Secure air implementation method integrating mobile edge computing and block chain technology

Abstract

The invention discloses a secure air implementation method integrating mobile edge computing and block chain technology, and the method constructs a secure edge computing framework integrating block chain assisted unmanned aerial vehicles, in an unmanned aerial vehicle assisted edge computing system, a plurality of unmanned aerial vehicles carrying micro servers serve as air edge servers, and the unmanned aerial vehicles carry micro servers. And a calculation unloading service is provided for the Internet of Things equipment. Meanwhile, block chain tasks are deployed on the edge servers, so that the edge servers also serve as consensus nodes of a block chain model; the system uses a practical Byzantine consensus mechanism to increase fault tolerance of blockchain nodes, an unmanned aerial vehicle edge server needs to mount received calculation tasks and corresponding calculation results on a blockchain platform as transactions, and blocks are packaged into the blockchain system through verification of the consensus nodes, so that privacy of users is protected. According to the method, a more rational and efficient unloading strategy is realized, so that the energy consumption and the time delay of the system are reduced, and the effectiveness of the system is improved.

Description

Technical Field

[0001] The present invention relates to a secure air implementation method integrating mobile edge computing and blockchain technology, and belongs to the technical field of wireless communications. Background Art

[0002] With the explosive growth of the number of mobile devices in the Industrial Internet of Things (IoT) and the development of 5G / 6G wireless communication technology. GSMA pointed out that by the end of 2023, nearly 5.6 billion people (69% of the total population) will subscribe to mobile services, and the number of mobile devices connected globally has exceeded 49 billion. The widespread use of IoT provides opportunities for computing-intensive tasks such as telemedicine, smart cars, automatic navigation and games. At the same time, the interactive information functions implemented by these applications have also continuously improved the communication quality and network requirements between mobile users over time. However, the limited battery capacity and low computing power of mobile users make it challenging for them to handle these computing-intensive tasks, such as large storage capacity, high communication quality and ultra-low communication latency. With the development of mobile edge computing (MEC) technology, computing-intensive tasks that mobile users cannot handle themselves can be directly or relayed to MEC servers for processing, which greatly reduces the pressure on mobile terminals to process data. Unlike traditional cloud computing systems, MEC servers are usually placed at the edge of the entire communication network, closer to mobile devices. By assigning computing tasks to MEC servers, the transmission delay and congestion between mobile terminals and MEC servers can be effectively reduced, and unnecessary energy consumption in the network can be avoided.

[0003] However, traditional ground MEC networks are more vulnerable to natural disasters such as earthquakes, and cannot provide computing services to remote areas without available communication equipment infrastructure coverage. In order to solve this problem, more and more scholars have begun to introduce MEC networks supported by unmanned aerial vehicles (UAVs). Due to the unique high controllable mobility and easy deployment and expansion of UAVs, they can effectively expand the coverage of wireless communication networks and conveniently expand to remote areas to provide computing services, provide network capacity and reduce network communication latency. Therefore, deploying MEC servers on UAVs can not only provide aerial computing services for user terminals, but also assist user terminals in processing their computing tasks. For this reason, UAV networks deployed with MEC servers are used as the core framework of wireless communication networks, which has aroused great interest in academia and the industrial Internet of Things community. However, the information interaction process of transmitting the user terminal's own offloaded tasks to the UAV, the migration of the transmitted computing tasks between the user terminal and the UAV, and the environmental differences of mobile users' movements may lead to privacy leakage problems of user terminals.

[0004] It is worth noting that some scholars have introduced blockchain technology into the collaborative network of MEC-UAV, which can solve the above-mentioned privacy issues and ensure the security and privacy of the aerial computing offload tasks transmitted from the user terminal to the UAV. Blockchain is a promising confidentiality technology for building a trust platform. Unlike traditional digital ledgers that rely on third-party neutral institutions, blockchain is a new type of distributed billing technology that relies on point-to-point (P2P) networks. The blockchain system uses its unique encryption and confidentiality technologies to store the communication tasks between the user terminal and the UAV in the blockchain. Due to the immutability of the block, the security and privacy of the communication between the user and the UAV are guaranteed.

[0005] However, some existing scholars who study the joint work of blockchain and MEC-UAV believe that the blockchain layer model and the UAV-MEC layer model are separate. But in fact, the block transaction records in the blockchain need to go through processes such as block generation, block verification and consensus. Since the drone acts as a node in the blockchain, these processes all require the computing resources on the UAV to complete these operations. The offloading of computing tasks also requires the computing resources on the UAV to complete. Therefore, the blockchain layer and the UAV-MEC layer are inseparable. The two restrict and influence each other. How to allocate the computing resources of the UAV to balance the energy consumption and latency between the two parts is an existing challenge. In addition, existing work has hardly considered the optimization of the UAV trajectory, which leads to excessive energy consumption of the transmission data between the UAV and the user terminal, resulting in unnecessary resource loss. Summary of the invention

[0006] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art and propose a secure aerial implementation method integrating mobile edge computing and blockchain technology. The method constructs a secure edge computing framework that integrates blockchain-assisted drones. In the drone-assisted edge computing system, multiple drones equipped with micro servers serve as aerial edge servers to provide computing offload services for IoT devices; at the same time, blockchain tasks are deployed on these edge servers so that they also serve as consensus nodes of the blockchain model. The blockchain nodes of the present invention have a certain degree of fault tolerance, which improves the data security of end users; at the same time, by jointly considering the mutual trade-offs and influences between the two optimization variables of UAV trajectory changes and user associations, a more rational and efficient offloading strategy is made, thereby reducing the energy consumption and delay of the system and improving the utility of the system.

[0007] The technical solution adopted by the present invention to solve the technical problem is: a secure air implementation method integrating mobile edge computing and blockchain technology, the method comprising the following steps:

[0008] Step 1: Establish a system model including M drones, multiple terminal devices and blockchain, where each UAV handles the computing offload tasks of K user terminals in its service area, and deploys a mobile edge computing (MEC) server on each UAV, so that the UAV acts as a MEC server. At the same time, the blockchain server is deployed on the UAV, making it act as a blockchain node. The consensus process will be completed between M UAVs (i.e., M blockchain nodes).

[0009] Step 2: Each UAV acts as an aerial node, and the computing tasks of mobile users are offloaded to the computing nodes for processing. This process determines the computing energy consumption and computing latency of the UAV-MEC model system.

[0010] While receiving tasks from the mobile terminal, the UAV will move in trajectory in order to reduce the transmission energy consumption and delay of communication between it and the user terminal. This part generates the propulsion energy consumption of the UAV.

[0011] Step 3: When the UAV completes the user's computing task, it will collect the computing offload records it generates as "transactions". The transaction (also known as blockchain task) is verified by M UAVs, and a new block will be generated to record the transaction. When more than two-thirds of the drones pass the verification, that is, the entire blockchain network reaches a consensus, the block will be added to the blockchain. This process generates the computing energy consumption and computing latency of the blockchain model.

[0012] Step 4: According to the security offloading energy consumption and latency of the UAV-MEC layer and the blockchain layer, a joint optimization problem of user association, UAV trajectory, and computing resource allocation is constructed to minimize the weighted sum of latency and energy consumption of the entire aerial security computing process.

[0013] Further, the joint optimization problem of step 4 of the present invention is to first decompose the optimization problem of the UAV-MEC layer and the optimization problem of the blockchain model, and the solution step of the UAV-MEC layer optimization problem. First, for the impact of the task offloading scheduling of the terminal user on the delay, combined with the given trajectory and computing resource allocation, the optimal solution of the user's offloading scheduling is solved, and the optimal solution of the user task offloading scheduling is obtained, and then combined with the given computing resource allocation, the continuous graph approximation SCA algorithm is used to optimize and solve the trajectory of the UAV, and then the user task offloading scheduling and UAV trajectory solved before are used to optimize and solve the computing resources allocated to the edge server by the UAV to process the offloading task, and the user task offloading scheduling, UAV trajectory and MEC are obtained. The computing resources allocated to the unmanned aerial vehicle for processing the offloading task are iteratively updated to the convergence of the user task offloading scheduling, UAV trajectory and computing resources through the fast coordinate descent method, and the minimum energy consumption and delay of the MEC-UAV layer are obtained. At the same time, the computing resources required for the edge computing server to allocate to the unmanned aerial vehicle to act as a MEC node have been solved in the above steps, and the computing resources required for the edge computing server to allocate to the unmanned aerial vehicle to act as a blockchain node can be solved later.

[0014] Furthermore, the method of the present invention includes: within the entire flight cycle time T of the UAV, T is evenly divided into n time slots, and the ground user terminal k uses orthogonal frequency division multiple access OFDMA transmission to unload information, assuming that the computing power of the user itself is insufficient to support them in processing computing tasks. However, considering that the MEC server is deployed on the drone, the computing tasks of mobile users can be transmitted to the drone for processing through a wireless link. It is worth noting that the drone can only unload computing tasks for one ground user terminal in each time slot. Define the amount of task data that the user terminal k needs to transmit to the UAV in the nth time slot as l k [n],a k [n] is a user-associated variable, indicating whether user terminal k has a computing task in the nth time slot that needs to be transferred to the UAV for offloading, and its value is [0, 1]. The UAV has a computing task to offload in each time slot, and when a k [n] = 1 means that the offload task processed by the UAV in time slot n is that of end user k, k [n] = 0 means that the task processed by the UAV in time slot n is not for user k, so the amount of task data unloaded from user terminal k to the UAV is a k [n]*l k [n].

[0015] When the kth user terminal offloads the computing task to the UAV in time slot n, the channel gain between the user terminal and the UAV is:

[0016]

[0017] In the present invention, it is assumed that the system adopts OFDMA, that is, the total available bandwidth B is divided equally into K subcarriers for all users, with a size of The transmission rate between mobile user k and UAV in the nth time slot can be expressed as:

[0018]

[0019] in, represents the reference received signal-to-noise ratio (SNR) of the UAV when d0 = 1 m, and N0 represents the noise power density of the UAV. The transmit power of TD k used to offload the task to the UAV for calculation in each time slot n is given by p k [n] indicates.

[0020] According to the fact that the amount of data unloaded by user terminal k should be smaller than the amount of information transmitted, the following calculation causal relationship conditions can be obtained:

[0021] a k [n]l k [n]≤δ t R k [n]

[0022] where δ t is the size of each time slot, a k [n] represents the user-associated variable, l k [n] represents the amount of task data unloaded by user k to UAV in time slot n, R k [n] represents the transmission rate of user k to the UAV in time slot n.

[0023] In addition, the computation task T{n} of user k is characterized by two parameters (l k [n],C k [n]), l k [n] represents the amount of task data unloaded by user k to UAV in time slot n, C k [n] refers to the number of CPU cycles required to complete the computation of the computational task T{n} in time slot n.

[0024] The delay in completing the computing task processing includes the transmission delay of user terminal k offloading the computing task to the UAV and the computing delay of the UAV processing the computing task. The transmission delay of mobile user k offloading the computing task to the UAV is:

[0025]

[0026] In addition, the delay of the drone processing the computing task of mobile user k is:

[0027]

[0028] where f k [n] represents the computing power allocated by the edge computing server to the UAV to process the offloaded tasks of user k. The transmission energy consumption of mobile user k offloading the computing tasks to the UAV is:

[0029]

[0030] Set the CPU power calculation for time slot n to k comp represents the effective energy coefficient of the drone, which depends on the chip architecture of the MEC server equipped on the drone. Therefore, the computational energy consumption of the drone processing the computing task of mobile user k in the nth time slot can be expressed as:

[0031]

[0032] Based on the above analysis, the energy consumption of mobile user k completing the computing task processing in time slot n can be expressed as:

[0033]

[0034] Similarly, the delay for the system to complete the computational task processing of mobile user k in time slot n is:

[0035]

[0036] Furthermore, the method of the present invention includes: a blockchain server is provided on each UAV in the model, and the drone also acts as a blockchain node. There are M drones in total, that is, there are M consensus nodes in the blockchain model, and the Practical Byzantine Consensus Protocol (PBFT) with fault tolerance is used in this blockchain architecture. In addition to supporting fault-tolerant fault nodes, this algorithm can also tolerate malicious nodes. A malicious node refers to a node that may make a completely opposite response to a received message or even forge a message. There are only two roles of drones acting as nodes in the PBFT algorithm, one is a drone acting as a primary node and the other is a drone acting as a replica node. The two roles between drones can be converted to each other.

[0037] For greater fault tolerance, the maximum number of fault-tolerant nodes in the PBFT algorithm is f = (M-1) / 3, where M is the total number of nodes in the entire blockchain system. That is to say, when M is 4, at most one drone can be tolerated as a malicious node or a faulty node in the drone cluster. At this time, even if there is a drone acting as a malicious node, the blockchain system can still reach a consensus, thereby protecting the security and reliability of data unloaded from the user terminal.

[0038] Using the PBFT consensus protocol to reach a blockchain consensus requires four stages, namely the pre-prepare stage of all drone nodes receiving client requests, the preparation stage between drone nodes, the commit stage, and the drone node's reply stage to the client. The consensus process of the blockchain system in the present invention is as follows:

[0039] The client sends a petition to the master node, indicating the start of the request phase. Nodes in the blockchain system collect transactions from the MEC system, including encrypted user identities, computing task results, and computing offload records. The master node needs to verify the signature and message authentication code (MAC) when receiving the transaction. Assume that all nodes require β1 and β2 CPU cycles to generate or verify a signature and a MAC, respectively. Consider that φ represents the maximum number of transactions stored in a block, and θ corresponds to the proportion of correct transactions. Therefore, the computational cost of the master node can be expressed as:

[0040]

[0041] When receiving the pre-prepare message from the master node, the slave node must continue to verify the signature and MAC of the transaction when receiving a new block. Then generate a signature and M-1 MACs. The computational cost of the slave node can be calculated as:

[0042] e2=β1+β2(M-1)+φ(β1+β2).

[0043] In the preparation phase, each replica node needs to send a preparation message to other nodes. The node must verify the signatures and MACs from 2f (f = (M-1) / 3) nodes (including the replica node itself), and then generate a signature and M-1 MACs. Therefore, in this process, the computational cost of the replica node can be expressed as:

[0044] e3=β1+β2(M-1)+2f(β1+β2).

[0045] In the submission phase, each node must verify the 2f signature and MAC when receiving the submission message from other nodes, and then generate a signature and MAC. Therefore, in this process, the computational cost of the node can be expressed as:

[0046] e4=(2f+1)(β1+β2).

[0047] In the reply phase, when f+1 matching commit messages are collected, the new block will become a valid block and will then be broadcast to the blockchain system. In this process, the computational cost can be expressed as:

[0048] e5=(f+1)(β1+β2).

[0049] The computing resource allocation strategy in the blockchain system is represented as F2 = {f m |m∈M}. f m It is the computing power of the blockchain nodes.

[0050] Based on the above analysis, the total delay cost of completing tasks in the blockchain system can be given by the following formula:

[0051]

[0052] Similarly, in the blockchain system, the computing power of the CPU Therefore, the total energy cost of completing a task in the blockchain system can be expressed as:

[0053]

[0054] Furthermore, the method of the present invention includes: in the system, the UAV flies from an initial position to a final position. According to the speed v[n] of the UAV during the flight, the flight power consumption of the UAV can be modeled as:

[0055]

[0056] Where P0 and P i They represent the profile power and induced power of the rotor of the UAV in the hovering state. And Utip, v0, d0, ρ, s and A are parameters related to the aerodynamics of the UAV. Therefore, the flight energy consumption of the UAV during this period is expressed as:

[0057]

[0058] Furthermore, the method of the present invention includes: the total delay cost in the entire drone network system can be expressed as:

[0059]

[0060] The total energy consumption to complete computing tasks and blockchain task processing can be expressed as:

[0061]

[0062] Where ξ is a mapping factor that ensures that the propulsion energy consumption of the drone is at the same level as the computing energy consumption of the drone and the computing energy consumption of the blockchain. Then, according to a secure air implementation method integrating mobile edge computing and blockchain technology as claimed in the above claims, it is characterized in that an optimization problem is constructed with the goal of minimizing the energy consumption and delay of the entire system security computing. The formulated problem to be optimized can be specifically expressed as:

[0063]

[0064] stC1:q[1]=q0,q[N]=q F

[0065] C2:

[0066] C3:

[0067] C4:

[0068] C5:

[0069] C6: l h，k [n]≥0,

[0070] C7:

[0071] C8:

[0072] C9:

[0073] In the above optimization problem, the objective function is the weighted sum of the energy consumption of the entire network and the total delay of completing the computing task and blockchain task processing. w1 (0≤w1≤1) is the weight factor, which integrates the objective function into a whole and balances the weights of delay and energy consumption. F represents the maximum computing power of the drone. By adjusting the value of w1, the preference between energy consumption and delay can be dynamically controlled. When the battery of the drone is low, the value of w1 can be increased to save more computing and communication energy. At the same time, if the network requirements are sensitive to delay, the value of w1 should be reduced to reduce delay. Constraint C1 corresponds to the initial / end position of the drone, constraint C2 is the maximum speed constraint satisfied by the drone, and constraint C3 represents the speed constraint of the drone, whose speed depends on the front and back positions of the drone. Constraints C4-C5 represent the task offloading relationship between the drone and the end user. Constraints C6-C7 represent the causal relationship between the size of the offloaded task of each end user and the transmission capacity between the user and the drone. Constraint C8 represents the guarantee that the computing resources allocated to the offload task and blockchain computing task are less than the maximum computing power of the drone.

[0074] Furthermore, the method of the present invention includes: the convex optimization problem is a MINLP (Mixed-Integer Nonlinear Programming) problem, which combines the user-associated variable A=( k =a k [n]), drone trajectory Computational resource allocation in the UAV-MEC layer F1 = (f k[n]) and the computing resource allocation in the blockchain system F2 = {f m |m∈M}, determine the optimal solution. In order to solve this thorny problem, the above variables are decoupled, and the entire problem is decomposed into several sub-optimization problems for alternate optimization and solution, so as to reduce the algorithm complexity and the difficulty of solution. And the present invention uses the designed user association and computing resource allocation algorithm based on fast coordinate descent (BCD) and the UAV trajectory optimization algorithm based on continuous graph approximation (SCA) to solve the optimal solution of the problem.

[0076] Beneficial effects:

[0077] 1. The present invention proposes a secure aerial implementation method integrating mobile edge computing and blockchain technology. The method improves the communication quality between UAV and terminal users in the UAV-MEC mobile Internet of Things network by introducing a blockchain model. The personal privacy of the terminal is stored in blocks by utilizing the tamperability and privacy characteristics of blockchain technology, thereby ensuring the security and privacy of communication between the user terminal and the UAV.

[0078] 2. The present invention realizes that by optimizing the flight trajectory of UAV in the blockchain-MEC-UAV model, the communication between the terminal user and UAV under secure communication can reduce more transmission delay and energy consumption, balance the weight between the communication loss and computing loss of the system, improve system performance, and reduce unnecessary resource loss.

[0079] 3. The present invention improves and derives the processes of block generation, verification and consensus in block transactions in the blockchain, and enables the server carried by the UAV to act as both a MEC node and a blockchain node, integrating the cross-system of the blockchain and the UAV-MEC, and achieving mutual trade-offs. At the same time, the present invention improves the PBFT consensus algorithm, which can achieve consensus between blockchain nodes even in a dangerous network with malicious nodes, and complete privacy protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 This is a model diagram of the present invention.

[0081] Figure 2 The present invention is a flow chart of the method.

[0082] Figure 3 This is an algorithm flow chart of the practical Byzantine fault-tolerant consensus algorithm adopted in the blockchain model of the present invention. DETAILED DESCRIPTION

[0083] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0084] like Figure 1 As shown, the present invention proposes a secure air implementation method integrating mobile edge computing and blockchain technology, including the following contents:

[0085] Assume that there are K end users, and the UAV is limited by its maximum speed and initial / end position. The total time of the UAV's air flight is assumed to be T. The horizontal plane projection position of the UAV at any time t∈[0, T] can be expressed as q(t)={(x(t), y(t)}, and the height of the UAV is fixed to H. During its trajectory movement, the UAV unloads the user terminal with computing tasks. For convenience, the entire architecture is divided into two layers: UAV-MEC layer and blockchain layer. The horizontal position w of user k k =[x k ,y k ]. This invention uses a sufficiently small constant δ t The UAV flight cycle T is divided into N time slots of equal size, where δ t =T / N. In each time slot, the drone is considered static. Therefore, the horizontal position of the drone at time slot n can be expressed as: q[n] = {x[n], y[n]}, where q(t) = q(δ t n)=q[n]. Then the trajectory of the drone is: And the distance between the user and the UAV:

[0086] 1. Task upload model: Since the size of the result is much smaller than the size of the offloaded data, the delay and energy consumption of sending the return result from the drone to the mobile terminal are omitted.

[0087] Line-of-sight (LOS) propagation is assumed. In addition, the Doppler effect caused by the mobility of the UAV can be effectively compensated at the receiver. The channel power gain between mobile user k and the UAV in the nth time slot is:

[0088]

[0089] β0 represents the channel gain when the reference distance d0=1m.

[0090] Assume that the system adopts OFDMA. The total available bandwidth B is divided equally into K subcarriers for each user, with a size of The transmission rate between mobile user k and UAV in time slot n can be expressed as:

[0091]

[0092] represents the reference received signal-to-noise ratio (SNR) of the UAV when d0 = 1 m, and N0 represents the noise power density of the UAV. The transmit power of mobile user k used to offload the task to the UAV for calculation in each time slot n is given by p k [n] indicates.

[0093] Furthermore, it is assumed that the UAV is able to store the unprocessed offload bits from the mobile users in its memory if the offload rate exceeds its computational capacity.

[0094] According to the assumption that the amount of user uninstalls should be less than the amount of information transmitted, the following causal relationship calculation conditions can be obtained:

[0095]

[0096] l k [n]>0 indicates that user k offloads the computational load to the drone in time slot n.

[0097] 2. MEC computing model:

[0098] In addition, each computing task T{n} is characterized by two parameters (l k [n],C k [n]). Here l k [n] represents the size of the input computing task from device k to the UAV in time slot n, C k [n] refers to the amount of computation required to complete the computational task T{n} in time slot n.

[0099] 1) The delay in completing the computing task processing includes the transmission delay of the mobile user offloading the computing task to the UAV and the computational delay of the UAV processing the computing task. The transmission delay of mobile user k offloading the computing task to the UAV is:

[0100]

[0101] In addition, the computational delay of the UAV in processing the computational task of mobile user k is:

[0102]

[0103] where f k [n] is the computing power (CPU cycles) of the UAV allocated to mobile user k to process the computing task in the nth time slot.

[0104] period / s).

[0105] 2) Transmission and computing energy consumption:

[0106]

[0107] Assume that each drone supports dynamic voltage and frequency scaling (DVFS) technology, which can dynamically adjust the CPU frequency, which can save energy consumption when the drone processes computing tasks. Set the power calculation of the CPU in time slot n to k comp represents the effective energy coefficient of the drone, which depends on the chip architecture of the MEC server equipped on the drone. Therefore, the computing energy consumption of the drone processing the computing task of mobile user k in time slot n can be expressed as:

[0108]

[0109] 3) Total cost:

[0110] Let A[n]=(a k [n]) represents a user-associated variable, where a k [n] = 1 indicates that the mobile user k is associated with the UAV in the nth time slot; otherwise a k [n] = 0. Please note that each mobile user can only offload one computing task to a drone in each time slot, and each drone can only process one computing task of an end user in the same time slot, so the following constraints need to be met:

[0111]

[0112] Based on the above analysis, the energy consumption of mobile user k completing the computing task processing in time slot n can be expressed as:

[0113]

[0114] Similarly, the delay for the system to complete the computational task processing of mobile user k in time slot n is:

[0115]

[0116] 3. Blockchain consensus model:

[0117] The consensus process of the present invention is as follows Figure 3 As shown in the figure, the total number of nodes in the blockchain system is M=4, that is, there are M drones in the entire air safety implementation method, and the fault-tolerant practical Byzantine consensus protocol (PBFT) is used. In addition to supporting fault-tolerant fault nodes, this algorithm can also tolerate malicious nodes. Malicious nodes refer to nodes that may make completely opposite replies to received messages or even forge messages.

[0118] For more fault tolerance, the maximum number of fault-tolerant nodes in the PBFT algorithm is f = (M-1) / 3, where M is the total number of nodes in the entire blockchain system. That is to say, when M is 4, it means that the blockchain network can tolerate f = 1 faulty nodes or malicious nodes. That is, even if node R3 is a faulty or malicious node (R3 in the figure only receives transactions and does not verify sent transactions), the network can still reach a consensus.

[0119] The client sends a petition to the master node, indicating the start of the request phase. Nodes in the blockchain system collect transactions from the MEC system, including encrypted user identities, computing task results, and computing offload records. The master node needs to verify the signature and message authentication code (MAC) when receiving the transaction. Assume that all nodes require β1 and β2 CPU cycles to generate or verify a signature and a MAC, respectively. Considering that = represents the maximum number of transactions stored in a block, and θ corresponds to the proportion of correct transactions. Therefore, the computational cost of the master node can be expressed as:

[0120]

[0121] When receiving the pre-prepare message from the master node, the slave node must continue to verify the signature and MAC of the transaction when receiving a new block. Then generate a signature and M-1 MACs. The computational cost of the slave node can be calculated as:

[0122] e2=β1+β2(M-1)+φ(β1+β2).

[0123] In the preparation phase, each replica node needs to send a preparation message to other nodes. The node must verify the signatures and MACs from 2f (f = (M-1) / 3) nodes (including the replica node itself), and then generate a signature and M-1 MACs. Therefore, in this process, the computational cost of the replica node can be expressed as:

[0124] e3=β1+β2(M-1)+2f(β1+β2).

[0125] In the submission phase, each node must verify the 2f signature and MAC when receiving the submission message from other nodes, and then generate a signature and MAC. Therefore, in this process, the computational cost of the node can be expressed as:

[0126] e4=(2f+1)(β1+β2).

[0127] In the reply phase, when f+1 matching commit messages are collected, the new block will become a valid block and will then be broadcast to the blockchain system. In this process, the computational cost can be expressed as:

[0128] e5=(f+1)(β1+β2).

[0129] The computing resource allocation strategy in the blockchain system is represented as F2 = {f m |m∈M}. f m It is the computing power of the blockchain nodes.

[0130] Based on the above analysis, the total delay cost of completing tasks in the blockchain system can be given by the following formula:

[0131]

[0132] Similarly, in the blockchain system, the computing power of the CPU Therefore, the total energy cost of completing a task in the blockchain system can be expressed as:

[0133]

[0134] 4. UAV propulsion energy consumption:

[0135] In the proposed system, a fixed-altitude rotary-wing UAV is considered. In practice, such a UAV flies from an initial position q0 to a final position q F During this period, its speed is subject to the maximum speed V max Therefore:

[0136] q[1]=q0

[0137] q[N]=q F

[0138]

[0139] The flight power consumption of the rotorcraft is modeled as:

[0140]

[0141] Where P0 and P i They represent the blade profile power and induced power in the hovering state respectively. Utip, v0, d0, ρ, s and A are other parameters related to the aerodynamics of the drone. In order to achieve the speed v[n] of the drone, there is a formula:

[0142]

[0143] Therefore, the flight energy consumption of the drone during this period is expressed as:

[0144]

[0145] 5. Problem analysis:

[0146] According to the analysis, the total delay cost in the entire drone network system can be expressed as:

[0147]

[0148] The total energy consumption to complete computing tasks and blockchain task processing can be expressed as:

[0149]

[0150] Where ξ is a mapping factor that ensures that the propulsion energy consumption of the drone is at the same level as the computing energy consumption of the drone and the computing energy consumption of the blockchain. Then, the present invention proposes a user uninstallation association variable A=(a k [n]), UAV trajectory optimization variables And the computing resource allocation variable F1 in the UAV-MEC layer = (f k [n]) and the computing resource allocation variable F2 in the blockchain system = {f m |m∈M}, so as to minimize the weighted sum of the energy consumption and delay of the entire system. Therefore, mathematically, the problem to be optimized can be formulated as:

[0151] P:

[0152] stC1:q[1]=q0,q[N]=q F

[0153] C2:

[0154] C3:

[0155] C4:

[0156] C5:

[0157] C6: l h，k [n]≥0,

[0158] C7:

[0159] C8:

[0160] C9:

[0161] In problem P, the objective function is the weighted sum of the energy consumption of the entire network and the total delay of completing the computing task and blockchain task processing. w1 (0≤w1≤1) is the weight factor that integrates the objective function into a whole and balances the weights of latency and energy consumption. F represents the maximum computing power of the drone. By adjusting the value of w1, the preference between energy consumption and latency can be dynamically controlled. For example, when the drone's battery is low, the value of w1 can be increased to save more computing and communication energy. At the same time, if the network requirements are sensitive to latency, the value of w1 should be reduced to reduce latency. Constraint C1 corresponds to the initial / end position of the drone, constraint C2 is the maximum speed constraint satisfied by the drone, and constraint C3 represents the speed constraint of the drone, whose speed depends on the front and back positions of the drone. Constraints C4-C5 represent the task offloading relationship between the drone and the end user. Constraints C6-C7 represent the causal relationship between the size of the offloaded task of each end user and the transmission capacity between the user and the drone. Constraint C8 represents the guarantee that the computing resources allocated to the offload task and blockchain computing task are less than the maximum computing power of the drone.

[0162] It can be seen that the optimization problem proposed by the present invention is a MINLP (Mixed-Integer Nonlinear Programming) problem, and its challenge and difficulty lies in the coupling of variables between the objective function and the constraints, which makes it more difficult to obtain the optimal solution to the problem.

[0163] 6. Optimization solution:

[0164] By analyzing the structure of the objective function and the constraints of the proposed optimization problem, it can be seen that the binary variable a k The existence of [n] makes the problem non-convex. Due to the discreteness of constraint C5, the optimization of this problem cannot adopt continuous optimization theory, and discrete optimization is not suitable for large-scale problems, which greatly increases the computational burden. Therefore, for convenience, we first convert the binary variable a k [n]∈{0,1} is relaxed to a k [n] = [0, 1], which transforms the 0-1 discrete programming problem into a continuous optimization method based on duality, thus facilitating the construction and implementation of the algorithm. Given all parameters, the optimization problem in the formula can be specifically expressed as P1, as shown below:

[0165] IPl):

[0166] stC1-C4,C6-C9 C10:

[0167] Due to the coupling of variables in the objective function and constraints, the optimization problem P1 is still a non-convex problem. In order to solve this thorny problem, the present invention considers decomposing the formulated problem into several sub-problems. For the computing resource allocation F2 of a given blockchain, the problem P1 can be rewritten as the optimization problem P2. Therefore, the task offloading scheduling optimization problem A of the MEC system, the drone trajectory optimization problem Q and the computing resource allocation optimization problem F1 can be solved alternately as three sub-problems. The specific process is as follows:

[0168] In the above optimization problem P1, the objective function contains In order to solve the non-convexity of problem P1, the present invention sets an auxiliary variable t k ={t k [n]}, so that it satisfies t k [n]≥(R k [n]) -1 Therefore, problem P1 can be modified into the following optimization problem P2:

[0169] oeLh

[0170] stC1-C4,C6-C10 C11:

[0171] 1) End user and UAV association:

[0172] For a given UVA trajectory Q, the computational resource allocation F1 of the UAV-MEC layer, and the auxiliary variable t k , the optimization problem of user association can be expressed as P2.1:

[0173] QPa.U:

[0174] st

[0175]

[0176]

[0177] The objective function and constraints of problem (P2.1) are about the variable a k is a convex function of [n]. Therefore, Problem (P2.1) is a convex problem and can be solved by a standard convex program solver such as CVX.

[0178] 2) UAV trajectory optimization:

[0179] For a given end-user association variable A and the computational resource allocation F1 of the UAV-MEC layer, the UAV trajectory optimization problem Q can be expressed as the following problem P2.2:

[0180] (P2.2):

[0181] st(C1)-(C3),(C6)-(C7),C11

[0182] The objective function is non-convex, and the constraint C7 is non-convex with respect to q[n]. To solve, the SCA technique is applied.

[0183] First, in order to solve the non-convexity of the objective function P(||v[n]||), the slack variable v is introduced. n ≥||v[n]||, so the UAV propulsion energy consumption expression can be rewritten as:

[0184]

[0185] Note that the second term in the above equation is still non-convex with respect to q[n]. Here we introduce another slack variable: The expression can be easily obtained as:

[0186]

[0187] Then for any given local point {v n，j ,u n，j} (j represents the jth iteration), the right-hand side (RHS) of the above expression can be expanded to a lower bound by a first-order Taylor expansion, because it is relative to u n and v n is jointly convex. Let Represent this lower bound function:

[0188]

[0189] Based on the above discussion, the flight energy consumption of the UAV can be approximately expressed as a convex function, namely:

[0190]

[0191] Secondly, consider the expression of RHS in constraint C7 is non-convex with respect to q[n]. However, if |q[n]-w k [n]| 2 As a whole, it can be considered as a convex expression. Therefore, for any given local point {q j [n]}, the lower bound of the RHS of the expression in C7 can be derived previously express:

[0192]

[0193] By substituting the derived lower bound function and the approximate convex expression into the problem, we can obtain the new constraint:

[0194] C12:

[0195] C13:

[0196] C14:

[0197] Since constraint C11 is non-convex, the problem is still a non-convex problem. Finally, in order to solve the non-convexity of constraint C11, the SCA algorithm is used to convert constraint C11 into a convex constraint, and an approximate solution to the problem can be obtained. By using the SCA algorithm, Theorem 1 is as follows.

[0198] Theorem 1: For any given local trajectory at the rth iteration have:

[0199]

[0200] When the equation holds, we have

[0201] According to the content of Theorem 1, constraint C11 is rewritten as constraint C15:

[0202] C15: (t k，m [n])-1≤B k，m φ lb {q m [n]}.

[0203] Therefore, the problem can be rewritten as P2.3. The new problem is convex with respect to both Q and t, indicating that the new problem is convex and can be solved by iterative optimization.

[0204] (P2.3):

[0205] stC1-C3,C6 C12-C15

[0206] 3) Computing resource allocation at the MEC-UAV layer

[0207] After obtaining user associations A, drone trajectories Q, and auxiliary variables t, the computational resource allocation optimization problem in the drone-MEC layer can be expressed as problem P2.4:

[0208] (P2.4):

[0209] St

[0210]

[0211] In summary, the objective function and constraints of problem P2.4 are convex relative to F1. Therefore, the problem is a convex optimization problem, and the optimal computing resource allocation can also be solved by CVX.

[0212] Computing resource allocation at the blockchain layer:

[0213] By solving the above problems, we can obtain the task offloading schedule A, the resource allocation F1 in the MEC-UAV system, and the UAV trajectory Q.

[0214] Therefore, the allocation of computing resources in the blockchain can be written as an optimization problem P3:

[0215] (P3):

[0216] St

[0217]

[0218] In addition, it can be observed that the objective function and the two constraints are convex with respect to F2. Therefore, the allocation of computing resources in the optimization problem of blockchain system is a convex problem.

[0219] Based on the above analysis, the convex subproblem can be solved alternatively using the CVX optimization tool.

[0220] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A secure air implementation method integrating mobile edge computing and blockchain technology, characterized in that: The following steps are involved: Step 1: Establish a system model including M drones, multiple terminal devices, and blockchain, where each UAV handles the computing offload tasks of K user terminals in its service area, and deploys a mobile edge computing (MEC) server on each UAV, so that the UAV acts as a MEC server. At the same time, the blockchain server is deployed on the UAV, so that it acts as a blockchain node. The consensus process will be completed between M UAVs (i.e., M blockchain nodes); Step 2: Each UAV acts as an aerial node, and the computing tasks of mobile users are offloaded to the computing nodes for processing. This process determines the computing energy consumption and computing latency of the UAV-MEC model system; While receiving the task from the mobile terminal, the UAV will move in trajectory, which reduces the transmission energy consumption and delay of the communication between it and the user terminal, and generates the propulsion energy consumption of the UAV; Step 3: When the UAV completes the user's computing task, it will collect the computing offloading records it generates as "transactions". The transaction (also known as blockchain task) is verified by M UAVs, and a new block will be generated to record the transaction. When more than two-thirds of the drones pass the verification, that is, the entire blockchain network reaches a consensus, the block will be added to the blockchain. This process generates the computing energy consumption and computing latency of the blockchain model. Step 4: According to the security offloading energy consumption and latency of the UAV-MEC layer and the blockchain layer, a joint optimization problem of user association, UAV trajectory, and computing resource allocation is constructed to minimize the weighted sum of latency and energy consumption of the entire aerial security computing process.

2. A secure air implementation method integrating mobile edge computing and blockchain technology according to claim 1, characterized in that: The joint optimization problem of step 4 is to first decompose the optimization problem of the UAV-MEC layer and the optimization problem of the blockchain model. The steps for solving the optimization problem of the UAV-MEC layer are as follows: first, for the impact of the task offloading scheduling of the terminal user on the delay, combined with the given trajectory and computing resource allocation, the optimal solution of the user's offloading scheduling is solved, and the optimal solution of the user task offloading scheduling is obtained. Then, combined with the given computing resource allocation, the continuous graph approximation SCA algorithm is used to optimize and solve the trajectory of the UAV, and then the previously solved user task offloading scheduling and UAV trajectory are used to optimize and solve the computing resources allocated by the edge server to the UAV to process the offloading task. The obtained user task offloading scheduling, UAV trajectory and the computing resources allocated by MEC to the drone to process the offloading task are iteratively updated to the convergence of the user task offloading scheduling, UAV trajectory and computing resources through the fast coordinate descent method, and the minimum energy consumption and delay of the MEC-UAV layer are obtained. At the same time, the computing resources required for the edge computing server to allocate to the drone to act as a MEC node have been solved in the above steps, and the computing resources required for the edge computing server to allocate to the drone to act as a blockchain node can be solved later.

3. A secure air implementation method integrating mobile edge computing and blockchain technology according to claim 1, characterized in that: The method includes: during the entire flight cycle time T of the UAV, T is evenly divided into n time slots, and the ground user terminal k uses orthogonal frequency division multiple access OFDMA transmission mode to unload information. It is assumed that the computing power of the user itself is not enough to support them to process computing tasks, but considering that the MEC server is deployed on the UAV, the computing tasks of the mobile users can be transmitted to the UAV for processing through the wireless link. It is worth noting that the UAV can only unload computing tasks for one ground user terminal in each time slot, and the task data amount that the user terminal k needs to transmit to the UAV in the nth time slot is defined as l k [n],a k [n] is a user-associated variable, indicating whether user terminal k has a computing task in the nth time slot that needs to be transferred to the UAV for offloading. The value is [0, 1]. The UAV has a computing task to be offloaded in each time slot. k [n] = 1 means that the offload task processed by the UAV in time slot n is that of end user k, k [n] = 0 means that the task processed by the UAV in time slot n is not for user k, so the amount of task data unloaded from user terminal k to the UAV is a k [n]*l k [n]; When the kth user terminal offloads the computing task to the UAV in time slot n, the channel gain between the user terminal and the UAV is: Assume that the system uses OFDMA, that is, the total available bandwidth B is divided equally into K subcarriers for all users, with a size of The transmission rate between mobile user k and UAV in the nth time slot can be expressed as: in, represents the reference received signal-to-noise ratio (SNR) of the UAV when d0 = 1 m, N0 represents the noise power density of the UAV, and the transmit power of TD k used to offload the task to the UAV for calculation in each time slot n is given by p k [n] means; According to the fact that the amount of data unloaded by user terminal k should be smaller than the amount of information transmitted, the following calculation causal relationship conditions can be obtained: q k [n]l k [n]≤δ t R k [n] where δ t is the size of each time slot, a k [n] represents the user-associated variable, l k [n] represents the amount of task data unloaded by user k to UAV in time slot n, R k [n] represents the transmission rate of user k to the UAV in time slot n; In addition, the computation task T{n} of user k is characterized by two parameters (l k [n],C k [n],l k [n] represents the amount of task data unloaded by user k to UAV in time slot n, C k [n] refers to the number of CPU cycles required to complete the computation of task T{n} in time slot n; The delay in completing the computing task processing includes the transmission delay of user terminal k offloading the computing task to UAV and the computing delay of UAV processing the computing task. The transmission delay of mobile user k offloading the computing task to UAV is: In addition, the delay of the drone processing the computing task of mobile user k is: where f k [n] represents the computing power allocated by the edge computing server to the UAV to process the offloaded tasks of user k. The transmission energy consumption of mobile user k offloading the computing tasks to the UAV is: Set the CPU power calculation for time slot n to k comp represents the effective energy coefficient of the drone, which depends on the chip architecture of the MEC server equipped on the drone. The computing energy consumption of the drone processing the computing task of mobile user k in the nth time slot can be expressed as: The energy consumption of mobile user k completing the computing task processing in time slot n can be expressed as: The delay of the system completing the computational task processing of mobile user k in time slot n is:

4. A secure air implementation method integrating mobile edge computing and blockchain technology according to claim 1, characterized in that: The method includes: in the model, a blockchain server is provided on each UAV, and the UAV also acts as a blockchain node, wherein there are M UAVs in total, that is, there are M consensus nodes in the blockchain model in total, and the Practical Byzantine Consensus Protocol (PBFT) with fault tolerance is used in this blockchain architecture, and there are only two roles of the UAV acting as a node in the PBFT algorithm, one is a UAV acting as a primary node (primary) and the other is a UAV acting as a replica node (replica), and the two roles between the UAVs can be converted to each other; The maximum number of fault-tolerant nodes of the PBFT algorithm is f = (M-1) / 3, where M is the total number of nodes in the entire blockchain system. That is to say, when M is 4, at this time, at most one drone in the drone cluster can be tolerated as a malicious node or a faulty node. At this time, even if there is a drone as a malicious node, the blockchain system can still reach a consensus, thereby protecting the security and reliability of data unloaded by the user terminal; Using the PBFT consensus protocol to reach a blockchain consensus requires four stages, namely the pre-prepare stage in which all drone nodes receive client requests, the prepare stage between drone nodes, the commit stage, and the reply stage of the drone nodes to the client. The consensus process of the blockchain system is as follows: The client sends a petition to the master node, indicating the start of the request phase. The nodes in the blockchain system collect transactions from the MEC system, including encrypted user identities, computing task results, and computing offload records. The master node needs to verify the signature and message authentication code (MAC) when receiving the transaction. Assume that all nodes require β1 and β2 CPU cycles to generate or verify a signature and a MAC, respectively. Considering that φ represents the maximum number of transactions stored in a block, and θ corresponds to the proportion of correct transactions, the computational cost of the master node can be expressed as: When receiving the pre-prepare message from the master node, the slave node must continue to verify the signature and MAC of the transaction when receiving a new block, and then generate a signature and M-1 MACs. The computational cost of the slave node can be calculated as: e2=β1+β2(M-1)+φ(β1+β2). In the preparation phase, each replica node needs to send a preparation message to other nodes. The node must verify the signatures and MACs from 2f (f = (M-1) / 3) nodes (including the replica node itself), and then generate a signature and M-1 MACs. In this process, the computational cost of the replica node can be expressed as: e3=β1+β2(M-1)+2f(β1+β2). In the submission phase, each node must verify the 2f signature and MAC when receiving the submission message from other nodes, and then generate a signature and MAC. In this process, the computational cost of the node can be expressed as: e4=(2f+1)(β1+β2). In the reply phase, when f+1 matching commit messages are collected, the new block will become a valid block and will then be broadcast to the blockchain system. During this process, the computational cost can be expressed as: e5=(f+1)(β1+β2). The computing resource allocation strategy in the blockchain system is represented as F2 = {f m |m∈M},f m It is the computing power of the blockchain node; Based on the above analysis, the total delay cost of completing tasks in the blockchain system can be given by the following formula: Similarly, in the blockchain system, the computing power of the CPU Therefore, the total energy cost of completing a task in the blockchain system can be expressed as:

5. A secure air implementation method integrating mobile edge computing and blockchain technology according to claim 1, characterized in that: The method includes: in a system, a UAV flies from an initial position to a final position, and according to the speed v[n] of the UAV during the flight, the flight power consumption of the UAV can be modeled as: Where P0 and P i They represent the profile power and induced power of the rotor of the UAV in the hovering state, and Utip, v0, d0, ρ, s and A are parameters related to the aerodynamics of the UAV. Therefore, the flight energy consumption of the UAV during this period is expressed as:

6. A secure air implementation method integrating mobile edge computing and blockchain technology according to claim 1, characterized in that: The method includes: the total delay cost in the entire drone network system can be expressed as: The total energy consumption to complete computing tasks and blockchain task processing can be expressed as: Where ξ is a mapping factor that ensures that the propulsion energy consumption of the drone is at the same level as the computing energy consumption of the drone and the computing energy consumption of the blockchain. Then, according to a secure aerial computing method integrating mobile edge computing and blockchain technology as claimed in the above claims, it is characterized in that an optimization problem with the goal of minimizing the energy consumption and delay of the entire system security computing is constructed. The formulated problem to be optimized can be specifically expressed as: s.t.C1:q[1]=q0,q[N]=q F C2： C3： C4： C5： C6:l h，k [n]≥0, C7： C8： C9： In the above optimization problem, the objective function is the weighted sum of the energy consumption of the entire network and the total delay of completing the computing task and blockchain task processing. w1 (0≤w1≤1) is the weight factor, which integrates the objective function into a whole and balances the weights of delay and energy consumption. F represents the maximum computing power of the drone. By adjusting the value of w1, the preference between energy consumption and delay can be dynamically controlled. When the battery of the drone is insufficient, the value of w1 can be increased to save more computing and communication energy consumption. At the same time, if the network requirements are sensitive to delay, then the value of w1 should be reduced to reduce delay. Constraint C1 corresponds to the initial / end position of the drone respectively. Constraint C2 is the maximum speed constraint satisfied by the drone. Constraint C3 represents the speed constraint of the drone, and its speed depends on the front and back positions of the drone. Constraints C4-C5 represent the task offloading relationship between the drone and the end user. Constraints C6-C7 represent the causal relationship between the size of the offloaded task of each end user and the transmission capacity between the user and the drone. Constraint C8 represents the computing resources allocated to offload tasks and blockchain computing tasks to ensure that they are less than the maximum computing power of the drone.

7. A secure air implementation method integrating mobile edge computing and blockchain technology according to claim 1, characterized in that: The method includes: the convex optimization problem is a mixed integer nonlinear programming MINLP problem, which combines the user associated variable A=(a k [n]), drone trajectory Computational resource allocation in the UAV-MEC layer F1 = (f k [n]) and the computing resource allocation in the blockchain system F2 = {f m |m∈M}, determine the optimal solution, decouple the above variables, decompose the whole problem into several sub-optimization problems for alternate optimization and solution, and use the designed user association and computing resource allocation algorithm based on fast coordinate descent (BCD) and the UAV trajectory optimization algorithm based on continuous graph approximation (SCA) to solve the optimal solution of the problem.

Images