A high-energy-efficiency adaptive dual-mode data acquisition method, device, equipment and medium

Abstract

This invention discloses a high-efficiency adaptive dual-mode data acquisition method, device, equipment, and medium, belonging to the field of communication technology for novel power distribution systems. The method first constructs an objective function and determines constraints with the goal of maximizing network utility, defined as the difference between network operating benefits and network operating costs. Then, a hierarchical decoupling strategy is employed to solve the objective function, including: solving for the optimal network topology under a given set of hovering points using the block coordinate descent method to minimize network operating costs; solving for the optimal computing resource allocation scheme using the Lagrange function to maximize operating benefits; and searching for the optimal set of hovering points using the number of hovering points as the action variable based on a deep reinforcement learning framework. Finally, a fixed or mobile data acquisition mode is adaptively selected based on the solution results. This invention resolves the contradiction between energy consumption and data integrity in UAV data acquisition, achieving synergistic optimization of system energy efficiency and reliability.

Description

Technical Field

[0001] This invention belongs to the field of communication technology for novel power distribution systems, and specifically relates to a high-efficiency adaptive dual-mode data acquisition method, device, equipment, and medium. Background Technology

[0002] Unmanned aerial vehicle (UAV)-assisted renewable wireless sensor network (RWSN) data acquisition technology has attracted widespread attention in recent years, primarily employing two basic modes: hover-collection and mobile-collection. In hover-collection mode, the UAV hovers at a fixed location for an extended period, receiving data transmitted by sensor nodes via multi-hop transmission. This method offers stable communication quality, but the high energy consumption of hovering UAVs severely limits their mission duration. In mobile-collection mode, the UAV flies along a predetermined trajectory, sequentially visiting multiple nodes for near-field data acquisition. While this extends coverage, the short communication windows with each node make data packet loss or incomplete transmission prone to occur, compromising reliability.

[0003] In existing technologies, the two modes mentioned above are usually independent of each other, operating with pre-set fixed strategies and lacking the ability to adaptively adjust according to the dynamic state of the network. Especially in RWSNs, nodes rely on energy harvesting for power, and their available energy fluctuates significantly over time and space. Fixed acquisition modes cannot effectively cope with such dynamic changes. Therefore, current technologies have a core problem: they cannot simultaneously guarantee the integrity of data acquisition and system energy efficiency in complex and dynamic network environments, making it difficult to achieve the optimal balance between energy consumption and reliability during UAV data acquisition. Summary of the Invention

[0004] The purpose of this invention is to provide a high-efficiency adaptive dual-mode data acquisition method, device, equipment and medium to solve the problem that the existing technology cannot simultaneously guarantee the integrity of data acquisition and system energy efficiency in complex dynamic network environments, and it is difficult to achieve the optimal balance between energy consumption and reliability in the process of UAV data acquisition.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a high-efficiency adaptive dual-mode data acquisition method, comprising: An objective function is constructed with the goal of maximizing network utility, and constraints are determined; wherein, network utility is the difference between network operating revenue and network operating cost. Based on constraints, a sampling hierarchical decoupling strategy is used to solve the objective function, including: for a given set of hovering points, solving for the optimal network topology using block coordinate descent to minimize network operating costs; for the optimal network topology, solving for the optimal computing resource allocation scheme using the Lagrangian function to maximize operating benefits; and based on a deep reinforcement learning framework, searching for the optimal number of hovering points with the number of hovering points as the action variable and network energy efficiency as the reward value, thereby determining the optimal set of hovering points. Based on the optimal hovering point set, optimal network topology, and optimal computing resource allocation scheme, the UAV is controlled to execute either a fixed data acquisition mode or a mobile data acquisition mode.

[0006] Furthermore, the optimal network topology is solved using the block coordinate descent method, including: The next-hop selection problem for each sensor node in the network is modeled as an independent subproblem; wherein, the set of available next hops for a node is determined by the set of available forward neighbors, which is constructed according to the forward transmission principle; Based on the current network topology, nodes are sorted according to their influence on network operating costs, and the next hop selection of each node is optimized in turn; this process is iterated until the network operating cost converges, thus obtaining the optimal network topology.

[0007] Furthermore, the optimal computing resource allocation scheme is solved using the Lagrange function, including: Construct a Lagrangian function that includes constraints on the allocation of computing resources; The Lagrangian function is solved using the KKT conditions to obtain the resource allocation ratio for each business that maximizes the operational benefits.

[0008] Furthermore, based on a deep reinforcement learning framework, the optimal number of hover points is searched using the number of hover points as the action variable and network efficiency as the reward value, including: The drone is considered an intelligent agent, the network energy state is considered the environmental state, the number of hovering points is considered the agent's action, and the network energy efficiency is considered the reward value. The intelligent agent learns through trial and error by interacting with the environment and outputs the number of hovering points that can obtain the maximum cumulative reward, which is the optimal number of hovering points. The environmental state is the remaining energy level of all sensor nodes and energy storage stations after completing a round of data collection, and the remaining energy level is normalized by their battery capacity.

[0009] Furthermore, an objective function is constructed with the goal of maximizing network utility, and constraints are determined, including: The objective function is expressed as:

[0010] The constraints are expressed as follows:

[0011] in, To optimize the variables of the problem; For computing resources, This represents the number of hover points; For business serial number; For a set of business functions; For network operating costs; For business Corresponding operating benefits; For the set of hovering points; For energy storage stations; This is a set of all nodes recorded in ascending order based on their remaining energy. The total number of nodes; i , j These represent the node index numbers; It is the set of sensor nodes in the network; For nodes The available forward neighbor set; Represents 0-1 variables; To be assigned to business Computing resources; This refers to the maximum operating frequency of the CPU per round; For the propulsion energy consumption of UAVs; Energy consumption for data preprocessing for UAVs; Energy consumption for UAV data reception; This represents the total battery capacity of the UAV.

[0012] Furthermore, network operating costs It is expressed as follows:

[0013]

[0014] in, A constant greater than 0; This represents the energy consumption level of each energy storage station. This refers to the energy consumption of the energy storage station in each cycle; The total energy of the energy storage station; The overall energy consumption level of the network is extremely poor. For nodes Energy consumption level after each round of work is completed.

[0015] Furthermore, the fixed data acquisition mode is as follows: the node uploads the data to the energy storage station where the UAV is stationed through self-organizing network multi-hop transmission; The mobile data acquisition mode is as follows: the drone hovers sequentially above the selected node to collect data at close range.

[0016] In a second aspect, the present invention provides a drone-assisted RWSN adaptive data acquisition device, comprising: The model building module is used to construct an objective function and determine constraints with the goal of maximizing network utility; wherein, network utility is defined as the difference between network operating revenue and network operating cost; The solution module is used to solve the objective function based on constraints and a sampling hierarchical decoupling strategy. This includes: finding the optimal network topology using block coordinate descent for a given set of hovering points to minimize network operating costs; finding the optimal computing resource allocation scheme using the Lagrangian function for the optimal network topology to maximize operational benefits; and searching for the optimal number of hovering points based on a deep reinforcement learning framework, using the number of hovering points as the action variable and network energy efficiency as the reward value, thereby determining the optimal set of hovering points. The execution module is used to control the UAV to perform either a fixed data acquisition mode or a mobile data acquisition mode based on the optimal hovering point set, the optimal network topology, and the optimal computing power resource allocation scheme.

[0017] In a third aspect, the present invention provides an electronic device including a processor and a memory, the processor being configured to execute a computer program stored in the memory to implement the high-efficiency adaptive dual-mode data acquisition method described above.

[0018] In a fourth aspect, the present invention provides a computer-readable storage medium storing at least one instruction that, when executed by a processor, implements the high-efficiency adaptive dual-mode data acquisition method described above.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows: By constructing an optimization problem aimed at maximizing network utility and employing a hierarchical decoupling strategy to solve it, the complex joint optimization problem is decomposed into three manageable sub-problems (topology optimization, resource allocation, and mode selection). Finally, the most suitable data acquisition mode is adaptively selected based on the solution results. This solves the problem of tight coupling between mode selection, path planning, and resource allocation in UAV data acquisition, achieving improved overall system energy efficiency and extended network lifetime while ensuring reliable data acquisition.

[0020] By employing a block coordinate descent method and modeling the next-hop selection of each node as an independent subproblem for iterative optimization, the computational complexity of network topology optimization is significantly reduced. This method allows the network to converge quickly to a low-cost topology based on real-time states, ensuring the real-time performance and feasibility of the algorithm in large-scale networks and avoiding the energy void problem.

[0021] By constructing a Lagrangian function and applying the KKT conditions to solve for the optimal resource allocation scheme, a globally optimal computing power resource allocation ratio can be obtained under given conditions with rigorous mathematical assurance. This ensures that limited airborne computing power resources are accurately allocated to the most profitable operations, maximizing the data processing benefits of UAVs and the utilization efficiency of computing power resources.

[0022] By utilizing a deep reinforcement learning framework, the number of hovering points is treated as an action and network energy efficiency as a reward for learning. This enables the drone to learn to make optimal decisions based on real-time network energy states without the need for complex mathematical models. This endows the system with powerful online learning and adaptive capabilities, allowing it to dynamically adapt to the spatiotemporal instability of network energy harvesting and maintain high-performance operation at all times.

[0023] The constructed objective function and constraints explicitly define the network operating cost as an expression that includes the energy consumption of energy storage stations and the energy consumption range of the network. This directly quantifies and integrates the two key objectives of extending network lifetime and balancing node energy consumption into the optimization model. The optimization process is guided to spontaneously choose solutions that reduce global energy consumption and prevent premature failure of some nodes, thereby achieving green and sustainable network operation.

[0024] Through methodological optimization, intelligent switching between the two acquisition modes was ultimately achieved at the control level. This combined the advantages of fixed data acquisition mode in terms of coverage and mobile data acquisition mode in terms of short-range communication reliability, overcoming the inherent limitations of a single mode and enabling the system to flexibly respond to different scenario requirements. Attached Figure Description

[0025] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart of a high-efficiency adaptive dual-mode data acquisition method according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the average energy consumption of each node in different environments according to embodiments of the present invention; Figure 3 This is a schematic diagram illustrating the range of average remaining network energy levels in each round under different environments according to embodiments of the present invention. Figure 4 This is a schematic diagram comparing the overall energy consumption of different algorithm networks in embodiments of the present invention; Figure 5 This is a schematic diagram showing the comparison of the range of remaining energy levels of different algorithm networks in an embodiment of the present invention; Figure 6 This is a structural block diagram of a UAV-assisted RWSN adaptive data acquisition device according to an embodiment of the present invention; Figure 7 This is a structural block diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0026] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0027] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this invention is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0028] Glossary: ​​Dual-mode data acquisition method includes two modes: fixed data acquisition and mobile data acquisition. In fixed data acquisition mode, nodes upload the collected data to a UAV (Unmanned Aerial Vehicle) via a flexible self-organizing network for aggregation, while the UAV only resides at the energy storage station to perform data preprocessing. In mobile data acquisition mode, UAVs hover above selected nodes one by one, collecting data from nearby nodes, and then return to the energy storage station to complete data preprocessing and replenish energy. After completing data acquisition, the UAV allocates available computing resources to various services based on the principle of maximizing network energy efficiency, and typically can only preprocess a portion of the data.

[0029] Example 1 like Figure 1 As shown, a high-energy-efficiency adaptive dual-mode data acquisition method includes: S1. Construct an objective function with the goal of maximizing network utility and determine the constraints; wherein, network utility is defined as the difference between network operating revenue and network operating cost; S2. Based on constraints, a sampling hierarchical decoupling strategy is used to solve the objective function, including: for a given set of hovering points, solving for the optimal network topology using the block coordinate descent method to minimize network operating costs; for the optimal network topology, solving for the optimal computing power resource allocation scheme using the Lagrangian function to maximize operating benefits; and based on a deep reinforcement learning framework, searching for the optimal number of hovering points with the number of hovering points as the action variable and network energy efficiency as the reward value, thereby determining the optimal set of hovering points. S3. Based on the optimal hovering point set, optimal network topology, and optimal computing resource allocation scheme, control the UAV to execute either a fixed data acquisition mode or a mobile data acquisition mode.

[0030] The aforementioned scheme, by constructing a function model aimed at maximizing network utility, unifies multiple conflicting optimization objectives into a comprehensive quantitative index. This avoids the limitations of traditional methods with their singular optimization objectives, guiding the solution at the system level to simultaneously consider data acquisition volume and network operational quality, achieving global optimization of network energy efficiency and reliability. Through a layered decoupling solution strategy, it overcomes the challenge of complex coupling, decomposing the three highly coupled complex variables of hovering point planning, topology construction, and resource allocation, and designing block coordinate descent, Lagrangian functions, and deep reinforcement learning for each. This reduces the problem's solution complexity, making previously difficult-to-solve non-convex optimization problems solvable. It ensures that the adaptive method possesses real-time computation and rapid response capabilities in large-scale network environments, demonstrating strong engineering practical value.

[0031] It is important to emphasize that deep reinforcement learning is used in step S2 to determine the optimal number of hovering points. The drone is treated as an intelligent agent, which can learn autonomously and adapt to the dynamic changes in the network's energy state by interacting with the environment. It can output the most efficient collection strategy in real time to improve the current network energy efficiency. This allows the system to no longer rely on preset fixed rules, but can intelligently and seamlessly switch between fixed and mobile collection modes, effectively cope with the randomness and instability of renewable energy power supply, and always maintain the optimal working state.

[0032] In summary, this invention systematically addresses the core challenges of UAV-assisted RWSN data acquisition in terms of energy efficiency, reliability, and adaptability through a unified target modeling, hierarchical solution, and intelligent decision execution technical approach, providing key technical support for realizing a green and sustainable wireless sensing network.

[0033] In one embodiment of the present invention, a high-efficiency adaptive dual-mode data acquisition method is also provided, comprising: S100. Construct an objective function with the goal of maximizing network utility and determine the constraints; wherein, network utility is defined as the difference between network operating revenue and network operating cost.

[0034] It should be noted that, theoretically, any adjacent nodes can establish a relay relationship in a network. However, to avoid reverse transmission or self-loops, nodes... The available forward and backward neighbors can be represented according to the forward transmission principle as follows:

[0035]

[0036] in, For nodes The available forward neighbor set, including nodes All possible next jumps; For energy storage stations and Center front The set of hover point numbers corresponding to each node. The set of node numbers obtained by sorting each node in ascending order of remaining energy; This represents the corresponding hover point, specifically a node. The designated spatial position above for UAV hovering; Represents a set Mid-distance nodes The nearest node's corresponding number That is, the distance node The nearest hovering point, Indicates the corresponding distance; Represents a node The set of backward neighbors, that is, the set of nodes that use it as the next hop; Represents a node For nodes The back neighbor, otherwise ; It is the set of sensor nodes in the network; j Indicates the node index number.

[0037] For a given network topology, nodes The flow load can be calculated as follows:

[0038] in, For nodes The amount of data that needs to be sent to its next hop. For each round of nodes The amount of raw data perceived. For backward nodes Send to node The amount of data.

[0039] With nodes For example, its total energy consumption for data transmission It is expressed as follows:

[0040] in, For nodes Energy consumption for sending one bit of data The energy consumption for receiving one bit of data. Energy consumption for sensing a unit bit of data.

[0041] Therefore, node Energy consumption level after each round of work is completed It can be represented as follows:

[0042] Among them, the battery capacity of each node in the network same.

[0043] UAV data reception power consumption The calculation method is as follows:

[0044] in, The energy consumption per bit of data received by a UAV. For business sets, For business Total amount of data.

[0045] Energy consumption of UAV data preprocessing The calculation method is as follows:

[0046] in, This is the power consumption factor, the value of which depends on the UAV's CPU chip architecture. The preprocessing time allocated to the UAV in each round, Representing the size of each data collection cycle and the given hover point set, respectively. The optimal flight path and speed of the UAV are calculated by a genetic algorithm. The maximum operating frequency of the CPU per round is calculated as follows:

[0047] in, This refers to the total CPU frequency of the UAV. The propulsion energy consumption of a UAV includes two parts: mobility energy consumption and hovering energy consumption. This represents the total battery capacity of the UAV.

[0048] After completing the adaptive allocation of computing resources, the business Size of preprocessed data It can be represented as follows:

[0049] in, To be assigned to business computing resources Indicates processing unit bit service Number of CPU cycles required to process the data.

[0050] In summary, business Corresponding operating benefits It can be represented as follows:

[0051] Network operating costs It is expressed as follows:

[0052]

[0053] in, A constant greater than 0 Representing energy storage stations, The energy consumption level of each energy storage station. This refers to the energy consumption of the energy storage station in each cycle. The total energy of the energy storage station. The total number of nodes. The overall energy consumption level of the network is extremely poor. For nodes Energy consumption level after each round of work is completed.

[0054] Considering three influencing factors—the number of hovering points, network topology, and UAV computing resource allocation—an optimization problem is proposed with the goal of maximizing network utility, as detailed below:

[0055] In the formula, network energy efficiency is the ratio of network operating revenue to operating costs. To optimize the variables of the problem; For computing resources, This represents the number of hover points; For business serial number; For a set of business functions; For network operating costs; For business Corresponding operating benefits; For the set of hovering points; For energy storage stations; This is a set of all nodes recorded in ascending order based on their remaining energy. The total number of nodes; i , j These represent the node index numbers; It is the set of sensor nodes in the network; For nodes The available forward neighbor set; Represents 0-1 variables; To be assigned to business Computing resources; This refers to the maximum operating frequency of the CPU per round; For the propulsion energy consumption of UAVs; Energy consumption for data preprocessing for UAVs; Energy consumption for UAV data reception; This represents the total battery capacity of the UAV.

[0056] constraint C 1 represents the set of hovering points of the UAV. For set The former The hover point above each node, Represents a fixed data acquisition pattern; constraints C 2 and C 3 indicates the selection of the data transmission path for the sensor node; constraints C 4 indicates the computing resource constraints of the UAV; constraints C 5 indicates the energy constraint of the UAV.

[0057] S200. Based on constraints, a sampling hierarchical decoupling strategy is used to solve the objective function, including: for a given set of hovering points, solving for the optimal network topology using the block coordinate descent method to minimize network operating costs; for the optimal network topology, solving for the optimal computing power resource allocation scheme using the Lagrangian function to maximize operating benefits; and based on a deep reinforcement learning framework, searching for the optimal number of hovering points with the number of hovering points as the action variable and network energy efficiency as the reward value, thereby determining the optimal set of hovering points.

[0058] It should be noted that in optimization problems (OPs), variables... It determines the generation of network topology, and thus affects network cost; The amount of data preprocessing for each service affects network revenue; while the number of hover points... The magnitude of the value affects the entire data acquisition and preprocessing process. Therefore, considering the coupling relationship between various variables in the constraints, a hierarchical decoupling method is used to solve the problem.

[0059] In one embodiment, solving for the optimal network topology using the block coordinate descent method includes: The next-hop selection problem for each sensor node in the network is modeled as an independent subproblem. The available next-hop set of a node is determined by its available forward neighbor set, which is constructed according to the forward transmission principle to avoid data back transmission or the formation of self-loops. Based on the current network topology, the next-hop selection of each node is optimized sequentially by sorting the nodes according to their influence on the network operating cost. The process is iterated until the network operating cost converges, thus obtaining the optimal network topology.

[0060] In one embodiment, solving for the optimal computing resource allocation scheme using the Lagrange function includes: Construct a Lagrangian function containing constraints on the allocation of computing resources; apply the KKT conditions to solve the Lagrangian function to obtain the computing resource allocation ratio for each business that maximizes the operational benefits.

[0061] In one embodiment, based on a deep reinforcement learning framework, the optimal number of hover points is searched using the number of hover points as the action variable and network efficiency as the reward value, including: The drone is considered an intelligent agent, the network energy state is considered the environmental state, the number of hovering points is considered the agent's action, and the network energy efficiency is considered the reward value. The intelligent agent learns through trial and error by interacting with the environment and outputs the number of hovering points that can obtain the maximum cumulative reward, which is the optimal number of hovering points. The environmental state is the remaining energy level of all sensor nodes and energy storage stations after completing a round of data collection, and the remaining energy level is normalized by their battery capacity.

[0062] As an example, network topology solving aims to address a given set of hovering points. Generating a low-cost network topology corresponds to the optimization problem. It can be represented as follows:

[0063] in, For a given The network operating cost below In optimizing the problem The variable is a constant, and the variable is the next hop of each node. A block coordinate descent method is used to quickly search for low-cost network topologies. The link states from each node to its available forward neighbor set can be constructed as corresponding "block coordinates". It can be decomposed into multiple subproblems for rapid solution. (Based on nodes) For example, its corresponding subproblems As shown below.

[0064]

[0065]

[0066]

[0067] Subproblems This is the optimal next hop problem for a single node. For a given set of hover points, only nodes Impact on network operating costs This indicates the node's ranking in ascending order based on its real-time network operating cost and influence. 1 node denoted as the number of iterations. The next hop for the remaining nodes is fixed, and the next hop selection for each node is optimized one by one based on the real-time topology. The calculation of network operating cost converges through continuous iteration, thereby obtaining the optimal topology structure.

[0068] Given a set of hover points Generating the optimal computing resource allocation scheme corresponds to the optimization problem. as follows:

[0069] in, For a given set of hover points through The calculated operational revenue of each service under the obtained network topology, and the specific influencing factors, are the maximum operating frequency. Given values, the optimization problem The computing resources allocated to each business It is strictly concave, therefore it has a unique extremum point that satisfies the KKT solution conditions, and the corresponding Lagrangian function is shown below:

[0070] in, , ,and These are the corresponding constraints. C 2 and C 3 Lagrange multipliers.

[0071] By solving and This allows for network topology reconstruction and adaptive adjustment of computing resource allocation under a given set of hovering points, thereby maximizing network energy efficiency. Based on the above two sub-problems, an optimization problem is constructed. Find the number of hovering points corresponding to the maximum network energy efficiency. Specifically, it is expressed as follows:

[0072] The solution can be achieved by treating the UAV as an agent and modeling the problem using TD3-based deep reinforcement learning. The agent and the environment interact with each other through a discrete number of hovering points. (State) To complete the first Network energy status after round of data collection This is the set of all states. Then, select state-based... Actions performed , It is the collection of all actions. As a result of an action, the agent enters the next state.

[0073] 1): State quantization

[0074] Including energy storage stations and nodes in completing the first Energy consumption after each data collection round is normalized based on the battery capacity of the energy storage station and nodes to facilitate network training.

[0075] 2): Action mapping

[0076] In the Wheel, the action of the network This corresponds to the number of UAV hover points. (Output value of the motion network) The range of values ​​is The actual output value is mapped to the number of UAV hover points using the above formula. Corresponds to a fixed data acquisition mode.

[0077] 3): Reward Calculation

[0078] The reward calculation based on the number of hover points is The optimization objective is to use the real-time energy efficiency of the network as a reward value, and finally select the hover point set with the largest corresponding reward value as the optimal hover point set, thereby achieving the same optimization effect as the objective.

[0079] In a real-world network environment, once the locations of each node are determined, a corresponding virtual environment is created on the server. The UAV learns the actions corresponding to high reward values ​​under different network environments through TD3 until the reward stabilizes. Based on the learning results, the sustainability of the RWSN is achieved through dynamic coordination of the data collection mode.

[0080] S300. Based on the optimal hovering point set, optimal network topology, and optimal computing power resource allocation scheme, control the UAV to execute a fixed data acquisition mode or a mobile data acquisition mode.

[0081] It should be noted that the fixed data acquisition mode is: the node uploads data to the energy storage station where the UAV is stationed through self-organizing network multi-hop transmission; the mobile data acquisition mode is: the UAV hovers in sequence directly above the selected node to collect data at close range.

[0082] To further explain and illustrate this scheme, a simulation example is provided below: Figures 2-5 Here is a comparison chart of specific simulation results, in which Figure 2 and Figure 3The figure shows the network energy consumption and the range of remaining energy levels for each round under different environments. Simulation results show that the network energy consumption and the range of remaining energy levels can adaptively adjust with changes in network state, thereby improving network energy efficiency while ensuring the overall balance of network energy consumption. According to the practical meaning of the reward function, as the reward increases, both energy consumption and the range should gradually decrease. Furthermore, network energy consumption increases with the number of nodes, while the range is less affected by the network environment. This is because the agent, based on the underlying solution, will consume as much energy as possible on data preprocessing, thus adaptively adjusting to maximize the reward.

[0083] Figure 4 and Figure 5 The figures show the range of network energy consumption and remaining energy levels for different algorithms with different numbers of nodes. SCCI-ADEDG is the algorithm proposed in this invention, PFDG represents the pure fixed data acquisition mode algorithm, and PMDG represents the pure mobile data acquisition mode algorithm. The remaining algorithmic details of PFDG and PMDG are the same as those of SCCI-ADEDG.

[0084] Depend on Figure 4 It can be seen that network energy consumption increases with the number of nodes. Since there is no UAV assistance, PFDG consumes the most energy, while SCCI-ADEDG needs to consider the overall balance of network energy consumption, so its corresponding energy consumption is slightly greater than that of PMDG. Figure 5 As shown, the range of remaining energy levels in SCCI-ADEDG exhibits a slow upward trend with the increase of the number of nodes, while the remaining energy level ranges of other algorithms are less affected by the number of nodes. In PFDG, UAVs reside only at energy storage stations, and the energy obtained by these stations from photovoltaic panels is sufficient to support data reception; therefore, the overall remaining energy level range of the network may be 1. Compared to PFDG, both PFDG and SCCI-ADEDG have ranges less than 0.2, demonstrating superior energy balance compared to PFDG, with SCCI-ADEDG showing a more significant advantage.

[0085] Example 2 like Figure 6 As shown, based on the same inventive concept as the above embodiments, the present invention also provides a UAV-assisted RWSN adaptive data acquisition device, comprising: The model building module is used to construct an objective function and determine constraints with the goal of maximizing network utility; wherein, network utility is defined as the difference between network operating revenue and network operating cost; The solution module is used to solve the objective function based on constraints and a sampling hierarchical decoupling strategy. This includes: finding the optimal network topology using block coordinate descent for a given set of hovering points to minimize network operating costs; finding the optimal computing resource allocation scheme using the Lagrangian function for the optimal network topology to maximize operational benefits; and searching for the optimal number of hovering points based on a deep reinforcement learning framework, using the number of hovering points as the action variable and network energy efficiency as the reward value, thereby determining the optimal set of hovering points. The execution module is used to control the UAV to perform either a fixed data acquisition mode or a mobile data acquisition mode based on the optimal hovering point set, the optimal network topology, and the optimal computing power resource allocation scheme.

[0086] Example 3 like Figure 7 As shown, the present invention also provides an electronic device 100 for implementing a high-efficiency adaptive dual-mode data acquisition method; The electronic device 100 includes a memory 101, at least one processor 102, a computer program 103 stored in the memory 101 and executable on at least one processor 102, and at least one communication bus 104.

[0087] The memory 101 can be used to store the computer program 103. The processor 102 implements the steps of the high-energy-efficiency adaptive dual-mode data acquisition method of Embodiment 1 by running or executing the computer program stored in the memory 101 and calling the data stored in the memory 101.

[0088] The memory 101 may primarily include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the electronic device 100 (such as audio data), etc. In addition, the memory 101 may include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.

[0089] At least one processor 102 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 102 may be a microprocessor or any conventional processor. Processor 102 is the control center of electronic device 100, connecting various parts of electronic device 100 via various interfaces and lines.

[0090] The memory 101 in the electronic device 100 stores multiple instructions to implement a high-energy-efficiency adaptive dual-mode data acquisition method, and the processor 102 can execute multiple instructions to achieve the following: An objective function is constructed with the goal of maximizing network utility, and constraints are determined; wherein, network utility is the difference between network operating revenue and network operating cost. Based on constraints, a sampling hierarchical decoupling strategy is used to solve the objective function, including: for a given set of hovering points, solving for the optimal network topology using block coordinate descent to minimize network operating costs; for the optimal network topology, solving for the optimal computing resource allocation scheme using the Lagrangian function to maximize operating benefits; and based on a deep reinforcement learning framework, searching for the optimal number of hovering points with the number of hovering points as the action variable and network energy efficiency as the reward value, thereby determining the optimal set of hovering points. Based on the optimal hovering point set, optimal network topology, and optimal computing resource allocation scheme, the UAV is controlled to execute either a fixed data acquisition mode or a mobile data acquisition mode.

[0091] Example 4 If the modules / units integrated in the electronic device 100 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, and read-only memory (ROM).

[0092] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0093] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0094] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0095] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0096] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0097] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A high-efficiency adaptive dual-mode data acquisition method, characterized in that, include: An objective function is constructed with the goal of maximizing network utility, and constraints are determined. The network utility is determined by network operating revenue and network operating cost. Network operating revenue is determined based on the amount of data preprocessed by each service. The amount of data preprocessed by each service is determined based on the UAV computing power resources allocated to the corresponding service and the number of CPU cycles required to process a unit bit of service data. Network operating cost is determined based on the energy consumption level of the energy storage station in each round, the energy consumption level of the sensor nodes in the network after each round of work, and the range of the overall network energy consumption level determined by the energy consumption levels of the energy storage station and the sensor nodes in each round of work. The constraints include hovering point set constraints, sensor node data transmission path selection constraints, UAV computing power resource constraints, and UAV energy constraints. The UAV energy constraints are determined by the UAV propulsion energy consumption, the UAV data preprocessing energy consumption, the UAV data reception energy consumption, and the total battery capacity of the UAV. Based on the constraints, a hierarchical decoupling strategy is adopted to solve the objective function, including: for a given set of hovering points, solving for the optimal network topology using block coordinate descent to minimize the network operating cost; for the optimal network topology, solving for the optimal computing resource allocation scheme using the Lagrangian function to maximize the network operating benefit; and based on a deep reinforcement learning framework, searching for the optimal number of hovering points with the number of hovering points as the action variable and network energy efficiency as the reward value, thereby determining the optimal set of hovering points. Based on the optimal number of hovering points, the optimal set of hovering points, the optimal network topology, and the optimal computing power resource allocation scheme, the UAV is controlled to execute either a fixed data acquisition mode or a mobile data acquisition mode, and the acquired data is preprocessed according to the optimal computing power resource allocation scheme. Specifically, when the optimal number of hovering points is 0, the fixed data acquisition mode is executed, allowing the nodes to upload data to the energy storage station where the UAV is stationed via self-organizing network multi-hop transmission; when the optimal number of hovering points is greater than 0, the mobile data acquisition mode is executed, allowing the UAV to hover sequentially above the selected nodes for close-range data acquisition.

2. The high-efficiency adaptive dual-mode data acquisition method according to claim 1, characterized in that, Solving for the optimal network topology using the block coordinate descent method includes: The next-hop selection problem for each sensor node in the network is modeled as an independent subproblem; wherein, the set of available next hops for a node is determined by the set of available forward neighbors, which is constructed according to the forward transmission principle; Based on the current network topology, nodes are sorted according to their influence on network operating costs, and the next hop selection of each node is optimized in turn; this process is iterated until the network operating cost converges, thus obtaining the optimal network topology.

3. The high-efficiency adaptive dual-mode data acquisition method according to claim 1, characterized in that, Solving for the optimal computing resource allocation scheme using the Lagrange function includes: Construct a Lagrangian function that includes constraints on the allocation of computing resources; The Lagrangian function is solved using the KKT conditions to obtain the resource allocation ratio for each business that maximizes the operational benefits.

4. The high-efficiency adaptive dual-mode data acquisition method according to claim 1, characterized in that, Based on a deep reinforcement learning framework, the optimal number of hover points is searched using the number of hover points as the action variable and network efficiency as the reward value, including: The drone is considered an intelligent agent, the network energy state is considered the environmental state, the number of hovering points is considered the agent's action, and the network energy efficiency is considered the reward value. The intelligent agent learns through trial and error by interacting with the environment and outputs the number of hovering points that can obtain the maximum cumulative reward, which is the optimal number of hovering points. The environmental state is the remaining energy level of all sensor nodes and energy storage stations after completing a round of data collection, and the remaining energy level is normalized by their battery capacity.

5. The high-efficiency adaptive dual-mode data acquisition method according to claim 1, characterized in that, The objective function is constructed with the goal of maximizing network utility, and the constraints are determined, including: The objective function is expressed as: The constraints are expressed as follows: in, To optimize the variables of the problem; For computing resources, This represents the number of hover points; For business serial number; For a set of business functions; For network operating costs; For business Corresponding operating benefits; For the set of hovering points; For energy storage stations; This is a set of all nodes recorded in ascending order based on their remaining energy. The total number of nodes; i , j These represent the node index numbers; It is the set of sensor nodes in the network; For nodes The available forward neighbor set; Represents 0-1 variables; To be assigned to business Computing resources; This refers to the maximum operating frequency of the CPU per round; For the propulsion energy consumption of UAVs; Energy consumption for data preprocessing for UAVs; Energy consumption for UAV data reception; This represents the total battery capacity of the UAV.

6. The high-efficiency adaptive dual-mode data acquisition method according to claim 5, characterized in that, Network operating costs It is expressed as follows: in, A constant greater than 0; This represents the energy consumption level of each energy storage station. This refers to the energy consumption of the energy storage station in each cycle; The total energy of the energy storage station; The overall energy consumption level of the network is extremely poor. For nodes Energy consumption level after each round of work is completed.

7. The high-efficiency adaptive dual-mode data acquisition method according to claim 1, characterized in that, The optimal hovering point set includes the energy storage station and hovering points above the first n nodes arranged in ascending order based on remaining energy, where n is the number of optimal hovering points; in the fixed data acquisition mode, the UAV stays at the energy storage station to complete data preprocessing; in the mobile data acquisition mode, the UAV hovers sequentially above the selected node to collect data from nearby nodes, and then returns to the energy storage station to complete data preprocessing and replenish energy.

8. A UAV-assisted RWSN adaptive data acquisition device, characterized in that, include: The model building module is used to construct an objective function and determine constraints with the goal of maximizing network utility. The network utility is determined by network operating revenue and network operating cost. The network operating revenue is determined based on the amount of data preprocessed by each service, which is determined based on the UAV computing power resources allocated to the corresponding service and the number of CPU cycles required to process a unit bit of service data. The network operating cost is determined based on the energy consumption level of the energy storage station in each round, the energy consumption level of the sensor nodes in the network after each round of work, and the range of the overall network energy consumption level determined by the energy consumption levels of the energy storage station and the sensor nodes in each round of work. The constraints include hovering point set constraints, sensor node data transmission path selection constraints, UAV computing power resource constraints, and UAV energy constraints. The UAV energy constraints are determined by the UAV propulsion energy consumption, the UAV data preprocessing energy consumption, the UAV data reception energy consumption, and the total battery capacity of the UAV. The solution module is used to solve the objective function based on the constraints and employing a hierarchical decoupling strategy. This includes: finding the optimal network topology using block coordinate descent for a given set of hovering points to minimize the network operating cost; finding the optimal computing resource allocation scheme using the Lagrangian function for the optimal network topology to maximize the network operating benefit; and searching for the optimal number of hovering points based on a deep reinforcement learning framework, using the number of hovering points as the action variable and network energy efficiency as the reward value, thereby determining the optimal set of hovering points. The execution module is used to control the UAV to execute a fixed data acquisition mode or a mobile data acquisition mode based on the optimal number of hovering points, the optimal set of hovering points, the optimal network topology, and the optimal computing power resource allocation scheme, and to preprocess the acquired data according to the optimal computing power resource allocation scheme. Specifically, when the optimal number of hovering points is 0, the fixed data acquisition mode is executed, so that the nodes upload data to the energy storage station where the UAV is stationed through self-organizing network multi-hop transmission; when the optimal number of hovering points is greater than 0, the mobile data acquisition mode is executed, so that the UAV hovers sequentially above the selected nodes to collect data at close range.

9. An electronic device, characterized in that, It includes a processor and a memory, the processor being used to execute a computer program stored in the memory to implement the high-energy-efficiency adaptive dual-mode data acquisition method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one instruction, which, when executed by a processor, implements the high-efficiency adaptive dual-mode data acquisition method as described in any one of claims 1 to 7.

Images