameter value, and the inner optimization problem is further iterated until the inner and outer double-layer iterative problems are solved.

[0074] based on Figure 1 The hidden aerial computing method based on a movable antenna is shown. The movable antenna is connected to the radio frequency unit through a flexible cable. Based on the position of the movable antenna as the optimization variable, the maximum transmission power of each sensor, the moving range of the movable antenna and the hidden communication constraint as the constraints, and the minimization of the mean square error of the base station as the optimization goal, the antenna position optimization model is constructed and solved, so as to determine the optimal position of the movable antenna relative to the sensor, so as to drive the antenna to move to the optimal position in the two-dimensional plane, improve the degree of freedom of the channel in the continuous space, effectively avoid the position where the channel state is in deep fading, further improve the system performance, and reduce the system error. At the same time, only one radio frequency unit is required, which reduces the cost and energy consumption.

[0075] The concealed aerial computing system of the present invention uses a single movable antenna to replace the traditional fixed multi-antenna system, which can effectively improve the problems of high hardware cost, high energy consumption and low utilization rate of spatial freedom of the traditional fixed multi-antenna system, and reduce the error of concealed aerial computing. Secondly, the joint optimization algorithm of transceiver and antenna position based on the penalty dual decomposition inner and outer double-layer iterative architecture proposed by the present invention, the inner layer consists of four independent sub-problems, and each sub-problem of the inner layer can be further decoupled into multiple sub-sub-problems that can run in parallel, and the optimal solution closed expression of each sub-sub-problem is derived, which makes the algorithm complexity extremely low and easy to implement in a real system.

[0076] When applying the concealed aerial computing method based on the movable antenna provided by the present invention, it is not necessary to Figure 1 The steps are executed in the order shown. The specific execution order of the steps can be determined according to needs, and the present invention does not limit this.

[0077] In addition, the present invention also provides a simulation embodiment of the covert aerial computing method based on a movable antenna, and the simulation parameters are set as follows: N = 4, K = 6, L = 3, all sensors are evenly distributed in a circle with a center of (0m, 0m) and a radius of 10m, the base station coordinates are (0m, 20m), the monitor coordinates are (0m, 10m), and the path response matrix is a diagonal matrix, and each diagonal component obeys a cyclic symmetric complex Gaussian distribution, that is, represents the distance from the kth sensor to the base station (detector), α b =2.8 and α w =3.5 is the corresponding path loss factor, All sensors have the same maximum transmit power constraint, that is, P1=P2=…=P K =P max , λ = 0.01 m, ε = 0.1, movable area of ​​the movable antenna of the sensor Γ k =[-λ,λ]×[-λ,λ], 1≤k≤K. MATLAB software was used for simulation, and the number of Monte Carlo simulations was 100.

[0078] Figure 4 The figure is a schematic diagram showing the comparison of mean square error between the method of the present invention and the antenna position fixed algorithm and the antenna position random algorithm. Figure 4 It can be seen that the mean square error of all algorithms increases with the sensor maximum transmit power constraint P max This is because the larger maximum P max This means that each sensor can transmit signals with greater power, thereby obtaining a lower mean square error. Figure 4 It can also be found that by optimizing the antenna position, the method of the present invention is superior to the two algorithms of fixed antenna position and random antenna position in terms of mean square error performance, which also proves the effectiveness of the method of the present invention and the necessity of optimizing the antenna position.

[0079] The above is a concealed aerial computing method based on a movable antenna provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding concealed aerial computing system based on a movable antenna, including:

[0080] A concealed aerial computing system of a base station, a monitor and sensors of multiple IoT devices; the base station is provided with multiple fixed antennas, the monitor is provided with a single fixed antenna; each of the sensors is provided with a single movable antenna and uploads information to the base station via the movable antenna; the movable antenna is movable relative to the sensor and is connected to a radio frequency unit via a flexible cable;

[0081] A calculation module is used to establish link channel models from each sensor to a base station and from each sensor to a monitor respectively, taking the position of a movable antenna as a variable; determining a mean square error representation of the base station according to the link channel model from each sensor to the base station; determining a lower bound representation of the false detection probability of the monitor according to the link channel model from each sensor to the monitor, and determining a covert communication constraint according to the lower bound representation of the false detection probability of the monitor; constructing an antenna position optimization model with the position of the movable antenna as an optimization variable, the maximum transmission power of each sensor, the moving range of the movable antenna and the covert communication constraint as constraints, and minimizing the mean square error of the base station as an optimization goal; solving the antenna position optimization model to determine the optimal position of each movable antenna relative to the sensor at which it is located;

[0082] The antennas of the sensors of each IoT device are moved to the optimal position obtained by the calculation module, and the information collected by each sensor is uploaded and integrated to the base station through hidden aerial calculation.

[0083] For the specific definition of the concealed aerial computing device based on the movable antenna, please refer to the definition of the concealed aerial computing method based on the movable antenna above, which will not be repeated here. Each module in the above-mentioned concealed aerial computing device based on the movable antenna can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0084] The present invention also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 A covert aerial computing method based on movable antenna is provided.

[0085] The present invention also provides a computer device. At the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 A covert aerial computing method based on movable antenna is provided.

[0086] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0087] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A covert aerial computing method based on a movable antenna, characterized in that: The method is applied to a concealed aerial computing system including a base station, a monitor and sensors of multiple IoT devices; the base station is arranged with multiple fixed antennas, the monitor is arranged with a single fixed antenna; each of the sensors is arranged with a single movable antenna and uploads information to the base station through the movable antenna; The movable antenna is movable relative to the sensor and is connected to the radio frequency unit via a flexible cable; the method comprises: Taking the position of the movable antenna as a variable, the link channel models from each sensor to the base station and from each sensor to the monitor are established respectively; Determine the mean square error representation of the base station according to the link channel model from each sensor to the base station; determine the lower bound representation of the false detection probability of the monitor according to the link channel model from each sensor to the monitor, and determine the covert communication constraint according to the lower bound representation of the false detection probability of the monitor; The antenna position optimization model is constructed by taking the position of the movable antenna as the optimization variable, the maximum transmission power of each sensor, the moving range of the movable antenna and the covert communication constraint as the constraints, and minimizing the mean square error of the base station as the optimization goal; Solve the antenna position optimization model to determine the optimal position of each movable antenna relative to the sensor where it is located, so that each movable antenna can upload and fuse the information collected by each sensor to the base station through covert aerial calculation at the optimal position.

2. The covert aerial computing method based on a movable antenna as claimed in claim 1, characterized in that: The link channel models from each sensor to the base station and from each sensor to the monitor are established respectively with the position of the movable antenna as a variable, specifically including: Taking the position of the movable antenna as a variable, the link channel model from each sensor to the base station is established by the following formula: Taking the position of the movable antenna as a variable, the link channel model from each sensor to the monitor is established by the following formula: Among them, t k is the movable antenna position of the kth sensor, is the receiving field response matrix of the base station, and the subscript H represents the conjugate transpose. is the receiving field response vector of the monitor, is the path response matrix from the kth sensor coordinate origin to the base station coordinate origin, is the path response matrix from the kth sensor coordinate origin to the monitor coordinate origin, is the base station’s transmit field response vector, is the transmitted field response vector of the monitor, is the link channel model from the kth sensor to the base station, is the link channel model from the kth sensor to the monitor.

3. The covert aerial computing method based on a movable antenna as claimed in claim 1, characterized in that: The mean square error of the base station determined according to the link channel model from each sensor to the base station is expressed as: in, is the signal misalignment error, u is the receiving beam of the base station, t k is the movable antenna position of the kth sensor, is the link channel model from the kth sensor to the base station, v k is the transmission beam of the kth sensor, and the subscript H represents the conjugate transpose. is the error caused by base station noise, is the base station noise power, and MSE is the mean square error of the base station.

4. The covert aerial computing method based on a movable antenna as claimed in claim 1, characterized in that: The lower bound of the false detection probability of the monitor is determined according to the link channel model from each sensor to the monitor: D(p0(y w )||p1(y w ))≤2ε 2 , Among them, y w is the received signal of the monitor, p0(y w ) is the null hypothesis The following y w The likelihood function of the null hypothesis Indicates that the sensor did not send information, p1(y w ) is the alternative hypothesis The following y w The likelihood function of Indicates that the sensor sends information, D(p0(y w )||p1(y w )) is p0(y w ) to p1(y w ), σ0 is the received signal power of the monitor under the null hypothesis, is the received noise power of the monitor, σ1 is the received signal power of the monitor under the alternative hypothesis, ε is the preset covert communication constraint threshold, is the link channel model from the kth sensor to the monitor, t k is the movable antenna position of the kth sensor, v k is the transmitting beam of the kth sensor.

5. The concealed aerial computing method based on a movable antenna as claimed in claim 3, characterized in that: The constructing of the antenna position optimization model specifically includes: The antenna position optimization model is constructed by the following formula: Among them, C1 is the covert communication constraint, C2 is the maximum transmission power constraint of the sensor, C3 is the movable area constraint of each movable antenna, Γ k is the movable area of ​​the movable antenna of the kth sensor, P k is the maximum transmission power of the kth sensor, p0(y w ) is the null hypothesis The following y w The likelihood function of the null hypothesis Indicates that the sensor did not send information, p1(y w ) is the alternative hypothesis The following y w The likelihood function of Indicates that the sensor sends information, D(p0(y w )||p1(y w )) is p0(y w ) to p1(y w ), and ε is the preset covert communication constraint threshold.

6. The covert aerial computing method based on a movable antenna as claimed in claim 5, characterized in that: The step of solving the antenna position optimization model to determine the optimal position of each movable antenna relative to the sensor in which it is located specifically includes: Based on the Lambert W function, the covert communication constraints in the antenna position optimization model are equivalently transformed, and based on the covert communication constraints after the equivalent transformation, the antenna position optimization model is equivalently transformed; The augmented Lagrangian function method is used to transform the antenna position optimization model after equivalent transformation into an inner and outer double-layer iterative problem that satisfies the penalty dual decomposition algorithm framework. The inner and outer double-layer iterative problem is solved to determine the optimal position of each movable antenna.

7. The covert aerial computing method based on a movable antenna as claimed in claim 6, characterized in that: The solving of the inner and outer double-layer iterative problem specifically includes: According to the penalty parameters and dual variables of the current inner and outer double-layer iterative problem, the block coordinate descent method is used to decompose the inner optimization problem, and the base station receiving beam optimization subproblem, covert communication optimization subproblem, sensor transmission beam optimization subproblem and movable antenna position optimization subproblem are obtained. Iteratively solve the four optimization sub-problems obtained by decomposition until the number of iterations reaches a preset number of iterations, and determine the current optimal inner optimization variable value of the solution containing the four optimization sub-problems; For the outer optimization problem, the equality constraint violation degree value and control parameter value of the current inner and outer double-layer iterative problem are determined according to the current optimal inner optimization variable value; Determine whether the equality constraint violation degree value of the current inner and outer double-layer iterative problem is less than the preset threshold; if so, take the current optimal inner layer optimization variable value as the solution to the inner and outer double-layer iterative problem; if not, update the penalty parameter and dual variable according to the current control parameter value, and iteratively solve the inner layer optimization problem.

8. A concealed aerial computing system based on a movable antenna, characterized in that: include: A covert aerial computing system with a base station, a monitor, and sensors for multiple IoT devices; The base station is provided with a plurality of fixed antennas, the monitor is provided with a single fixed antenna, each of the sensors is provided with a single movable antenna and information is uploaded to the base station via the movable antenna; the movable antenna is movable relative to the sensor and is connected to the radio frequency unit via a flexible cable; A calculation module is used to establish link channel models from each sensor to a base station and from each sensor to a monitor respectively, taking the position of the movable antenna as a variable; determining a mean square error representation of the base station according to the link channel model from each sensor to the base station; determining a lower bound representation of the false detection probability of the monitor according to the link channel model from each sensor to the monitor, and determining a covert communication constraint according to the lower bound representation of the false detection probability of the monitor; Taking the position of the movable antenna as the optimization variable, the maximum transmission power of each sensor, the moving range of the movable antenna and the covert communication constraint as the constraints, and minimizing the mean square error of the base station as the optimization goal, an antenna position optimization model is constructed; the antenna position optimization model is solved to determine the optimal position of each movable antenna relative to the sensor where it is located; The antennas of the sensors of each IoT device are moved to the optimal position obtained by the calculation module, and the information collected by each sensor is uploaded and integrated to the base station through hidden aerial calculation.

9. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

Images