An edge unloading method based on blockchain and wireless charging

Abstract

To address the energy constraints of mobile devices and ensure the integrity of critical data during the offloading process, this invention constructs an edge offloading method powered by blockchain and wireless charging, incorporating blockchain technology. Based on this method, task offloading and resource allocation strategies are jointly optimized, and a distributed parallel iterative algorithm based on convex optimization (DPIA-CO) is creatively designed to solve the problem. Furthermore, considering the energy consumption and latency impact of blockchain technology, this invention designs a RAFT (RAFT-RRLB) consensus mechanism based on resource reputation and load balancing. By sensing the remaining resources and load status of MEC nodes, it elects resource-rich nodes as leaders to achieve consensus, improving the resource utilization of the mobile edge computing system and reducing energy consumption and latency.

Description

Technical Field

[0001] This invention relates to an edge offloading model and method based on blockchain and wireless charging, belonging to the field of mobile edge computing. Background Technology

[0002] With the rapid development of IoT technology and the continuous advancement of communication technology, Mobile Edge Computing (MEC), as an emerging computing paradigm, is gradually becoming an effective way to solve the problem of limited computing power on mobile devices. MEC technology allows mobile devices to offload computing tasks to MEC servers at the edge, thereby achieving efficient utilization of computing resources and improving task processing efficiency.

[0003] However, with the increasing number of mobile devices and the growing complexity of application scenarios, the challenges facing MEC networks are becoming increasingly apparent. Firstly, mobile devices are often small in size and have limited resources such as power and computing power, making it impractical to install large-capacity batteries or frequently charge and replace them. If a device runs out of power, it will be unable to provide continuous service, leading to service interruptions and impacting user experience and the reliability of critical applications. Simultaneously, ensuring the immutability and security of critical information generated by the MEC network is a pressing issue. Blockchain technology, as a distributed ledger technology, offers a new solution to data security and trust issues due to its decentralized, immutable, and traceable characteristics. By applying blockchain technology to MEC networks, secure storage and transmission of critical information can be achieved, ensuring the integrity and credibility of information, thereby improving the security and reliability of the MEC network.

[0004] However, in blockchain-enabled MEC systems, MEC servers not only need to handle their own tasks but also those of the blockchain system, such as block generation and consensus execution, which makes model construction more complex. Furthermore, while ensuring the security of critical data, blockchain also increases system energy consumption. Therefore, it is necessary to balance the security of blockchain with its energy consumption impact and jointly optimize both the blockchain system and the MEC system.

[0005] In summary, this invention studies an edge offloading method based on blockchain and wireless charging. It also considers the energy consumption and latency introduced by the blockchain consensus mechanism, proposing a blockchain-enabled Wireless Power Transfer based Mobile Edge Computing (WPT-MEC) model. Furthermore, it designs a RAFT (RAFT based on Resource Reputation and Load Balancing) blockchain consensus mechanism based on resource reputation and load balancing to achieve secure resource transactions. While ensuring all tasks complete within a specified latency, the invention jointly optimizes task offloading, resource allocation, and bandwidth allocation strategies. A Distributed Parallel Iterative Algorithm based on Convex Optimization (DPIA-CO) is used to solve the energy consumption minimization problem of the blockchain and WPT-MEC models. Summary of the Invention

[0006] To address the issues of device energy constraints and the immutability and traceability of critical data in MEC scenarios, this invention proposes a blockchain-enabled WPT-MEC computation offloading model and designs a RAFT-RRLB blockchain consensus mechanism to reduce overall system energy consumption and latency.

[0007] This invention discloses an edge offloading method based on blockchain and wireless charging. The construction of this method includes the following steps:

[0008] Step 1: System Model Construction. This invention constructs a mobile edge computing model in a wireless charging scenario combined with blockchain. The MEC server can not only handle tasks offloaded from mobile devices, but also act as a wireless hybrid power source to provide energy to mobile devices. This hybrid power source is supported by renewable energy sources, such as wind and solar power, and electricity from the power grid. Simultaneously, the MEC server also assumes the responsibilities of a blockchain network node, responsible for collecting and processing transactions generated from WPT-MEC, and mining key information from these WPT-MEC systems into blocks for consensus. The mobile device integrates an EH unit, which captures radio frequency signal energy from the MEC server and converts it from electrical energy into chemical energy, storing it in its battery cell to provide energy for subsequent transmission and computing tasks.

[0009] Step 2: Construct the task offloading model. Within a time slot, each device generates one task, which is then completely offloaded to the MEC server for computation. Triples are used. Let L represent the task that needs to be unloaded from device n, where L n ρ n and These represent the task data size, computation density, and maximum tolerable latency, respectively. When device n generates a task and sends an offload request to the network, the MEC servers in the network will decide whether to join the cluster of device n.

[0010] Step 3: Design the RAFT-RRLB consensus mechanism. Since the leader in a blockchain consensus mechanism is responsible for generating and replicating data blocks, it consumes more resources compared to other followers, thus requiring careful leader selection. However, in the traditional RAFT mechanism, the leader is randomly assigned, which is clearly not optimal. Furthermore, RAFT is prone to having multiple candidates with the same number of votes during the election phase, leading to election failure. All nodes need to wait a long time to restart the election process, significantly increasing consensus overhead. Therefore, this invention proposes a RAFT based on resource reputation and load balancing to improve the performance of traditional RAFT.

[0011] Step 4: Construct a computational unloading problem model. This includes the following steps:

[0012] Step 4.1: Construct the MEC server computing model.

[0013] Since users divide tasks proportionally and transmit them to the corresponding MECs in the cluster for processing, for user device n, the computing tasks are offloaded to the MEC. m The data transmission delay at that time is shown in the following formula:

[0014]

[0015] When the MEC server receives a task, it will consume its own computing resources to process the task. The data processing latency when user device n offloads its task to the MEC is shown in the following formula:

[0016]

[0017] The energy consumption of a device is positively correlated with its operating time. When user equipment n sends data to MEC... m The energy consumption of uplink transmission during data transmission is shown in the following formula:

[0018]

[0019] Where, ∈ c (J / ms) is the transmission energy consumption coefficient.

[0020] Similarly, MEC mThe calculated energy consumption when processing the unloading task of device n is shown in the following formula:

[0021]

[0022] Where, ∈ p (J / ms) represents the calculated energy consumption coefficient.

[0023] Therefore, in MEC server computing mode, the total transmission energy consumption of the device and the total computing energy consumption of the server are expressed as follows:

[0024]

[0025] Step 4.2: Construct a local computing model.

[0026] In local computing mode, the computing power of user device n is F. n This represents the computing resources available for local task processing. The device can decide whether to process the task locally or offload it to the MEC cluster for processing. The latency for local task processing is shown in the following formula:

[0027]

[0028] Meanwhile, the energy consumption of local processing tasks is shown in the following formula:

[0029]

[0030] Among them, P l This represents the device's power consumption when processing tasks locally.

[0031] Step 4.3: Construct a blockchain consensus mechanism model.

[0032] In the election phase of RAFT-RRLB, the leader needs to perform (2+D) i (M-1) status confirmation communications, including (M-1) vote requests, (M-1) election confirmations, and D i (M-1) heartbeats. In addition, the leader needs to perform (M-1) block repetitions during the block verification phase.

[0033] It is worth noting that MEC requires bandwidth and energy for each data transmission, and according to the bandwidth allocation strategy mentioned earlier, the bandwidth for each state acknowledgment signal transmission is equal to W. l s / (2+Di ) Therefore, the energy consumption for the leader to transmit the status acknowledgment signal is as follows:

[0034]

[0035] Similarly, the energy consumption of the leader replica block can be expressed as:

[0036]

[0037] The energy consumption of the data transmission leader is shown in the following formula:

[0038]

[0039] In addition, the leader also needs to consume computing resources to generate blocks. The computational latency required for the leader to perform block generation is shown in the following formula:

[0040]

[0041] The energy consumption for the leader to generate this block is shown in the following formula:

[0042]

[0043] During the consensus process, each follower needs to send a voting response, D. i There are one heartbeat response and one block acknowledgment response, therefore there are a total of (D) i +2) confirmation transmissions.

[0044] The transmission delay of the follower is shown in the following formula:

[0045]

[0046] The energy consumption of data transmission by the follower is shown in the following formula:

[0047]

[0048] Therefore, the consensus on MEC m Energy consumption is expressed as:

[0049]

[0050] In summary, in a blockchain-based wireless charging MEC system, the total energy consumption includes the transmission energy consumption of all users, the computing task energy consumption, and the computing task energy consumption and consensus energy consumption of the MEC base station. The total energy consumption is shown in the following formula:

[0051]

[0052] During the uninstallation process, for user devices The task will be processed locally or in parallel by the MECs in the MEC cluster. Therefore, the offload latency for user n is the maximum of the service latencies of all MECs in the MEC cluster and the local service latency, as shown in the following formula:

[0053]

[0054] in, Local service is delayed.

[0055] MEC's ​​unloading decision is The computing resource allocation and bandwidth allocation for all MECs are denoted as follows: and The energy minimization problem in offloading and consensus is shown in the following equation:

[0056]

[0057] (a) indicates that the unloading delay cannot exceed the task delay threshold; (b) and (c) indicate that the computing resources and bandwidth allocated to the MEC cannot exceed its maximum computing capacity and bandwidth; (d) indicates that the decision variable is a continuous variable in [0,1]; (e) indicates that the task unloaded by each user can be processed in parallel by the MEC cluster or processed locally; and (f) indicates that the remaining energy of each device is greater than the minimum energy required for device operation but less than the maximum energy required for device storage.

[0058] Step 5: Solve the problem using a distributed parallel iterative algorithm based on convex optimization. To address the computational offloading problem mentioned above, this invention decouples the non-convex problem into a convex problem by introducing new variables, and then solves it using a distributed convex optimization parallel iterative algorithm based on blockchain. Each MEC node independently processes its local sub-problem, performing optimization calculations to solve the specific sub-problem based on the allocated data and constraints. After solving its own sub-problem, each MEC node shares the results or partial information with other MEC nodes through the blockchain network. This sharing process relies on the blockchain's consensus mechanism; for example, distributed consensus algorithms (such as Byzantine fault tolerance, PoW, PoS, etc.) may be used to ensure the consistency and correctness of information among nodes in the network. In this invention, an improved RAFT is used for consensus. Once a blockchain node has collected complete global information, it updates its local variables based on this information.

[0059] The beneficial effects of this invention are: by combining wireless power transfer technology with mobile edge computing, it provides a continuous and reliable power supply to terminal devices through wireless channels, solving problems such as service interruptions and data loss caused by insufficient device power, and further extending the life cycle of MEC. Simultaneously, this invention considers the traceability and immutability of key information in WPT-MEC, and constructs a WPT-MEC computation offloading model incorporating blockchain technology. Furthermore, it designs a RAFT (RAFT based on Resource Reputation and LoadBalancing, RAFT-RRLB) consensus mechanism based on resource reputation and load balancing. By sensing the remaining resources and load status of MEC nodes, it elects resource-rich nodes as leaders to achieve consensus, thereby improving the resource utilization rate of WPT-MEC. Attached Figure Description

[0060] Figure 1 Model diagram of edge offloading methods empowered by blockchain and wireless charging

[0061] Figure 2 Flowchart of the RAFT-RRLB consensus mechanism

[0062] Figure 3 A comparison chart of energy consumption performance of different solutions

[0063] Figure 4 A comparison chart of latency performance of different solutions

[0064] Figure 5 A comparison chart of task failure rates for different solutions. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0066] It should be noted that, in order to avoid unnecessary details from obscuring the important parts of the present invention, the accompanying drawings only show structures or processing steps closely related to the present invention, while omitting other details that are not closely related to the present invention.

[0067] Additionally, it should be noted that the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0068] Referring to the accompanying drawings, the steps in a specific embodiment are as follows:

[0069] Step 1: Construct a system model. For example... Figure 1 The system model diagram shown below assumes that there are M MEC base stations in the system model, and the set of base stations is represented as follows: Each base station is equipped with X antennas (X > N). The base stations in the model are heterogeneous, with each base station having different bandwidth and computing power, denoted by W. m (MHz) and C m (CPU cycle frequency) is the representation. There are N users in the scenario, and the user set is represented as... Each user will be served by a specific MEC cluster (i.e., a collection of several base stations). There are two types of wireless links in the model: Wireless link 1 between the MEC server and the device, and Wireless link 2 between MEC servers. Wireless link 1 is used for data transmission and wireless charging, while Wireless link 2 supports communication and data exchange between base stations. The performance of the wireless links is affected by factors such as bandwidth, channel conditions, and transmission power.

[0070] Step 2: Construct the task offloading model. The offloading decision for device n is constructed as a continuous variable a. m，n ∈[0,1], also known as MEC m Clustering decision. If a m，n =0, MEC m It will not be added to the MEC cluster of device n; if a m，n ∈(0,1],MEC m The device will be added to the MEC cluster of device n and will undertake a corresponding proportion of offloading tasks. MEC clustering decision provides a flexible task offloading mechanism that dynamically adjusts the degree of association between the device and the MEC cluster based on the heterogeneity of the MEC base station, thereby achieving optimal utilization of computing resources.

[0071] Φ n A MEC cluster defined as device n, consisting of Φ n The equipment provided by Ω n Devices that share the same MEC cluster as device n are called intra-cluster devices, while the remaining devices are called inter-cluster devices.

[0072] When MEC server m provides services to device n, the interference signals of devices within the cluster are as follows:

[0073]

[0074] in, g represents the transmitted signal power of other devices v. m，nx represents the channel coefficient between device v and MEC server m. v This indicates the signal sent by device v.

[0075] When MEC server m provides services to device n, the inter-cluster device interference signal is as follows:

[0076]

[0077] Based on the two formulas above, the total signal received by MEC server m when providing edge services to device n includes the useful signal from device n and the interference signals from devices within and outside the cluster. The total signal received by MEC server m when providing edge services to device n is shown in the following formula:

[0078]

[0079] In the formula, Let n be the transmitted signal power. For device n and MEC m The channel coefficients between them, z is Gaussian white noise with a mean of 0 and a variance of σ², The complex signal sent by device n.

[0080] Beamforming technology can completely eliminate interference with devices within the cluster. The beamforming vector for device n is calculated as follows:

[0081]

[0082] in

[0083] Assuming MEC m The computing resources allocated to device n are c. m,n The bandwidth is w m,n c m,n and w m,n All are continuous variables and cannot exceed MEC. m The computing power and bandwidth of MEC m Resources will not be allocated to devices that are not within the service range.

[0084] To allocate bandwidth and ensure that devices can send signals to different MECs within the MEC cluster, this invention employs the Orthogonal Frequency Division Multiple Access (OFDMA) method. When MECs... m When providing bandwidth for transmitting task data to device n, the uplink transmission rate of device n is as follows:

[0085]

[0086] After the MEC cluster completes processing of the device's tasks, the results are integrated and returned to the device. After verifying the returned data, the device pays the MEC cluster based on the amount of resources provided. Since the downlink data is relatively small, its overhead is ignored.

[0087] In the WPT-MEC model, the server not only processes task data from devices but also acts as a stable wireless power source, supplying power to each mobile device. Each device receives wireless power from the MEC server, stores the electricity, and powers its transmission, computation, and other operations. The power supply channel conditions are derived based on the Rayleigh fading channel model. Among them, A d =4.11, representing the antenna gain, f c =915MHz represents the carrier frequency, and γ represents the path loss exponent. d m,n It is the distance between the mobile device and the base station, α i It is an independent random channel fading factor that follows a unit-mean exponential distribution. Then, the energy obtained by device n through wireless power supply in each time slot... for:

[0088]

[0089] in, P represents the energy harvesting efficiency, and P represents the transmit power of the MEC when wirelessly powered.

[0090] Step 3: Design the RAFT-RRLB consensus mechanism. Figure 2 The process of the RAFT-RRLB consensus mechanism is demonstrated. RAFT-RRLB is more suitable for resource-constrained MEC scenarios, as it can dynamically adjust the priority of leader election and monitor node resource utilization. and The leader election weights are adjusted, and a load balancing mechanism is introduced to monitor the number of times a blockchain node is elected as the leader during the consensus process. When the load on certain nodes in the system is high, the probability of these nodes becoming leaders is reduced to avoid nodes being elected as leaders too frequently, thus ensuring system resource balance and stability.

[0091] In the improved RAFT consensus mechanism, MEC m After completing the computational task, the remaining computing resources and bandwidth are denoted as follows: and During the consensus process, leaders and followers need to send state confirmation signals (such as heartbeat packets, vote requests, confirmation replies, etc.) and transmit them via Wireless Link 2. To simplify the RAFT-RRLB communication model, assume the data size of the state confirmation is L. s Let L b and D i These represent the block size and generation interval, respectively. During the consensus phase, the leader requires significantly more bandwidth and computing resources than the followers. Therefore, a MEC server with abundant resources can be selected as the leader. This invention uses a variable called resource reputation to describe MEC. m Communication and computational capabilities during the consensus process. Assuming all remaining MECs... m If all resources are used for consensus, then MEC m Resource reputation is defined as:

[0092]

[0093] Meanwhile, the load status of the MEC server that is elected as the leader becomes a significant factor affecting the quality of the solution. If MEC m Frequent leader elections increase latency during consensus tasks, impacting user service quality. Based on this consideration, a load balancing operator is designed to address MEC (Multi-Level Capability) issues during the consensus process. m The problem is that the load is too high.

[0094] Evaluation of MEC m The method for determining node load balancing is shown in the following formula:

[0095]

[0096] Where m represents the MEC server number, Grade m For MEC m The balance level, num m For MEC m The number of times the leader is elected, num_avg is the average number of times the leader is elected in the current MEC cluster, and α1, α2 (α1 < α2) are the load balancing level thresholds.

[0097] According to MEC m The load balancing level determines its load balancing operator, as shown in the following formula:

[0098]

[0099] The mechanism of this invention sets three load balancing operators under different balancing levels, namely η1, η2, η3, where 0 < η3 < η2 < η1 < 1.

[0100] Simultaneously, a probability-based model is used to describe the behavior of MEC in the improved RAFT mechanism, and MEC... m The probability of becoming the leader in an election is defined as P:

[0101]

[0102] Where R is the maximum reputation value of all MECs, i.e.

[0103] In the improved RAFT, the priority of leader election can be dynamically adjusted, and the resource utilization of sensing nodes can be monitored. and RAFT adjusts the leader election weights and introduces a load balancing mechanism to monitor the number of times a blockchain node is elected as the leader during the consensus process. When the load on certain nodes in the system is high, RAFT reduces the probability of these nodes becoming leaders, preventing nodes from being elected as leaders too frequently, thus ensuring system stability and reducing system latency.

[0104] Since all MECs will become either a leader or a follower after the election, P(m = Follower) = 1 - P(m = Leader). Next, the transmission process between the leader and followers is modeled. It is assumed that the backhaul channels between MECs and the access channels between MECs and users use different spectrums. Therefore, a MEC only needs to consider interference from other MECs. (Assumptions...) For MEC m The transmitted signal power, For the interference signal power of other MECs, For MEC m1 and MEC m2 The channel coefficients between them.

[0105] can be Generate, where γ is the path loss exponent. For MEC m1 With MEC m2 The distance between them. Considering single-hop communication between MECs, calculate the distance between MECs. m1 and MEC m2 The signal-to-noise ratio between them is shown in the following formula:

[0106]

[0107] In a probability-based model, to calculate the transmission overhead of a follower, when a specific follower sends a signal to the leader, the average signal-to-noise ratio (SNR) between the follower and all other consensus nodes is used as the SNR of that follower i with respect to other nodes, as shown in the following formula:

[0108]

[0109] It is worth noting that the size of the state confirmation data is much smaller than the size of the consensus block data. To utilize limited bandwidth as efficiently as possible, this invention assumes that the MEC server transmits two different types of data at a fixed bandwidth based on their data size. Therefore, the bandwidth allocation strategy can be expressed as follows:

[0110]

[0111] in, MEC m A fixed bandwidth is used to transmit status confirmation signals and blocks. When the leader receives a block confirmation signal from a follower, the block is allowed and appended to the blockchain. All transaction information in this block is completely transparent and cannot be altered.

[0112] Step 4: Construct a computation offloading problem model. Solve for the total energy consumption and latency in MEC server mode, the total energy consumption and latency in local computing mode, and the total energy consumption and latency of the consensus mechanism.

[0113] Step 5: Solve the problem using a distributed parallel iterative algorithm based on convex optimization. First, by introducing new variables to decouple the non-convex problem into a convex problem, a distributed parallel iterative algorithm based on convex optimization is used to solve it. This invention designs a distributed parallel iterative algorithm based on convex optimization to solve the above problem. To reduce complexity, this framework decomposes the original problem into M identical subproblems, and each MEC server can solve one subproblem in parallel.

[0114] To verify the effectiveness of the proposed model and the RAFT-RRLB consensus mechanism, this invention designed three schemes for comparative experiments:

[0115] (1) Binary Computing Offloading (BCO): To verify the effectiveness of the partial offloading scheme compared to the traditional binary offloading scheme, in this scheme, each user offloads computing tasks to a single MEC. Therefore, the clustering decision variable A satisfies...

[0116] (2) Separate Optimization (SO): In order to verify the necessity of joint optimization of the offloading process and consensus process, the optimization target of this scheme only includes the offloading energy consumption of all MECs and users.

[0117] (3) RAFT-based Offloading (RO): In order to verify the effectiveness of the improved RAFT consensus mechanism proposed in this invention, this invention selected the traditional RAFT consensus algorithm for comparison. Therefore, the probability of each node becoming the leader is completely random.

[0118] (4) RAFT-RRLB based on resource reputation and load balancing: Joint optimization is performed based on the RAFT-RRLB consensus mechanism proposed in this invention.

[0119] like Figure 3 As shown, total energy consumption increases with the number of user devices. For different numbers of devices, the proposed scheme has lower total energy consumption than SO, BCO, and RO. When N is 5, 6, and 7, the total energy consumption of the proposed scheme is similar to that of the RO scheme, but as the number of devices increases, the energy consumption of the proposed scheme is significantly lower than that of the RO scheme, and gradually becomes better. Through the RAFT-RRLB consensus mechanism and joint optimization of computation offloading strategy and energy consumption during the consensus process, lower total energy consumption is achieved compared to other schemes such as SO, BCO, and RO, even with an increasing number of user devices. This is particularly evident in environments with dense user devices, demonstrating significant energy efficiency advantages, which is especially important for WPT-MEC applications with strict energy consumption requirements.

[0120] like Figure 4 As shown, the total latency increases with the number of user devices. The solution proposed in this invention is resource-aware and simultaneously optimizes offloading energy consumption and consensus energy consumption. The solution of this invention can be applied to latency-sensitive applications while maintaining low latency.

[0121] like Figure 5 As shown, the task offloading failure rate increases with the number of user devices. This trend is because more devices generate a higher volume of task requests, increasing the energy consumption for transmitting and processing tasks, thus raising the likelihood of task offloading failure. Compared to the above solutions, the solution proposed in this invention replenishes the energy of user devices by introducing wireless power transmission technology, ensuring sufficient energy for devices to perform tasks, thereby significantly reducing the task offloading failure rate. Furthermore, the RAFT-RRLB consensus mechanism is employed to effectively manage the resource utilization and load status of the MEC server, maintaining a low task offloading failure rate even with a significant increase in the number of users.

[0122] In summary, this invention constructs an edge offloading method based on blockchain and wireless charging. It deeply considers the energy consumption and latency impacts of combining blockchain technology, proposes a blockchain consensus mechanism based on resource reputation and load balancing, jointly optimizes task offloading strategies, resource allocation strategies, and bandwidth allocation strategies, and utilizes a distributed parallel iterative algorithm based on convex optimization for solution. Finally, simulation experiments are conducted to analyze energy consumption performance, latency performance, and failure offloading rate performance, verifying the effectiveness and superiority of the proposed method and the RAFT-RRLB consensus mechanism.

[0123] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An edge offloading method based on blockchain and wireless charging, characterized by comprising the following steps: Step 1: Constructing a System Model: In the wireless charging scenario, a mobile edge computing (MEC) model incorporating blockchain is constructed. The MEC server not only handles device offloading tasks but also acts as a wireless hybrid power source to provide energy to the devices. The hybrid power source is supported by both renewable energy and the power grid. Simultaneously, the MEC server also acts as a blockchain network node, responsible for collecting and processing transactions generated in the wireless power transfer-based mobile edge computing (WPT-MEC) system, and mining these key information into blocks for consensus. The devices integrate an energy harvesting (EH) unit to capture radio frequency signal energy from the MEC server and convert it into electrical energy stored in their battery cells to provide energy for subsequent transmission and computing tasks. Step 2: Construct a task offloading model, and construct the offloading decision of device n as a continuous variable a. m,n ∈[0,1], also known as the m-th MEC server MEC m Clustering decision; if a m,n =0, MEC m It will not be added to the MEC cluster of device n; if a m,n ∈(0,1], MEC m It will be added to the MEC cluster of device n and will undertake the corresponding proportion of offloading tasks; Step 3: Design a RAFT (RAFT based on Resource Reputation and Load Balancing) consensus mechanism; RAFT-RRLB improves consensus performance by sensing the remaining resources and load status of MEC servers and electing resource-rich nodes as leaders. Step 4: Construct a computation offloading problem model to solve for the total energy consumption and latency in MEC server computing mode, the total energy consumption and latency in local computing mode, and the total energy consumption and latency of the consensus mechanism. Step 5: Solve the problem using a distributed parallel iterative algorithm based on convex optimization.

2. The edge offloading method based on blockchain and wireless charging as described in claim 1, characterized in that: Build a task unloading model; use triples Let L represent the task that needs to be unloaded from device n, where L n ρ n and These represent the task data size, computation density, and maximum tolerable latency, respectively. When device n generates a task and sends an offload request to the network, the MEC server in the network will decide whether to join the MEC cluster of device n.

3. The edge offloading method based on blockchain and wireless charging as described in claim 1, characterized in that: A new computational unloading problem model was constructed, and the specific steps are as follows: Step 4.1: Construct the MEC server computing model; Since users proportionally divide tasks and transmit them to the corresponding MEC servers in the MEC cluster for processing, for device n, the computing tasks are offloaded to the MEC. m The data transmission delay at that time is shown in the following formula: Where, r m,n It is the uplink transmission rate of device n; When the MEC server receives a task, it will consume its own computing resources to process the task; the data processing latency when device n offloads its task to the MEC server is shown in the following formula: Among them, c m,n It is MEC m The computing resources allocated to device n; The energy consumption of a device is positively correlated with its operating time. When device n moves to MEC m The energy consumption of uplink transmission during data transmission is shown in the following formula: Where, ∈ c (J / ms) is the transmission energy consumption coefficient; Similarly, MEC m The calculated energy consumption when processing the unloading task of device n is shown in the following formula: Where, ∈ p (J / ms) represents the calculated energy consumption coefficient; Therefore, in MEC server computing mode, the total computing energy consumption of the server and the total transmission energy consumption of the device are expressed as follows: in, It is a set of N devices; It is a collection of M MEC servers; Step 4.2: Construct a local computing model; In local computing mode, the computing power of device n is F. n This indicates the computing resources available for local task processing; the device can decide whether to process the task locally or offload it to the MEC cluster for processing. The latency for local task processing is shown in the following formula: The energy consumption for local task processing is shown in the following formula: Among them, P l This represents the device's power consumption when processing tasks locally. Step 4.3: Construct a blockchain consensus mechanism model; In the election phase of RAFT-RRLB, the leader needs to perform (2+D) i (M-1) status confirmation communications, including (M-1) vote requests, (M-1) election confirmations, and D i (M-1) heartbeats; in addition, the leader needs to perform (M-1) block repetitions during the block verification phase; D i It is the block generation interval; The MEC server requires bandwidth and energy for each data transmission. Therefore, the total energy consumption for data transmission should equal the sum of the energy consumed in each transmission. The bandwidth for each status acknowledgment signal transmission is equal to... Therefore, the energy consumption for the leader to transmit the status confirmation signal is as follows: in, The fixed bandwidth used by the leader to transmit status acknowledgment signals. It is the signal-to-noise ratio (SNR) between the leader MEC server and other MEC servers, L s It is the data size of the status confirmation signal; Similarly, the energy consumption of the leader replica block can be expressed as: in, L is the fixed bandwidth used by the leader to transmit blocks. b It refers to the size of the block; The energy consumption of the data transmission leader is shown in the following formula: In addition, the leader also needs to consume computing resources to generate blocks. The computing latency required for the leader to perform block generation is shown in the following formula; in The remaining computing resources for the leader; The energy consumption for the leader to generate this block is shown in the following formula: During the consensus process, each follower needs to send a voting response, D. i There are one heartbeat response and one block acknowledgment response, therefore there are a total of (D) i +2) confirmation transmissions; The transmission delay of the follower is shown in the following formula: in, It is MEC m The remaining bandwidth after completing the computational task. It is the average signal-to-noise ratio between the follower and all other consensus nodes; The energy consumption of data transmission by the follower is shown in the following formula: Therefore, the consensus on MEC m Energy consumption is expressed as: Where P(m=Leader) is MEC m The probability of becoming the leader in an election is P(m=Follower)=1-P(m=Leader); In summary, in a blockchain-based wireless charging MEC system, the total energy consumption includes the transmission energy consumption of all users, the computing task energy consumption, and the computing task energy consumption and consensus energy consumption of the MEC server; the total energy consumption is shown in the following formula: During the unloading process, for device n, Its task will be processed locally or in parallel by the MEC servers in the MEC cluster. Therefore, the offload latency of device n is the maximum of the service latencies of all MECs in the MEC cluster and the local service latency, as described by the following formula: in, Delayed local service; MEC's ​​unloading decision is The computing resource allocation and bandwidth allocation for all MECs are denoted as C = and The energy minimization problem in unloading and consensus is as follows: (a) indicates that the unloading delay cannot exceed the task delay threshold; (b) and (c) indicate that the computing resources and bandwidth allocated to the MEC cannot exceed its maximum computing capacity and bandwidth; (d) indicates that the decision variable is a continuous variable in [0,1]; (e) indicates that the task unloaded by each user can be processed in parallel by the MEC cluster or processed locally; and (f) indicates that the remaining energy of each device is greater than the minimum energy required for device operation but less than the maximum energy required for device storage.

4. The edge offloading method based on blockchain and wireless charging as described in claim 1, characterized in that: The solution is obtained using a distributed parallel iterative algorithm based on convex optimization. To address the computational offloading problem, a new variable is introduced to decouple the non-convex problem into a convex one, and then a distributed convex optimization parallel iterative algorithm based on blockchain is used for solving the problem. Each MEC server independently processes its local sub-problems, performing optimization calculations to solve specific sub-problems based on the allocated data and constraints. After solving its own sub-problem, each MEC server shares the results or partial information with other MEC servers through the blockchain network. This sharing process relies on the blockchain's consensus mechanism to ensure the consistency and correctness of information among nodes in the network. This invention uses an improved RAFT for consensus. Once a blockchain node has collected complete global information, it updates its local variables based on this information.

Images