A physical layer security based method and device for aerial computing of unmanned aerial vehicles

Abstract

The application belongs to the technical field of aerial computing, and specifically discloses a kind of unmanned aerial vehicle aerial computing method and device based on physical layer security, the method is: constructing unmanned aerial vehicle aerial computing model;Unmanned aerial vehicle mean square error function and eavesdropper mean square error function are constructed;With minimizing the mean square error at the unmanned aerial vehicle as the optimization goal, with minimizing the mean square error at the unmanned aerial vehicle as the optimization goal, with the receiving coefficient of unmanned aerial vehicle, the transmission coefficient of sensor and the trajectory of unmanned aerial vehicle as the optimization variable, and the mean square error function of eavesdropper is constrained, it is greater than or equal to the preset threshold, to this target optimization problem is constructed;The constructed target optimization problem is solved by using alternating optimization algorithm, and the optimal transmission coefficient, the optimal receiving coefficient of sensor and the optimal path of unmanned aerial vehicle are obtained;Wireless signal is transmitted based on the optimal transmission coefficient, the optimal receiving coefficient of sensor and the optimal path of unmanned aerial vehicle.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of aerial computing, in particular to a physical layer security based unmanned aerial vehicle aerial computing method and device. BACKGROUND

[0002] The aerial computing technology adopts the superposition characteristics of the wireless channel, and utilizes the signal superposition characteristics of the wireless multiple access channel to perform the calculation of the mathematical function, so that the calculation is completed in the transmission process. Multiple nodes access the data center node in a non-orthogonal multiple access manner, and the receiving node can receive the transmitted superposition signal. By processing the superposition signal, the required data can be directly obtained. However, in actual situations, the transmission process will be affected by noise. In addition, the power of the wireless signal sent by the sensor is limited. Therefore, by constraining the receiving factor, the transmission coefficient and the unmanned aerial vehicle trajectory under different conditions, the power of the entire system is controlled, so as to obtain the minimum mean square error at the receiving base station, so as to obtain more accurate calculation values and improve the reception accuracy of the wireless signal.

[0003] In recent years, research on aerial computing has focused on power allocation, with the goal of optimizing system performance by minimizing mean square error. However, the performance of aerial computing is often limited by the worst channel conditions among all links between users and the fusion center; this also means that even if most links have good channel conditions, the poor quality of individual links can still significantly affect overall performance. To solve this problem, unmanned aerial vehicles can be used as fusion centers.

[0004] Unmanned aerial vehicles can meet the specific application requirements of low latency, reliable command control and high-speed data transmission. Related research covers multiple fields, including low-latency communication, reliable command control, high-speed data transmission, spectrum efficiency and network coordination. In aerial computing, using unmanned aerial vehicles as fusion centers has the advantages of flexible deployment and dynamic adjustment, which can effectively avoid ground obstacles and interference and optimize channel conditions. However, ensuring the security of unmanned aerial vehicle communication and preventing eavesdroppers from intercepting messages sent by users to unmanned aerial vehicles is a challenging problem.

[0005] In actual aerial computing transmission processes, due to the broadcast nature of wireless channels, malicious eavesdropping nodes are inevitable in the signal transmission process. The eavesdropping nodes steal the messages sent by the sensors, which makes the signal transmission process have a high risk of being eavesdropped.

[0006] Traditional communication technology usually adopts a mode of "communication first and calculation second", that is, data in a distributed terminal device is transmitted to a central server through a wireless network, and centralized calculation is completed on the server; this mode completely separates communication and calculation process, which can meet the demand in early simple application scenarios, but its limitations become increasingly obvious with the increase of large-scale device access and complex intelligent tasks. SUMMARY

[0007] The present application provides a physical layer security based unmanned aerial vehicle air computing method, which optimizes the receiving coefficient of the unmanned aerial vehicle, the transmission coefficient of the sensor and the trajectory of the unmanned aerial vehicle, so that the quality of the wireless signal received by the unmanned aerial vehicle is improved, and the safety of communication and the safety of wireless signal transmission are ensured.

[0008] The second object of the present application is to provide a physical layer security based unmanned aerial vehicle air computing device.

[0009] The technical solution of the present application to solve the above technical problems is:

[0010] A physical layer security based unmanned aerial vehicle air computing method, comprising the following steps:

[0011] Step S1: constructing an unmanned aerial vehicle air computing model comprising an unmanned aerial vehicle, a sensor and an eavesdropper;

[0012] Step S2: constructing a mean square error function of the unmanned aerial vehicle according to the wireless signal received by the unmanned aerial vehicle, and constructing a mean square error function of the eavesdropper according to the wireless signal received by the eavesdropper;

[0013] Step S3: taking minimization of the mean square error at the unmanned aerial vehicle as the optimization objective, taking the receiving coefficient a0[n] of the unmanned aerial vehicle, the transmission coefficient b k [n] of the sensor and the trajectory q[n] of the unmanned aerial vehicle as optimization variables, and imposing a constraint on the mean square error function of the eavesdropper, so that it is greater than or equal to a preset threshold, to construct a target optimization problem;

[0014] Step S4: solving the constructed target optimization problem by using an alternating optimization algorithm to obtain the optimal transmission coefficient of the sensor, the optimal receiving coefficient of the sensor and the optimal path of the unmanned aerial vehicle;

[0015] Step S5: transmitting the wireless signal based on the optimal transmission coefficient of the sensor, the optimal receiving coefficient of the unmanned aerial vehicle and the optimal path of the unmanned aerial vehicle.

[0016] Furthermore, in step S1, the UAV aerial computing model consists of a UAV, K sensors, and an eavesdropper, wherein each sensor is equipped with a single antenna; the eavesdropper is equipped with a single antenna; each sensor is used to send wireless signals to the UAV; the UAV receives the superimposed signal of the wireless signals sent by all sensors; at the same time, the eavesdropper also intercepts the wireless signals sent by the sensors.

[0017] Assuming the drone's flight time is T, and dividing the flight time T into n∈N={1,2,...,N} time slots, then each time slot is... Assume the drone's flight altitude remains constant at H0, and represent the drone's horizontal position as q[n] = [x[n], y[n]]. T The following constraints are then imposed on the trajectory of the drone:

[0018]

[0019] q[0]=q0,q[N+1]=q F (2)

[0020] In the formula: L=Vt s This represents the maximum displacement of the UAV within a time slot, where V is the maximum velocity of the UAV; q0 and q F These represent the initial and final positions of the drone, respectively.

[0021] Furthermore, in step S2, the steps for constructing the mean square error function of the drone and the mean square error function of the eavesdropper are as follows:

[0022] Step S201: Calculate the average value of the wireless signals transmitted by all sensors:

[0023]

[0024] Where, x k [n] represents the wireless signal transmitted by sensor k;

[0025] Step S202: Calculate the wireless signal received by the drone and the wireless signal received by the eavesdropper, where...

[0026] The wireless signal received by the drone is:

[0027]

[0028] In the formula: b k [n] represents the transmission coefficient of the wireless signal transmitted by sensor k; n0[n] represents the mean and variance of the UAV, which is 0. 2 Additive white Gaussian noise, i.e., n0[n]~CN(0,σ) 2) ; h k [n] is a channel coefficient from sensor k to the UAV, where,

[0029]

[0030] where: w k = [x k , y k ] T is the position of sensor k; p0represents the channel gain at the reference distance d0= 1 m between sensor k and the UAV;

[0031] The wireless signal received by the eavesdropper is:

[0032]

[0033] where: n e [n] is an additive Gaussian white noise with mean 0 and variance s 2 , i.e., n e [n] ~ CN(0, s 2 ); g k [n] is a channel coefficient from sensor to the eavesdropper, where,

[0034]

[0035] where: w e = [x e , y e ] T is the position of the eavesdropper; p e represents the channel gain at the reference distance d0= 1 m between the eavesdropper and the UAV;

[0036] Step S203: imposing a constraint on the transmission power of sensor k in any time slot n;

[0037]

[0038] where: P max is the maximum transmission power of sensor k;

[0039] Step S204: obtaining the estimation function of the UAV and the eavesdropper, where,

[0040] The estimation function of the UAV is:

[0041]

[0042] The estimation function of the eavesdropper is:

[0043]

[0044] where: a0[n] and ae [n] represent the receiving coefficients of the UAV and the eavesdropper, respectively;

[0045] Step S205: obtaining the mean square error functions of the UAV and the eavesdropper, wherein,

[0046] The mean square error function of the UAV is:

[0047]

[0048] The mean square error function of the eavesdropper is:

[0049]

[0050] Step S206: constructing a target optimization problem by the obtained mean square error functions of the UAV and the eavesdropper.

[0051] Further, in step S3, the constructed target optimization problem (P0) is:

[0052]

[0053] In the formula, ξ is the difference threshold of the mean square error function of the eavesdropper at each time slot.

[0054] Further, in step S4, the target optimization problem is decomposed into sub-problem one, sub-problem two and sub-problem three, and the optimization variables are fixed alternately by an alternating optimization algorithm to perform iterative solution; wherein,

[0055] The sub-problem one is: fixing the receiving coefficient a0[n] of the UAV / the receiving coefficient a e [n] of the eavesdropper, the transmission coefficient b k [n] of the sensor k and the trajectory q[n] of the UAV, obtaining the optimal solution of a e [n] / a0[n]:

[0056] The sub-problem two is: fixing the receiving coefficient a0[n] of the UAV and the trajectory q[n] of the UAV, optimizing the transmission coefficient b k [n] of the sensor k.

[0057] The sub-problem three is: fixing the receiving coefficient a0[n] of the UAV and the transmission coefficient b k [n] of the sensor k, optimizing the trajectory q[n] of the UAV.

[0058] Further, the specific solution process of the sub-problem one is:

[0059] Fixing the receiving coefficient a0[n] of the UAV, the transmission coefficient b k [n] of the sensor k and the trajectory q[n] of the UAV, obtaining the optimal solution of a MSEe ({a e [n],b k The first derivative of [n], q[n]}) yields a e The optimal solution for [n]:

[0060]

[0061] Substituting equation (12) into MSE e ({a e [n],b k In [n], q[n]}, and rearranged, we get:

[0062]

[0063] The receiver's reception coefficient a e [n], the transmission coefficient b of sensor k k Given [n] and the trajectory q[n] of the drone, we obtain the optimal solution for a0[n].

[0064]

[0065] Substituting equation (13) into equation (11), the original objective optimization problem (P0) is rewritten as:

[0066]

[0067] Then, based on the objective optimization problem (P1), subproblems two and three are solved.

[0068] Furthermore, for subproblem two, the specific solution process is as follows:

[0069] The receiving coefficient a0[n] at the fixed UAV location, the UAV trajectory q[n], and the transmission coefficient b to sensor k. k [n] is optimized; equation (13) is transformed into:

[0070]

[0071] right b k [n] is expanded and simplified using a first-order Taylor series to obtain:

[0072] Therefore, subproblem two is:

[0073]

[0074] Solve the second subproblem mentioned above.

[0075] Furthermore, for subproblem three, the specific solution process is as follows:

[0076] The receiving coefficient a0[n] of the fixed unmanned aerial vehicle, the transmission coefficient b k [n] of the sensor k, optimize the trajectory q[n] of the unmanned aerial vehicle;

[0077] Each of the above is expanded to obtain:

[0078]

[0079] Introduce auxiliary variables:

[0080] L0={l k [n] = ||q[n]-w k || 2 ,n∈N,k∈K}

[0081] Let Obtain

[0082] Let And make a first-order Taylor expansion of f k [n] about ||q[n]-w k || 2 Obtain the lower bound And process the lower bound Obtain:

[0083]

[0084] From the above:

[0085]

[0086] Secondly, for the auxiliary variable L0={l k [n] = ||q[n]-w k || 2 ,n∈N,k∈K}, the constraint condition is l k [n]≤||q[n]-w k || 2 , where ||q[n]-w k || 2 For q[n], it is a convex constraint, but the original constraint l k [n]≤||q[n]-w k || 2 is a non-convex constraint; similarly, Taylor expansion is performed under the condition that q (r) [n] is given to obtain:

[0087] ||q[n]-w k || 2 ≥||q (r) [n]-w​k || 2 +2(q (r) [n]-w k ) T (q[n]-q (r) [n])

[0088] The arrangement obtains:

[0089] l k [n]≤||q (r) [n]-w k || 2 +2(q (r) [n]-w k ) T (q[n]-q (r) [n]),

[0090] Then the third sub-problem is:

[0091]

[0092] Solving the third sub-problem above.

[0093] Further, in step S4, the receiving coefficient a0[n] of the UAV, the transmission coefficient b k [n] of the sensor k and the trajectory q[n] of the UAV are solved by an alternating optimization algorithm to obtain the solution of the target optimization problem (P1), wherein the optimization process of the alternating optimization algorithm is:

[0094] Step S401: setting a preset difference threshold ξ, the number of sensors K, initializing the transmission coefficient b k [n] of the sensor k, substituting and calculating the initial value of a0[n] and a e [n], initializing the trajectory q[n] of the UAV, and calculating the initial MSE0({a0[n],b k [n],q[n]}) according to formula (9);

[0095] Step S402: let t=1,...,T, start the tth iteration;

[0096] Step S403: fixing According to formula (14), the optimal solution

[0097] Step S404: fixing According to formula (17), the optimal solution

[0098] Step S405: fixing The optimal solution q at this time is calculated according to formula (19) (t+1) [n];

[0099] Step S406: Given to calculate the mean square error of the UAV, and the mean square error of the eavesdropper is also calculated;

[0100] Step S407: Calculate the relative difference of the mean square error of the UAV and the mean square error of the eavesdropper, and judge whether the relative difference is less than the difference threshold ξ;

[0101] If the relative difference is greater than the difference threshold ξ, then t=t+1, and steps S403-S406 are repeated;

[0102] If the relative difference is less than the difference threshold ξ, then the loop is exited, and the output is the result.

[0103] A UAV air computing device based on a UAV air computing method based on physical layer security, comprising:

[0104] A model construction module for constructing a UAV air computing model;

[0105] A target function construction module for constructing a target optimization function with the minimization of the mean square error function at the UAV as the optimization target, and the receiving coefficient at the UAV, the transmission coefficient, and the trajectory of the UAV as the optimization variables;

[0106] An alternating optimization module for solving the constructed target optimization function using an alternating optimization algorithm to obtain an optimal solution.

[0107] Compared with the prior art, the present application has the following advantages and beneficial effects:

[0108] 1. The UAV air computing method based on physical layer security of the present application models the minimization of the mean square error at the UAV by deriving the functional relationship between the receiving coefficient of the UAV, the transmission coefficient of the sensor, and the system parameters such as the trajectory of the UAV and the mean square error, and optimizes the transmission coefficient and the trajectory of the UAV using an alternating optimization algorithm to minimize the mean square error, thereby ensuring secure transmission of information.

[0109] 2. The UAV air computing method based on physical layer security of the present application uses a UAV as a fusion center, which can effectively avoid ground obstacles and interference and optimize channel conditions; and by optimizing the transmission coefficient of the sensor and the trajectory of the UAV, the mean square error at the UAV end is minimized, so that the UAV can receive as much accurate signal as possible, and at the same time, by processing the mean square error at the eavesdropper end, the eavesdropper can only intercept as little signal as possible, thereby ensuring secure communication.

[0110] 3、The unmanned aerial vehicle air computing method based on physical layer security adopts air computing technology, sensors send wireless signals to the unmanned aerial vehicle, the unmanned aerial vehicle can quickly receive superimposed signals sent by all sensors, but the existence of eavesdroppers needs to be considered at the same time, the eavesdroppers need to steal as few signals as possible by setting a threshold to ensure safe transmission. In this scenario, the transmission coefficient of the sensor, the receiving coefficient of the unmanned aerial vehicle and the trajectory of the unmanned aerial vehicle are optimized by using an alternating optimization algorithm, the mean square error minimization problem of the unmanned aerial vehicle is modeled, the signal quality received by the legal receiver is ensured, the signal quality received by the eavesdropper is reduced, and the safety performance of wireless signal transmission is improved. BRIEF DESCRIPTION OF DRAWINGS

[0111] Figure 1 It is a flowchart of the unmanned aerial vehicle air computing method based on physical layer security.

[0112] Figure 2 It is a schematic diagram of the unmanned aerial vehicle air computing model.

[0113] Figure 3 It is a comparison diagram of the minimum mean square error at the unmanned aerial vehicle under different flight times.

[0114] Figure 4 It is a comparison diagram of the minimum mean square error at the unmanned aerial vehicle under different transmission powers P.

[0115] Figure 5 It is a structure block diagram of the unmanned aerial vehicle air computing device based on physical layer security.

[0116] Figure 6 It is a structure schematic diagram of the computer equipment. DETAILED DESCRIPTION

[0117] The present application will be further described in conjunction with the embodiments and drawings, but the embodiments of the present application are not limited thereto.

[0118] Those skilled in the art can understand that, unless specifically stated, the singular forms "a", "an" and "the" used herein also include the plural forms. It should be further understood that the use of the term "including" in the specification of the present application means that the features, integers, steps, operations, elements and / or components exist, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we say that an element is "connected" or "coupled" to another element, it can be directly connected or coupled to the other element, or there can be intermediate elements. In addition, the "connection" or "coupling" used herein can include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any single unit and all combinations of the associated listed items.

[0119] Those skilled in the art of the technology will understand that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art in the field to which this application belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.

[0120] Those skilled in the art of the technology will understand that, as used herein, "client," "terminal," and "terminal device" include both devices that are solely wireless signal receivers and devices that have both receiving and transmitting hardware capable of two-way communication over a two-way communication link. Such devices can include cellular or other communication devices with single-line or multiple-line displays, or no display, Personal Communications Service (PCS) devices that can combine a voice and data function, Personal Digital Assistants (PDAs) that can include a radio frequency receiver, pagers, Internet / Intranet access, Web browsers, organizers, calendars, and / or a Global Positioning System (GPS) receiver, conventional laptop and / or palmtop computers or other devices that have a radio frequency receiver. As used herein, "client," "terminal," and "terminal device" can be portable, transportable, installed in a vehicle (aeronautical, maritime, and / or land), or adapted and / or configured for local and / or distributed operation on Earth and / or any other location in space. As used herein, "client," "terminal," and "terminal device" can also be a communication terminal, an Internet terminal, a music / video playing terminal, such as a PDA, a Mobile Internet Device (MID), and / or a mobile phone with music / video playing function, a smart television, a set-top box, and / or the like.

[0121] The hardware referred to by the names "server", "client", "service node" and the like in the present application is essentially an electronic device with the equivalent capabilities of a personal computer, and is a hardware device with a central processing unit (including an arithmetic unit and a controller), a memory, an input device, and an output device, and the like necessary components disclosed by the von Neumann principle. A computer program is stored in the memory, the central processing unit calls the program stored in the external storage into the memory and runs it, executes the instructions in the program, and interacts with the input and output devices, thereby completing a specific function.

[0122] It should be noted that the concept of "server" in the present application can also be extended to the case of a server cluster. According to the network deployment principle understood by those skilled in the art, the servers should be logically divided, and in physical space, these servers can be independent of each other but can be called through an interface, or they can be integrated into a physical computer or a computer cluster. Those skilled in the art should understand this variation and should not be restricted by the implementation of the network deployment of the present application.

[0123] One or more technical features of the present application, unless explicitly specified, can be deployed on a server and accessed by a client remotely calling the online service interface provided by the server, or can be directly deployed and run on a client to implement access.

[0124] The neural network model referred to or possibly referred to in the present application, unless explicitly specified, can be deployed on a remote server and remotely called by a client, or can be deployed on a client with sufficient device capability and directly called by the client. In some embodiments, when it is run on a client, its corresponding intelligence can be obtained through transfer learning in order to reduce the requirement for client hardware running resources and avoid excessive occupation of client hardware running resources.

[0125] The various data involved in the present application, unless explicitly specified, can be remotely stored on a server or stored on a local terminal device, as long as it is suitable for being called by the technical solutions of the present application.

[0126] Those skilled in the art should know that the various methods of the present application, although based on the same concept and described to present commonality among them, are independently executable unless otherwise specified. Similarly, for each embodiment disclosed by the present application, it is based on the same inventive concept, and therefore, for the same concept of expression, and although the concept of expression is different, it is only for the convenience of appropriately transforming the concept, and should be understood as equivalent.

[0127] Unless otherwise indicated herein, the various disclosed embodiments can be combined in any and all permutations. It is intended that the following claims be construed to include all such embodiments.

[0128] Referring to Figure 1 and Figure 2 , the physical layer security based UAV aerial computing method of the present application comprises the following steps:

[0129] Step S1: constructing a UAV aerial computing model comprising a UAV, sensors and an eavesdropper (EVE);

[0130] In this embodiment, the UAV aerial computing model comprises one UAV, K sensors and one eavesdropper, wherein each sensor is equipped with a single antenna; the eavesdropper is equipped with a single antenna; each sensor is used to send wireless signals to the UAV; the UAV receives the superimposed signals of the wireless signals sent by all sensors; at the same time, the eavesdropper also eavesdrops the wireless signals sent by the sensors;

[0131] Assuming that the flight time of the UAV is T, the flight time T is divided into n∈Ν={1,2,...,N} time slots, then each time slot is Assuming that the flight height of the UAV remains unchanged at H0, and the horizontal position of the UAV is represented as q[n]=[x[n],y[n]] T , the following constraints are imposed on the trajectory of the UAV:

[0132]

[0133] q[0]=q0,q[N+1]=q F ;(2)

[0134] In the formula: L=Vt s represents the maximum displacement of the UAV in a time slot, wherein V is the maximum speed of the UAV; q0 and q F represent the initial position and the final position of the UAV, respectively.

[0135] Step S2: constructing a mean square error function of the UAV according to the wireless signals received by the UAV, and constructing a mean square error function of the eavesdropper according to the wireless signals received by the eavesdropper;

[0136] In this embodiment, the construction steps of the mean square error function of the UAV and the mean square error function of the eavesdropper are as follows:

[0137] Step S201: Calculate the average of the wireless signals sent by all sensors:

[0138]

[0139] where x k [n] represents the wireless signal sent by sensor k;

[0140] Step S202: Calculate the wireless signals received by the UAV and the eavesdropper, where

[0141] The wireless signal received by the UAV is:

[0142]

[0143] where b k [n] is the transmission coefficient of the wireless signal sent by sensor k; n0[n] is the additive Gaussian white noise with mean 0 and variance σ 2 of the UAV, i.e., n0[n] ~ CN(0, σ 2 ); h k [n] is the channel coefficient from sensor k to the UAV, where

[0144]

[0145] where w k = [x k , y k ] T is the position of sensor k; p0 represents the channel gain at the reference distance d0 = 1 m between sensor k and the UAV;

[0146] The wireless signal received by the eavesdropper is:

[0147]

[0148] where n e [n] is the additive Gaussian white noise with mean 0 and variance σ 2 at the eavesdropper, i.e., n e [n] ~ CN(0, σ 2 ); g k [n] is the channel coefficient from the sensor to the eavesdropper, where

[0149]

[0150] where w e = [x e , y e ] T is the position of the eavesdropper; p eH0(d0) represents the channel gain of the eavesdropper to the reference distance d0= 1 m of the UAV;

[0151] Step S203: imposing a constraint on the transmission power of the sensor k in any time slot n;

[0152]

[0153] P = Pmax k max is the maximum transmission power of the sensor k;

[0154] Step S204: obtaining the estimation function of the UAV and the eavesdropper, wherein,

[0155] The estimation function of the UAV is:

[0156]

[0157] The estimation function of the eavesdropper is:

[0158]

[0159] a0[n] and a e [n] represent the receiving coefficients of the UAV and the eavesdropper, respectively;

[0160] Step S205: obtaining the mean square error function of the UAV and the eavesdropper, wherein,

[0161] The mean square error function of the UAV is:

[0162]

[0163] The mean square error function of the eavesdropper is:

[0164]

[0165] Step S3: constructing a target optimization problem with the minimum mean square error at the UAV as the optimization objective, the receiving coefficient a0[n] of the UAV, the transmission coefficient b k [n] of the sensor, and the trajectory q[n] of the UAV as the optimization variables; and reducing the information stealing ability of the eavesdropper and enhancing the communication security by constraining the mean square error function of the eavesdropper to be greater than or equal to a preset threshold (i.e., the difference threshold of the mean square error function of the eavesdropper in each time slot);

[0166]

[0167] ξ is the difference threshold of the mean square error function of the eavesdropper in each time slot.

[0168] Step S4: the constructed target optimization problem is solved by using an alternating optimization algorithm to obtain optimal transmission coefficients of the sensors, optimal receiving coefficients of the sensors and optimal paths of the unmanned aerial vehicles;

[0169] The specific idea is: the transmission coefficients of the sensors, the receiving coefficients of the unmanned aerial vehicles and the receiving coefficients of the eavesdroppers are initialized, the adjusted values of the transmission coefficients of the sensors, the receiving coefficients of the unmanned aerial vehicles and the receiving coefficients of the eavesdroppers are determined according to the transmission coefficients of the sensors before adjustment, the receiving coefficients of the unmanned aerial vehicles before adjustment, the receiving coefficients of the eavesdroppers before adjustment, the mean square error function of the unmanned aerial vehicles and the mean square error function of the eavesdroppers, and the transmission coefficients of the sensors and the receiving coefficients of the unmanned aerial vehicles are adjusted; the optimal transmission coefficients of the sensors, the optimal receiving coefficients of the unmanned aerial vehicles and the optimal trajectories of the unmanned aerial vehicles are obtained according to the corresponding transmission coefficients after adjustment, the receiving coefficients of the unmanned aerial vehicles after adjustment and the mean square error function of the unmanned aerial vehicles after adjustment.

[0170] Firstly, the target optimization problem (P0) is decomposed into subproblem one, subproblem two and subproblem three, and the optimization variables are fixed alternately by using an alternating optimization algorithm to perform iterative solving; wherein,

[0171] The subproblem one is: the receiving coefficients a0[n] of the unmanned aerial vehicles / the receiving coefficients a e [n] of the eavesdroppers, the transmission coefficients b k [n] of the sensors k and the trajectories q[n] of the unmanned aerial vehicles are fixed, and the optimal solution of a e [n] / a0[n] is obtained:

[0172] The subproblem two is: the receiving coefficients a0[n] of the unmanned aerial vehicles and the trajectories q[n] of the unmanned aerial vehicles are fixed, and the transmission coefficients b k [n] of the sensors k are optimized;

[0173] The subproblem three is: the receiving coefficients a0[n] of the unmanned aerial vehicles and the transmission coefficients b k [n] of the sensors k are fixed, and the trajectories q[n] of the unmanned aerial vehicles are optimized.

[0174] The specific solving process of the subproblem one is:

[0175] 1) the receiving coefficients a0[n] of the unmanned aerial vehicles, the transmission coefficients b k [n] of the sensors k and the trajectories q[n] of the unmanned aerial vehicles are fixed, the first-order derivative of the MSE e ({a e [n],b k [n],q[n]}) is solved, and the optimal solution of a e [n] is obtained:

[0176]

[0177] Substituting equation (12) into MSE e ({a e [n],b k In [n], q[n]}, and rearranged, we get:

[0178]

[0179] 2) The reception coefficient a of the fixed eavesdropper e [n], the transmission coefficient b of sensor k k Given [n] and the trajectory q[n] of the drone, we obtain the optimal solution for a0[n].

[0180]

[0181] Substituting equation (13) into equation (11), the original objective optimization problem (P0) is rewritten as:

[0182]

[0183] Then, based on the objective optimization problem (P1), subproblems two and three are solved.

[0184] The specific solution process for subproblem two is as follows:

[0185] The transmission coefficient b of sensor k k [n] is optimized by fixing the receiving coefficient a0[n] at the UAV and the trajectory q[n] of the UAV. Since equation (13) is a non-convex constraint, equation (13) is transformed into:

[0186]

[0187] right b k [n] is expanded and simplified using a first-order Taylor series to obtain:

[0188]

[0189] Therefore, subproblem two is:

[0190]

[0191] Solve the second subproblem.

[0192] The specific solution process for subproblem three is as follows:

[0193] The receiver coefficient a0[n] of the fixed UAV and the transmission coefficient b of sensor k. k [n], optimize the drone trajectory q[n];

[0194] Will Expanding each item in the table, we get:

[0195]

[0196] Introduce auxiliary variables:

[0197] L0={l k [n] = ||q[n] - w k || 2 ,n∈N,k∈K}

[0198] make Can be rewritten as This item is for l k For [n], it is a convex constraint;

[0199] make

[0200] Then in The second term is -f k [n ] This term is a non-convex constraint for q[n], which makes the optimization problem (P1) a non-convex problem.

[0201] Although f k [n] is a non-convex constraint for q[n], but for ||q[n]-w k || 2 In other words, it is a convex constraint; therefore, given q (r) In the case of [n], for f k [n] about ||q[n]-w k || 2 Performing a first-order Taylor expansion yields a lower bound, which is denoted as . After processing, we can obtain:

[0202]

[0203] The result of this step If it is a concave function, then The function is convex, thus transforming the optimization problem into a convex optimization problem;

[0204] From the above:

[0205]

[0206] Secondly, for the auxiliary variable L0={l k [n] = ||q[n] - w k || 2 The constraint is l, n∈N, k∈K}. k[n]≤||q[n]-w k || 2 , where ||q[n]-w k || 2 For q[n], it is a convex constraint, but the original constraint l k [n]≤||q[n]-w k || 2 It is a non-convex constraint; similarly, given q (r) Performing a Taylor expansion in the case of [n], we get ||q[n]-w k || 2 ≥||q (r) [n]-w k || 2 +2(q (r) [n]-w k ) T (q[n]-q (r) [n]), then the constraint can be rewritten as l k [n]≤||q (r) [n]-w k || 2 +2(q (r) [n]-w k ) T (q[n]-q (r) [n]),

[0207] Finally, subproblem three can be rewritten as:

[0208]

[0209] Solve the third subproblem mentioned above.

[0210] In summary, the alternating optimization algorithm was used to solve for the receiving coefficient a0[n] at the UAV and the transmission coefficient b of sensor k. k Given [n] and the trajectory q[n] of the UAV, we obtain the solution to the objective optimization problem (P1). The optimization process of the alternating optimization algorithm is as follows:

[0211] Step S401: Set the preset difference threshold ξ, the number of sensors K, and initialize the transmission coefficient b of sensor k. k Substitute [n] and calculate a0[n] and a e The initial value of [n] is determined; the trajectory q[n] of the UAV is initialized, and the initial MSE0({a0[n],b) is calculated according to equation (9). k [n],q[n]});

[0212] Step S402: Let t = 1, ..., T, and begin the t-th iteration;

[0213] Step S403: fixing The optimal solution at this time is calculated according to formula (14)

[0214] Step S404: fixing The optimal solution at this time is calculated according to formula (17)

[0215] Step S405: fixing The optimal solution q at this time is calculated according to formula (19) (t+1) [n];

[0216] Step S406: giving to calculate the mean square error of the UAV and the eavesdropper (i.e. );

[0217] Step S407: calculating the relative difference of the mean square error of the UAV and the mean square error of the eavesdropper (i.e. and judging whether the relative difference is less than the difference threshold value ξ;

[0218] If the relative difference is greater than the difference threshold value ξ, let t = t + 1, and repeat steps S403-S406;

[0219] If the relative difference is less than the difference threshold value ξ, exit the loop, and the output is the result sought.

[0220] Step S5: transmitting the wireless signal based on the optimal transmission coefficient of the sensor, the optimal receiving coefficient of the UAV, and the optimal path of the UAV.

[0221] In the simulation, in addition to the optimization scheme proposed in the present application, three cases of no optimization, only power optimization, and only trajectory optimization are considered as reference schemes to verify the effectiveness of the scheme of the present application. The minimum mean square error function at the UAV is constructed in MATLAB using the following parameters: the number of sensors k = 3, the height of the UAV H = 100 m, the reference channel power gain of the UAV ρ0= 10 -3 , the reference channel power gain of the eavesdropper ρ e = 10 -3 , the noise power σ 2 = 10 -9 w, the maximum speed of the UAV v max = 20 m / s, the initial power p0= 1 w, the position of the eavesdropper is [0, -300], the starting position of the UAV is [0, 500], the ending position of the UAV is [0, -500], the positions of the three sensors are [-50, 300], [0, 0], and [50, -300], and the difference threshold value ξ = 0.2.

[0222] First, consider the simulation at different flight times T, the range of flight time T is between 100s to 200s, and the simulation is carried out in four cases of no optimization, only power optimization, only trajectory optimization and simultaneous power optimization and trajectory optimization respectively, and the simulation results are as shown in Figure 3

[0223] From the simulation results of Figure 3 , it can be seen that, except for the no optimization scheme, the minimum mean square error of the unmanned aerial vehicle in the remaining schemes decreases with the increase of flight time T, and it can be seen from Figure 3 that the scheme of simultaneous trajectory optimization and power optimization is better, and the scheme of the application can realize the smallest mean square error MSE compared with the other two benchmark schemes, because the scheme of the application dynamically balances the transmission power and the distance of all links, fully utilizes the joint optimization capability of trajectory design and power control, so as to illustrate that the joint design of trajectory and power control is crucial to improve the system performance.

[0224] Secondly, consider the simulation at different transmission powers P, the range of transmission power P is between 1w to 9w, and the simulation is carried out in four cases of no optimization, only power optimization, only trajectory optimization and simultaneous power optimization and trajectory optimization respectively, and the simulation results are as shown in Figure 4

[0225] From the simulation results of Figure 4 , it can be seen that, with the increase of transmission power P, the minimum mean square error of the unmanned aerial vehicle in all schemes is decreasing; and within this range, the scheme of the application is always superior to the other three benchmark schemes, which illustrates the effectiveness of the scheme of the application.

[0226] In summary, the unmanned aerial vehicle air computing method based on physical layer security of the application adopts air computing technology, the sensor sends wireless signals to the unmanned aerial vehicle, the unmanned aerial vehicle can quickly receive all the superimposed signals sent by the sensors, but at the same time the existence of eavesdroppers needs to be considered, and the threshold needs to be set to make the eavesdroppers steal as little signal as possible to ensure safe transmission. In this scenario, the transmission coefficient of the sensor, the receiving coefficient of the unmanned aerial vehicle and the trajectory of the unmanned aerial vehicle are optimized by using the alternating optimization algorithm, the minimum mean square error minimization problem of the unmanned aerial vehicle is modeled, the signal quality received by the legal receiver is ensured, the signal quality received by the eavesdropper is reduced, and the safety performance of wireless signal transmission is improved.

[0227] Please refer to Figure 5 ​​, one of the purposes of the present application is to provide a physical layer security based unmanned aerial vehicle air computing device, comprising a model construction module, a target function construction module and an alternating optimization module: wherein the model construction module is used to construct an unmanned aerial vehicle air computing model; the target function construction module is used to construct a target optimization function with the minimization of the mean square error function at the unmanned aerial vehicle as the optimization target, and the receiving coefficient at the unmanned aerial vehicle, the receiving coefficient at the eavesdropper, the transmission coefficient and the trajectory of the unmanned aerial vehicle as the optimization variables; the alternating optimization module is used to solve the constructed target optimization function by using an alternating optimization algorithm to obtain an optimal solution.

[0228] On the basis of any embodiment of the present application, please refer to Figure 6 Another embodiment of the present application further provides an electronic device, which can be realized by a computer device, as shown in Figure 6 The internal structure diagram of the computer device. The computer device comprises a processor, a computer readable storage medium, a memory and a network interface connected by a system bus. Among them, the computer readable storage medium of the computer device stores an operating system, a database and a computer readable instruction, the database can store a control information sequence, and the computer readable instruction can make the processor realize a physical layer security based unmanned aerial vehicle air computing method when executed by the processor. The processor of the computer device is used to provide computing and control ability to support the operation of the whole computer device. The memory of the computer device can store computer readable instructions, which can make the processor execute the physical layer security based unmanned aerial vehicle air computing method of the present application when executed by the processor. The network interface of the computer device is used to connect and communicate with the terminal. Those skilled in the art can understand that Figure 6 The structure shown in the figure is only a block diagram of part of the structure related to the present application scheme, and does not constitute a limitation on the computer device to which the present application scheme is applied. The specific computer device can include more or less components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0229] The processor in the embodiment is used to execute the specific functions of each module in Figure 5 The memory stores the program codes and various data required for executing the above-mentioned modules. The network interface is used for data transmission between the user terminal or the server. The memory in the embodiment stores the program codes and data required for executing all modules / sub-modules in the physical layer security based unmanned aerial vehicle air computing device of the present application. The server can call the program codes and data of the server to execute the functions of all sub-modules.

[0230] The application further provides a storage medium storing computer readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the physical layer security based aerial computing method of any of the embodiments of the application.

[0231] The application further provides a computer program product comprising computer programs / instructions, which, when executed by one or more processors, implement the steps of the physical layer security based aerial computing method of any of the embodiments of the application.

[0232] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiments of the application can be completed by a computer program instructing relevant hardware, and the computer program can be stored in a computer readable storage medium. When the program is executed, it can include the processes of the above-mentioned embodiments of the method. The storage medium can be a computer readable storage medium such as a magnetic disc, an optical disc, a read-only memory (ROM), or a random access memory (RAM).

[0233] The above is the preferred embodiment of the application, but the embodiments of the application are not limited by the above, and any change, modification, replacement, combination, simplification, etc. made without departing from the spirit and principle of the application shall be equivalent replacement and included in the protection scope of the application.

Claims

1. A method for aerial computing of unmanned aerial vehicles based on physical layer security, characterized in that, Includes the following steps: Step S1: Construct an aerial computing model of the drone that includes the drone, sensors, and eavesdroppers; Step S2: Construct the mean square error function of the UAV based on the wireless signals received by the UAV; Based on the wireless signals received by the eavesdropper, construct the mean square error function of the eavesdropper; Step S3: The optimization objective is to minimize the mean square error at the UAV, using the UAV's receiving coefficient a0[n] and the sensor's transmission coefficient b. k [n] and the trajectory q[n] of the drone are used as optimization variables. Constraints are imposed on the mean square error function of the eavesdropper to make it greater than or equal to a preset threshold, thereby constructing the objective optimization problem. Step S4: Use the alternating optimization algorithm to solve the constructed target optimization problem to obtain the optimal transmission coefficient of the sensor, the optimal reception coefficient of the sensor, and the optimal path of the UAV; Step S5: Transmit the wireless signal based on the optimal transmission coefficient of the sensor, the optimal reception coefficient of the UAV, and the optimal path of the UAV.

2. The UAV aerial computing method based on physical layer security according to claim 1, characterized in that, In step S1, the UAV aerial computing model consists of a UAV, K sensors, and an eavesdropper. Each sensor is equipped with a single antenna; the eavesdropper is also equipped with a single antenna; each sensor is used to send wireless signals to the UAV; the UAV receives the superimposed signal of the wireless signals sent by all the sensors; at the same time, the eavesdropper also intercepts the wireless signals sent by the sensors. Assuming the drone's flight time is T, and dividing the flight time T into n∈N={1,2,...,N} time slots, then each time slot is... Assume the drone's flight altitude remains constant at H0, and represent the drone's horizontal position as q[n] = [x[n], y[n]]. T The following constraints are then imposed on the trajectory of the drone: q[0]=q0,q[N+1]=q F ; (2) In the formula: L=Vt s This represents the maximum displacement of the UAV within a time slot, where V is the maximum velocity of the UAV; q0 and q F These represent the initial and final positions of the drone, respectively.

3. The UAV aerial computing method based on physical layer security according to claim 2, characterized in that, In step S2, the steps for constructing the mean square error function of the drone and the mean square error function of the eavesdropper are as follows: Step S201: Calculate the average value of the wireless signals transmitted by all sensors: Where, x k [n] represents the wireless signal transmitted by sensor k; Step S202: Calculate the wireless signal received by the drone and the wireless signal received by the eavesdropper, where... The wireless signal received by the drone is: In the formula: b k [n] represents the transmission coefficient of the wireless signal transmitted by sensor k; n0[n] represents the mean and variance of the UAV, which is 0. 2 Additive white Gaussian noise, i.e., n0[n]~CN(0,σ) 2 );h k [n] represents the channel coefficients from sensor k to the UAV, where, In the formula: w k =[x k ,y k ] T ρ0 represents the position of sensor k; ρ0 represents the channel gain at the reference distance d0 = 1m between sensor k and the UAV. The wireless signal received by the eavesdropper is: Where: n e [n] represents the number of eavesdroppers with a mean of 0 and a variance of σ. 2 Additive white Gaussian noise, i.e., n e [n]~CN(0,σ 2 );g k [n] represents the channel coefficients from the sensor to the eavesdropper, where, In the formula: w e =[x e ,y e ] T The location of the eavesdropper; ρ e This represents the channel gain at a reference distance d0 = 1m from the eavesdropper to the drone. Step S203: Apply a constraint to the transmission power of sensor k in any time slot n; In the formula: P max The maximum transmit power of sensor k; Step S204: Obtain the estimation functions for the drone and the eavesdropper, where, The estimation function for the UAV is: The eavesdropper's estimation function is: In the formula: a0[n] and a e [n] represents the reception coefficients of the drone and the eavesdropper, respectively; Step S205: Obtain the mean square error functions of the drone and the eavesdropper, where, The mean square error function of the UAV is: The mean squared error function of the eavesdropper is: Step S206: Construct the target optimization problem using the mean square error functions of the drone and the eavesdropper.

4. The UAV aerial computing method based on physical layer security according to claim 3, characterized in that, In step S3, the objective optimization problem (P0) is constructed as follows: In the formula: ξ is the threshold of difference in the mean square error function of the eavesdropper in each time slot.

5. The UAV aerial computing method based on physical layer security according to claim 4, characterized in that, In step S4, the target optimization problem is decomposed into sub-problems one, two, and three, and an alternating optimization algorithm is used to alternately fix the optimization variables for iterative solution; wherein, The first sub-problem is: the fixed drone's reception coefficient a0[n] / the eavesdropper's reception coefficient a e [n], the transmission coefficient b of sensor k k Given [n] and the trajectory q[n] of the drone, we can obtain a e The optimal solution for [n] / a0[n]: The second sub-problem is: given a fixed receiver coefficient a0[n] and a drone trajectory q[n], and a transmission coefficient b for sensor k. k [n] is optimized; The third sub-problem is: the receiving coefficient a0[n] of the fixed UAV and the transmission coefficient b of sensor k. k [n], optimize the drone trajectory q[n].

6. The UAV aerial computing method based on physical layer security according to claim 5, characterized in that, The specific solution process for sub-problem one is as follows: The receiver coefficient a0[n] of the fixed UAV and the transmission coefficient b of sensor k. k [n] and the drone trajectory q[n] are obtained by calculating the MSE. e ({a e [n],b k The first derivative of [n], q[n]}) yields a e The optimal solution for [n]: Substituting equation (12) into MSE e ({a e [n],b k In [n], q[n]}, and rearranged, we get: The receiver's reception coefficient a e [n], the transmission coefficient b of sensor k k Given [n] and the trajectory q[n] of the drone, we obtain the optimal solution for a0[n]. Substituting equation (13) into equation (11), the original objective optimization problem (P0) is rewritten as: Then, based on the objective optimization problem (P1), subproblems two and three are solved.

7. The UAV aerial computing method based on physical layer security according to claim 6, characterized in that, For subproblem two, the specific solution process is as follows: The receiving coefficient a0[n] at the fixed UAV location, the UAV trajectory q[n], and the transmission coefficient b to sensor k. k [n] is optimized; equation (13) is transformed into: right b k [n] is expanded and simplified using a first-order Taylor series to obtain: Therefore, subproblem two is: Solve the second subproblem mentioned above.

8. The UAV aerial computing method based on physical layer security according to claim 7, characterized in that, For subproblem three, the specific solution process is as follows: The receiver coefficient a0[n] of the fixed UAV and the transmission coefficient b of sensor k. k [n], optimize the trajectory q[n] of the drone; Will Expanding each item in the table, we get: Introduce auxiliary variables: L0={l k [n]=||q[n]-w k || 2 ,n∈N,k∈K} make get make And for f k [n] about ||q[n]-w k || 2 Performing a first-order Taylor expansion yields the lower bound. And to the lower realm After processing, we get: From the above: Secondly, for the auxiliary variable L0={l k [n] = ||q[n] - w k || 2 The constraint is l, n∈N, k∈K}. k [n]≤||q[n]-w k || 2 , where ||q[n]-w k || 2 For q[n], it is a convex constraint, but the original constraint l k [n]≤||q[n]-w k || 2 It is a non-convex constraint; similarly, given q (r) Taylor expansion for [n] yields: ||q[n]-w k || 2 ≥||q (r) [n]-w k || 2 +2(q (r) [n]-w k ) T (q[n]-q (r) [n]) The results were: Then the third subproblem is: Solve the third subproblem mentioned above.

9. The UAV aerial computing method based on physical layer security according to claim 8, characterized in that, In step S4, the receiving coefficient a0[n] at the UAV and the transmission coefficient b of sensor k are solved by an alternating optimization algorithm. k Given [n] and the trajectory q[n] of the UAV, we obtain the solution to the objective optimization problem (P1). The optimization process of the alternating optimization algorithm is as follows: Step S401: Set the preset difference threshold ξ, the number of sensors K, and initialize the transmission coefficient b of sensor k. k Substitute [n] and calculate a0[n] and a e The initial value of [n] is determined; the trajectory q[n] of the UAV is initialized, and the initial MSE0({a0[n],b) is calculated according to equation (9). k [n],q[n]}); Step S402: Let t = 1, ..., T, and begin the t-th iteration; Step S403: Fix The optimal solution at this point is calculated according to equation (14). Step S404: Fix The optimal solution at this point is calculated using equation (17). Step S405: Fix The optimal solution q at this time is calculated according to equation (19). (t+1) [n]; Step S406: Given To calculate the mean square error of the drone, and at the same time, to calculate the mean square error of the eavesdropper; Step S407: Calculate the relative difference between the mean square error of the drone and the mean square error of the eavesdropper, and determine whether the relative difference is less than the difference threshold ξ. If the relative difference is greater than the difference threshold ξ, then let t = t + 1 and repeat steps S403-S406. If the relative difference is less than the difference threshold ξ, exit the loop and output the result. This is the desired result.

10. A UAV aerial computing device for the UAV aerial computing method based on physical layer security as described in any one of claims 1-9, characterized in that, include: Model building module: used to build aerial computational models of UAVs; Objective function construction module: Used to construct an objective optimization function with the goal of minimizing the mean square error function at the drone, and the receiving coefficient at the drone, the receiving coefficient at the eavesdropper, the transmission coefficient, and the drone's trajectory as optimization variables; Alternating Optimization Module: Used to solve the constructed objective function using the alternating optimization algorithm to obtain the optimal solution.

Images