Block chain enabled and IRS-UAV assisted Internet of Vehicles multi-target computing unloading method

Abstract

The invention discloses a block chain enabling and IRS-UAV (Internet of Vehicles-Unmanned Aerial Vehicle) assisted multi-target computing unloading method for the Internet of Vehicles (V2X). Combination of V2X and edge calculation is the trend of V2X development at present, which can reduce the load of a central network and reduce the traffic transmitted to a remote data center. However, mobile edge computing (MEC) has a centralization problem, which will amplify the risk of data leakage. Therefore, the block chain technology is introduced, and the problems of single-point failure, trust, safety, transparency and the like in a centralized system are solved. Meanwhile, in order to prevent eavesdropping, leakage of signals in the direction of an eavesdropper is reduced by adjusting a phase shift matrix of an intelligent reflecting surface (IRS). Specifically, a vehicle in the system generates a computing task, the computing task is reflected by a UAV equipped with an IRS, and finally the computing task is transmitted to a base station. And the base station allocates computing resources according to the cost provided by the vehicle and the task size. In the data transmission process, after transmission of one node is completed, a block is generated and recorded, and transmission traceability, non-repudiation and the like are achieved. According to the scheme, the time delay, the energy consumption and the cost generated in the transmission process are balanced with the communication secrecy rate through phase shift matrix adjustment and power optimization, and the weighted total utility of the system is maximized.

Description

Technical Field

[0001] This scheme designs a blockchain-enabled and IRS-UAV-assisted vehicle-to-everything (V2X) multi-target computing offloading method, which belongs to the field of wireless communications. Background Art

[0002] In recent years, the trend of V2X and mobile edge computing integration has shown a growing state, which effectively reduces the network load. By processing and analyzing data at the edge node, the traffic transmitted to the remote data center can be reduced, the burden on the core network can be reduced, and the efficiency and reliability of the overall network can be improved. However, the current edge computing architecture faces significant centralization issues, which bring many challenges. For example, the centralized edge computing architecture has problems with flexibility, data privacy, and security.

[0003] As a decentralized distributed ledger, blockchain technology can be used to solve this problem. The integration of V2X and blockchain further enhances the security and reliability of cross-network data transmission. Smart contracts run on the blockchain and are responsible for performing automatic and trustless task offloading processes between nodes. This mechanism ensures that transactions are not rejected and that malicious behavior can be tracked.

[0004] Intelligent Reflective Surface (IRS) is an emerging wireless communication technology. It is a uniform planar array reflective element composed of a large number of low-cost passive devices, each of which can be independently phase-adjusted, aiming to improve the performance of signal transmission by adjusting the phase and amplitude of electromagnetic waves. IRS technology can adjust the reflective surface in real time to adapt to the changing network environment when the vehicle moves quickly, providing a stable communication link. This effectively solves the problem of secure connection caused by the frequent changes in the topology of the V2X network. At the same time, we deploy IRS on an unmanned aerial vehicle (UAV) hovering at a certain height, and the mission data sent by the vehicle is reflected by the IRS equipped by the UAV. This can not only expand the signal transmission range so that it is not blocked by obstacles in the urban environment; it can also prevent the eavesdropper (Eve) from eavesdropping on the information by adjusting the phase shift matrix of the IRS, so as to achieve the purpose of secure transmission. Summary of the invention

[0005] Technical problem solved: In view of the shortcomings of existing solutions, the present invention provides a blockchain-enabled IRS-UAV-assisted multi-objective computing offloading solution, which comprehensively considers the communication confidentiality rate, energy consumption, latency and task cost, while considering the fairness of computing resource allocation to maximize the total utility of the system.

[0006] Technical solution:

[0007] The present invention introduces a blockchain-enabled IRS-UAV-assisted multi-objective computing offloading scheme in V2X, characterized in that the task offloading scheme includes the following steps:

[0008] Step 1, establish a scenario in which IRS-UAV assists mobile vehicles in task offloading in blockchain-oriented V2X. The ground facilities include K MEC servers deployed with the base station and I mobile vehicles. Eve tries to eavesdrop on the vehicle's unloading task, and IRS is deployed on the UAV to suppress eavesdropping. This improves the security performance of the system;

[0009] Step 2: Establish the communication model, energy consumption model and cost model of IRS phase shift matrix, transmission power adjustment and BS allocation computing resources respectively, and build the optimization model based on the communication model, energy consumption model and cost model. The communication model is used to calculate the total delay from the task sending to the task calculation completion and the confidentiality rate during the communication process; the energy consumption model is used for UAV hovering energy consumption, transmission energy consumption, computing energy consumption and block generation energy consumption; the cost model includes block fees and the fees required for MEC;

[0010] Step 3, set the initial values ​​of relevant optimization variables;

[0011] Step 4, according to the current optimal power value, the flow optimization method is used to solve the phase shift matrix of IRS;

[0012] Step 5: Based on the current optimal IRS phase shift matrix, the transmit power is solved using an analytical method.

[0013] Step 6: Use the alternating optimization algorithm to loop through Step 4 and Step 5 until the system utility value of the next iteration between two adjacent iterations is less than or equal to the system utility value of the previous iteration, or the maximum preset number of iterations is reached, and the iteration ends.

[0014] Step 7, based on the current state of the entire V2X network, apply the IRS phase shift matrix, power adjustment, and computing resource allocation strategy to offload tasks and perform calculations on the vehicles in the V2X with the goal of maximizing the system weighted total utility value. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is the scenario diagram of this solution.

[0016] Figure 2 This is the algorithm solution process diagram proposed in this solution.

[0017] Figure 3 It is an algorithm diagram for alternately optimizing the IRS phase shift matrix and the transmission power.

[0018] Figure 4 It is a comparison diagram of three schemes in the embodiment of this scheme. DETAILED DESCRIPTION

[0019] The present solution is further described below in conjunction with the accompanying drawings and specific implementation methods.

[0020] like Figure 1 As shown in the figure, an IRS-UAV-assisted multi-target computing offloading scheme in blockchain-enabled V2X is proposed. This scheme maximizes the total system effect by increasing the communication confidentiality rate and reducing energy consumption and latency, and allocates resources based on the balance between costs and resources. The specific steps are:

[0021] Step 1: Establish a blockchain-oriented drone-assisted V2X task offloading system equipped with IRS. K MEC servers provide computing services for I different mobile vehicles on the ground, where the mobile vehicles are represented by {N1, N2, ..., N I Eve attempts to eavesdrop on the task of unloading a mobile vehicle using N e It means that eavesdropping on the channel transmission information attempts to disrupt data transmission, which may cause the computing task to fail. The unloading task corresponding to the vehicle is represented by {Z1,Z2....Z i ...Z I} indicates. i represents the mission of the i-th vehicle. IRS is deployed on UAV to suppress eavesdropping, thereby improving the security performance of the system.

[0022] Assume that the physical location information of each device including the eavesdropper is known, the location of the car, the location of the drone, the location of the base station and the location of the eavesdropper are represented by (x i ,y i ), (x u ,y u ,h u ), (x b ,y b ), (x e ,y e ) is represented by the Cartesian coordinate system. The distance between nodes is calculated using the Cartesian coordinate system. i Distance from car to BS, N i The distance from the vehicle to the UAV, N i The distance from the car to Eve is expressed as:

[0023]

[0024] The distance from IRS-UAV to BS and the distance from IRS-UAV to Eve are expressed as:

[0025]

[0026] The distance from BS to Eve is expressed as:

[0027]

[0028] Assuming that signal transmission is carried out in millimeter wave communication, N i Channel gain link between vehicle and IRS-UAV express,

[0029]

[0030] Where N is the number of IRS reflective elements, Represent the approximate cosine values ​​of the deviation angles of the links from vehicle to IRS-UAV and vehicle to eavesdropper, respectively. λ and d represent the wavelength carrier and antenna separation distance, respectively, and ω is the path loss when the reference distance is 1 meter.

[0031] N i The channel gain from the vehicle to the eavesdropper is It can be expressed as:

[0032]

[0033] The channel gains from IRS-UAV to the base station and from IRS-UAV to the eavesdropper are expressed as and express:

[0034]

[0035] N i The channel gain from vehicle to base station is express:

[0036]

[0037] The diagonal matrix of IRS reflective elements N is denoted by ψ i =diag[exp(jθ i,1 ),…,exp(jθ i,n ),…exp(jθ i,N )],θ i,n ∈[0,2π), is the phase introduced by the nth phase shifter in the ith period of IRS.

[0038] N i The data transmission rate that can be achieved by the car is denoted by R i,b express:

[0039]

[0040] In formula (12), B is the bandwidth, w iis the vehicle-to-uplink transmission power, σ 2 is the noise power.

[0041] Step 2: Establish a weighted system total utility maximization optimization problem for the above system, and maximize the system utility by jointly optimizing IRS phase shift, vehicle transmission power, and resource allocation.

[0042] Specifically:

[0043] If N i The vehicle transmits the task to the BS via IRS-UAV reflection, and its data transmission delay can be expressed as: in is the data task size. In addition to considering the transmission delay, the entire system also needs to consider the calculation delay of the MEC server. express,

[0044]

[0045] in, is the number of CPU cycles required to calculate 1 bit. is the task size, The computing resources available to the MEC server.

[0046] The energy consumption generated by the system during task offloading. The energy consumption generated by each node will be introduced below;

[0047] The energy consumption of the vehicle includes the energy consumption of unloading data tasks and the energy consumption of generating blocks. We use and To express:

[0048]

[0049] The energy consumption generated by generating blocks for vehicles is a fixed value. Therefore, the total energy consumption generated by the vehicle is expressed as:

[0050]

[0051] The second is the energy consumption of UAV. Because UAV does not forward data, the IRS equipped on UAV only reflects data. During the process of IRS reflecting data, UAV hovers at a fixed height, so the energy consumption of UAV includes hovering energy consumption and block generation energy consumption. express,

[0052]

[0053] in is the hovering power, which depends on the mass of the drone, the radius of the propeller, gravity, and air density. In theory, these parameters are constant at a fixed altitude and stable environmental conditions. is the total time to complete the task, that is,

[0054] Like the energy consumption of vehicle blocks, the energy consumption of UAV blocks is also a fixed value, so the total energy consumption of UAV is express,

[0055]

[0056] Finally, let’s look at the energy consumption of BS and MEC. MEC nodes need to perform task calculations, which will generate corresponding energy consumption. After MEC completes the calculation, it will return the result to BS, and BS will record the calculation result and the identity information of the MEC that performed the calculation in the block. Therefore, their energy consumption includes the calculation energy consumption of MEC and the block generation energy consumption of BS. express,

[0057]

[0058] Where κ is the capacitance coefficient (set / fixed value), Assigned to N i Vehicle to Mission Z i The amount of computing resources available to the MEC server, It is the number of CPU cycles required for MEC to calculate 1 bit, and the energy consumption of generating blocks is express.

[0059] So complete Z i The total energy consumption required by the task BS and MEC is:

[0060]

[0061] Complete Mission Z i The total energy consumption required is the sum of vehicle energy consumption, UAV energy consumption, BS energy consumption and MEC server energy consumption, expressed as:

[0062]

[0063] In addition, there is the issue of fees. Because the system uses blockchain to record transactions, a portion of the gas fee must be paid. Not only the blockchain needs to pay, but MEC also needs to pay a portion of the fee. Among them, the reward paid to the blockchain for completing the task is

[0064]

[0065] Indicates that the payment to the server is

[0066]

[0067] So the total cost of completing the task is

[0068]

[0069] Among them, ρ Gas , ρ Mec They are the gas price and the cost per bit of Mec service. Based on latency, energy consumption and cost issues, the system utility is obtained.

[0070]

[0071] in,

[0072]

[0073] Through the model given above, the optimization problem is established, which is as follows:

[0074]

[0075] swt min ≤w≤w max , (26a)

[0076]

[0077] PPR≤r, (26c)

[0078]

[0079] ψ i =diag[exp(jθ i,1 ),…,exp(jθ i,n ),…exp(jθ i,N )]. (26e)

[0080] Since P1 is a non-convex problem and cannot be solved directly, the original problem is decomposed into two sub-problems, namely phase shift optimization and power optimization.

[0081] Step 3: Figure 2 As shown in the figure, an algorithm for finding the IRS phase shift matrix using a flow optimization method is proposed, specifically:

[0082] First, the subproblem of dependency matrix optimization is written as:

[0083]

[0084] st(26e).

[0085] Since the target parameter ψ i It is difficult to find the solution between the two CSI parameters. So P2 is equivalently transformed to:

[0086]

[0087] P3 is considered as a complex circular flow optimization problem. The Riemannian gradient descent algorithm is used, where the Riemannian gradient of a manifold point is derived from the Euclidean gradient and the manifold The tangent space at this point is jointly determined. For any point s on the manifold, the tangent space is composed of all tangent vectors that pass through s. Each tangent vector represents a direction, and the objective function can be optimized along this direction starting from s. For a complex circular manifold, the tangent space at s is

[0088]

[0089] Where v represents q i The tangent vector at (s), is the Hadamard product, is the real part of the complex variable, is the conjugate operation.

[0090] About q i The Euclidean gradient of (s) is

[0091]

[0092] Among all the tangent vectors, the fastest growing objective function is the Riemann gradient, that is, It is the process from Euclidean gradient to tangent space The projection of

[0093]

[0094] In the Riemann gradient descent process, q i The direction of the next point of (s) is t s j s , where t s is the step length, j s is the descending direction vector. The retraction function is introduced to ensure that the next point q i (s+1) is still popular On, it is expressed as:

[0095]

[0096] A typical shrinkage function is a normalization function, i.e. Based on the Riemann gradient and conjugate gradient descent algorithms to find the optimal matrix, the update rule for searching on the manifold is as follows:

[0097]

[0098] in, is the Polak-Ribiere parameter. is the tangent vector j s From the cut space To the tangent space The mapping function is:

[0099]

[0100] Because g(q i )'s second-order derivative ( and ) is positive semidefinite, so g(q i ) is concave, which makes the conjugate gradient descent algorithm 1 cannot guarantee convergence to the optimal point. Therefore, algorithm 1 gives a local optimal solution.

[0101] Step 4: After solving the phase shift matrix of the local optimal solution in Step 3, use the analytical method to solve the optimal power value in P4. Figure 3 The algorithm shown alternately optimizes the phase shift matrix and power to find the optimal system utility value. It can be expressed as:

[0102]

[0103] st(26a),(26d).

[0104] Step 5: Allocate computing resources to the tasks of the vehicles. In order to ensure that the ratio of the cost paid by each vehicle to the computing resources obtained by the corresponding task of the vehicle is not less than the threshold, that is, the constraint PPR ≥ r, N I Vehicles are classified according to the fees they have to pay. The requirements are sorted as {N 1ˊ ,N 2ˊ ,N 3ˊ ,...,N Iˊ}, MEC servers are sorted as {M 1ˊ ,M 2ˊ ,M 3ˊ ,...,M Kˊ}, the order of MEC servers follows Generate a bipartite graph that records all combinations of vehicles and MECs.

[0105] Set up a matching matrix W, where the (i,k) element represents an edge, such as W i,k=1 means that vertices i and m are connected. For the other edge (ii, kk), we have i≠ii and k≠kk to satisfy the constraint that any two edges do not share vertices. For W, the computing resource allocation problem can be equivalently transformed into a two-dimensional matching problem, which is

[0106]

[0107] Among them, W i,k is a binary variable. k Assign to N i It takes 1 when , otherwise it takes 0. According to the constraints, each vehicle can only be assigned to one MEC server, and each MEC server can only be assigned to one vehicle. Optimization is to find the optimal W to minimize P5 is convex, and the optimal W can be solved by the Kuhn-Munkres (KM) algorithm.

[0108] like Figure 4 As shown in the figure, three schemes are considered: (1) No IRS phase shift matrix optimization: During the optimization process, there is no IRS phase shift matrix optimization, and only the transmit power is optimized. (2) Fixed transmit power size: After the power value is given, only the IRS phase shift matrix is ​​optimized. (3) IRS phase shift optimization and power optimization: During the optimization process, one of the values ​​is fixed and the other value is optimized, and the two are performed alternately to achieve the maximum total effect value.

[0109] The specific implementation methods of the present scheme have been described in detail above in conjunction with the accompanying drawings, but the present scheme is not limited to the above implementation methods. Various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present scheme.

Claims

1. A blockchain-enabled and IRS-UAV-assisted multi-objective computing offloading method for Internet of Vehicles. Its features are as follows: The integration of vehicle task offloading and mobile edge computing (MEC) in the vehicle-to-everything (V2X) network effectively reduces the burden on the network center and the traffic of network transmission, but at the same time, MEC faces centralization problems, such as single point failure, which will increase the risk of data leakage and affect communication security; the distributed characteristics of blockchain can solve this problem. Therefore, all nodes in V2X will generate blocks for transaction records after the data transmission is completed, so as to achieve the purpose of secure communication; in addition, there are eavesdroppers in the environment. In order to reduce the impact of eavesdroppers on V2X communication, this solution uses the characteristics of the reflection unit of the intelligent reflecting surface (IRS) to adjust the phase and amplitude, so as to achieve the purpose of increasing the confidentiality rate of communication; specifically: Step 1: Establish a scenario in which IRS-UAV assists mobile vehicles in task offloading in blockchain-oriented V2X. The ground facilities include K MEC servers deployed with the base station, and A moving vehicle, Eve attempts to eavesdrop on the vehicle unloading task, IRS is deployed on the UAV to suppress eavesdropping, thereby improving the security performance of the system; Step 2: Establish the communication model, energy consumption model and cost model of IRS phase shift matrix, transmission power adjustment and BS allocation computing resources respectively. Build an optimization model based on the communication model, energy consumption model and cost model. The communication model is used to calculate the total delay from the task sending to the task calculation completion and the confidentiality rate during the communication process; the energy consumption model is used for UAV hovering energy consumption, transmission energy consumption, computing energy consumption and block generation energy consumption; the cost model includes block fees and fees required by MEC; Step 3, set the initial values ​​of relevant optimization variables; Step 4, according to the current optimal power value, the flow optimization method is used to solve the phase shift matrix of IRS; Step 5: Based on the current optimal IRS phase shift matrix, the transmit power is solved using an analytical method. Step 6, using the alternating optimization algorithm, cyclically execute Step 4 and Step 5 until the system utility value of the next iteration between two adjacent iterations is less than or equal to the system utility value of the previous iteration, or the maximum preset number of iterations is reached, and the iteration ends; Step 7, allocate computing resources. Considering that vehicles will provide corresponding fees when unloading tasks, in order to ensure fairness between computing resources and fees, a bipartite graph is used to record the relationship between vehicle fees and MEC computing resources. In addition, this scheme proposes a price-performance ratio (PPR) constraint. For vehicle-MEC pairs that do not meet this condition, the corresponding value in the bipartite graph is set to infinity, indicating that the MEC cannot be allocated to the vehicle. After the bipartite graph is generated, the KM algorithm is used to solve it.

Images