Optimization method for computing unloading and content caching based on air STAR-RIS assistance of wireless sensor network

Abstract

The invention relates to an optimization method for computing unloading and content caching based on air STAR-RIS assistance of a wireless sensor network, and the method comprises the steps: constructing a system model, and carrying out the distribution and distribution of sensor equipment, UAV and edge clouds in the network: constructing a network model, a communication model, a computing model and a caching model; designing a UAV energy harvesting mode, the edge cloud having a radio frequency energy transmitter, the UAV having a rechargeable battery, an energy harvesting circuit assembly, and a computing server; a path-finding algorithm of an intelligent task request is designed, and after energy is collected, when the UAV caches the task, a corresponding node directly returns a task result; when the calculation task is not cached, the task is subjected to parallel calculation on the sensor, the UAV and the edge cloud, and a result is finally returned; and constructing a total energy consumption equation, determining the total energy consumption minimization of the system through joint optimization of energy collection time, caching, calculation of unloading, UAV hovering position and STAR-RIS passive beam forming variables, and determining that continuous charging and endurance are maintained after the UAV executes a task each time.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and more specifically to an optimization method for computing offloading and content caching assisted by STAR-RIS in the air of a wireless sensor network. Background Art

[0002] Wireless Sensor Networks (WSNs) consist of a large number of sensor nodes and have demonstrated impressive capabilities in efficiently transmitting large amounts of data. Their compactness, cost-effectiveness, and ease of deployment make WSNs an efficient choice for a variety of real-time applications. With the rapid development of WSNs, the explosive growth of sensor devices has intensified the demand for high data rates and ultra-low latency services. Traditional cloud computing has difficulty meeting the needs of these delay-sensitive and computationally intensive tasks due to high transmission latency and low resource utilization. Compared with traditional cloud computing, Mobile Edge Computing (MEC) provides computing and caching services to mobile devices by deploying servers at the edge of the network, avoiding the need to send data to a remote cloud center, thereby reducing the cost of long-distance transmission. Studies have shown that by deploying MEC servers at access points or base stations, not only can popular content be cached on edge nodes to reduce the latency and energy consumption of sensor content requests, but also tasks that require computing can be offloaded.

[0003] Although MEC can provide low-cost computing and caching services for nearby sensors, MEC is usually deployed on fixed base stations, which limits its adaptability to sensor requirements, coverage, and service quality. To overcome these limitations, UAV (Unmanned Aerial Vehicle) communication networks have become important due to their mobility, LoS links, and rapid deployment capabilities. UAVs equipped with MEC servers can provide flexible computing and caching services and dynamically provide 3C resources. However, due to the limited endurance of UAVs, it is difficult to meet the needs of high-energy consumption tasks. Therefore, continuous low-power solutions are crucial to improving the performance of UAV systems based on wireless sensor networks. Summary of the invention

[0004] To solve the above-mentioned shortcomings, the present invention provides an optimization method for computing offloading and content caching assisted by STAR-RIS in the air based on a wireless sensor network, including: constructing a wireless sensor system model, wherein the system model is based on the distribution of sensor devices, UAVs with wireless rechargeable batteries and edge clouds in the sensor network, as well as the allocation of cache, computing and communication resources, including: constructing a wireless sensor network model, a communication model, a computing model and a caching model; designing a UAV energy collection mode, wherein the edge cloud has a radio frequency energy transmitter for broadcasting wireless energy to the UAV, and the UAV has a rechargeable battery, an energy collection circuit component and a computing server, and the rechargeable battery is used to store the collected energy for its operation; power; design a pathfinding algorithm for intelligent task requests. After the UAV has collected energy, when the UAV caches the task, the corresponding node directly returns the result of the task to the sensor; when the UAV does not cache the computing task, the task is calculated in parallel on the sensor, UAV and edge cloud, and finally the result of the computing task is returned to the sensor device; based on the wireless sensor system model, UAV energy collection mode and the pathfinding algorithm for task requests, the total energy consumption equation of the wireless sensor system is constructed, and the total energy consumption of the system is minimized by jointly optimizing the energy collection time variable, cache variable, calculation offloading variable, UAV hovering position variable and STAR-RIS passive beamforming variable, and it is determined that the UAV maintains continuous charging and endurance after each task execution.

[0005] Optionally, a wireless sensor network model is constructed, wherein the network model includes sensor devices, UAVs with wireless rechargeable batteries and edge cloud distribution, and STAR-RIS, wherein the sensors are connected to the UAVs and edge cloud via wireless links, and signals are transmitted between the sensors and edge cloud via STAR-RIS as relay nodes; a communication model is constructed, wherein the communication channel between the UAVs, STAR-RIS and edge cloud is set as a LoS link, and a channel gain model is constructed based on the characteristics of the LoS link to describe the energy consumption of signal transmission during the communication process; a computing model is constructed, wherein the computing model is at least used to describe the uplink communication energy consumption and the computing energy consumption of each network node when the task request is not cached by the wireless sensor system; a cache model is constructed, wherein a task cache strategy is designed to store completed partial tasks and data in the UAV or edge cloud, and when the sensor sends the same computing task request again, if the UAV or edge cloud has cached the task and data, the computing result is directly returned to the sensor from the cache, thereby reducing task processing delay and energy consumption.

[0006] Optionally, the total energy consumption equation of the wireless sensor system satisfies the following formula:

[0007]

[0008] The total energy consumption of the STAR-RIS-assisted wireless sensor system can be minimized as follows:

[0009]

[0010] Among them, T h is the energy harvesting time variable, Y is the cache variable, α is the computation offloading variable, and r u is the UAV hovering position variable, and Θ is the STAR-RIS passive beamforming variable.

[0011] Optionally, the optimization of the cache variables includes: dynamically optimizing the cache variables according to a DRL algorithm, and dynamically adjusting the cache content according to content popularity, historical access frequency and network topology.

[0012] Optionally, the path-finding algorithm based on the wireless sensor system model, UAV energy collection mode and task request constructs a total energy consumption equation for the wireless sensor system, and minimizes the total energy consumption of the system by jointly optimizing energy collection time variables, cache variables, computational offloading variables, UAV hovering position variables and STAR-RIS passive beamforming variables, and ensures that the UAV maintains continuous charging and endurance after each mission, including: optimizing the relative position of the UAV and the sensor, and the transmission and reflection coefficient matrix of STAR-RIS through the SCA algorithm.

[0013] Optionally, the computational offloading variable dynamically adjusts the offloading ratio of computing tasks among sensor devices, UAVs with wireless rechargeable batteries, and edge clouds through an SCA algorithm to achieve a balance between energy consumption and processing efficiency, and keep the UAV continuously charged and on-the-go.

[0014] The embodiments of the present invention improve the endurance of unmanned aerial vehicles (UAVs) by introducing wireless power transmission technology; at the same time, the STAR-RIS technology is introduced to further reduce the UAV relay energy consumption, assist sensor nodes in forwarding tasks, and reasonably allocate UAV-based WSN resources. In addition, by installing STAR-RIS on the UAV, the system can dynamically adjust the relative position between STAR-RIS and the sensor to further improve the efficiency of task transmission and processing. The present invention formulates an energy consumption minimization problem to jointly optimize energy collection time variables, cache variables, computational offloading variables, UAV hovering position variables, and STAR-RIS passive beamforming variables. The practicality and effectiveness of the present invention have been verified in wireless sensor networks, successfully solving the problem of minimizing the energy consumption of UAV systems without wireless sensor networks, and can be widely used in smart cities, environmental monitoring, UAV network communications and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The following are drawings of exemplary embodiments of the present invention, and the same or similar reference numerals are used in the drawings to represent the same or similar elements. In the drawings:

[0016] Figure 1 A flow chart of an optimization method for computation offloading and content caching assisted by STAR-RIS in the air of a wireless sensor network according to an exemplary embodiment of the present invention is shown.

[0017] Figure 2 A flowchart of constructing each sub-model of a wireless sensor system model according to an exemplary embodiment of the present invention is shown.

[0018] Figure 3 A schematic diagram of a STAR-RIS assisted wireless sensor system model according to an exemplary embodiment of the present invention is shown. DETAILED DESCRIPTION

[0019] In order to better explain the present invention and facilitate understanding, the present invention is described in detail below through specific implementation modes in conjunction with the accompanying drawings.

[0020] In the present invention, the term "and / or" is intended to cover all possible combinations and sub-combinations of the listed elements, including any one, any sub-combination or all of the elements listed separately, without excluding other elements. Unless otherwise specified, the terms "first", "second", etc. are used to describe various elements without intending to limit the positional relationship, timing relationship or importance relationship of these elements. Such terms are only used to distinguish one element from another. Unless otherwise specified, the orientation or positional relationship indicated by the terms "front, back, up, down, left, right", etc. is usually based on the orientation or positional relationship shown in the accompanying drawings, which is only for the convenience and simplification of description, and cannot be understood as limiting the scope of protection of the present invention.

[0021] Since UAVs have limited endurance and are unable to meet the needs of high-energy tasks, continuous low-power solutions are essential for improving the performance of UAV-based WSNs. To enhance the endurance of UAVs, RF energy transmitters can be integrated on the edge cloud side to broadcast energy to UAVs through wireless power transmission technology. At the same time, in order to further reduce the overall energy consumption of UAV-based WSNs, reconfigurable intelligent surfaces (STAR-RIS) that can transmit and reflect simultaneously are introduced to replace the traditional signal relay mode of UAVs. However, there are still key research gaps in the existing technology. First, the fixed deployment of STAR-RIS is not adaptable enough in the dynamic WSN environment, resulting in poor task offloading transmission performance. Second, even if STAR-RIS is installed on UAVs and wireless power transmission technology is used to achieve UAV energy reduction and continuous endurance, the original computing and caching capabilities of UAVs are ignored, resulting in low WSN resource utilization and reduced system efficiency. Finally, the relevant research mainly lacks a comprehensive joint optimization framework covering caching decisions, offloading strategies, UAV hovering positions, and STAR-RIS passive beamforming, which makes it difficult to fully utilize the capabilities of UAV-based WSNs in dynamic and heterogeneous resource environments.

[0022] In view of the above problems, an embodiment of the present invention proposes an aerial STAR-RIS-assisted computing offloading and content caching framework for wireless sensor networks. The framework significantly reduces the overall energy consumption of UAV-based WSNs by jointly optimizing the caching strategy, task offloading strategy, UAV hovering position and STAR-RIS passive beamforming. In this framework, the UAV has computing and caching functions, supplemented by wireless power transmission technology to improve its endurance; at the same time, STAR-RIS is introduced as a passive relay node to optimize the task transmission path and improve task processing efficiency. By dynamically adjusting the relative position of STAR-RIS and sensor nodes, the task transmission performance and system adaptability are further enhanced.

[0023] The present invention addresses the problems of high energy consumption and low resource utilization efficiency caused by the surge in the number of sensor nodes and the dynamic changes in complex scenarios in wireless sensor networks. To this end, a computational offloading and content caching mechanism based on wireless sensor network airborne STAR-RIS is proposed to reduce the system energy consumption of UAV-based WSN. Among them, a wireless sensor network energy consumption minimization problem is formulated to jointly optimize energy collection decisions (energy collection time variables), content caching decisions (caching variables), computational offloading decisions (computational offloading variables), UAV hovering positions (UAV hovering position variables) and STAR-RIS passive beamforming (STAR-RIS passive beamforming variables), and wireless power transmission technology is used to ensure that the UAV can achieve continuous charging and endurance after each mission. The embodiment of the present invention divides the problem through the block coordinate descent method, and at the same time refers to an optimization algorithm that combines deep reinforcement learning (DRL) with successive convex approximation (SCA), which can adaptively cope with dynamic network environments and achieve approximately optimal decision results with low-complexity iterations, thereby making full use of the heterogeneous resources of wireless sensor networks.

[0024] The embodiments of the present invention improve the endurance of unmanned aerial vehicles (UAVs) by introducing wireless power transmission technology; at the same time, the STAR-RIS technology is introduced to further reduce the UAV relay energy consumption, assist sensor nodes in forwarding tasks, and reasonably allocate UAV-based WSN resources. In addition, by installing STAR-RIS on the UAV, the system can dynamically adjust the relative position between STAR-RIS and the sensor to further improve the efficiency of task transmission and processing. The practicality and effectiveness of the embodiments of the present invention have been verified in wireless sensor networks, successfully solving the problem of minimizing WSN energy consumption, and can be widely used in smart cities, environmental monitoring, UAV network communications and other fields.

[0025] The present invention proposes an optimization method for aerial STAR-RIS-assisted computing offloading and content caching for wireless sensor networks. First, by introducing wireless power transmission technology, it is ensured that UAV (unmanned aerial vehicle) can achieve continuous charging and endurance after performing high-energy consumption tasks. Then, a wireless sensor network model with STAR-RIS installed on UAV is designed, and the difference in task transmission performance between fixed-position UAV and flexible mobile UAV is simulated and compared, and the advantage of flexible mobility is verified. Secondly, for the content caching problem, the present invention adopts the DRL method to dynamically adjust the cache strategy according to content popularity, historical access frequency and network topology, improve the cache hit rate and reduce the system task processing delay, and compare the scheme based on the intelligent cache strategy with the traditional non-caching scheme. Then, a task offloading optimization scheme based on the SCA method is proposed, which reduces the task computing energy consumption by adjusting the distribution (offloading) ratio of tasks between sensors, UAVs and edge clouds, and verifies the advantage of partial offloading by simulating and comparing the difference in computing task performance between partial offloading and full offloading. Finally, for the optimization of the hovering position of the UAV and the STAR-RIS resource allocation problem, the present invention uses the SCA algorithm to optimize the relative position of the UAV and the sensor and the transmission and reflection coefficient matrix of STAR-RIS, thereby improving the task transmission efficiency and reducing the total energy consumption of the system. The simulation results show that the optimization framework proposed in the present invention significantly reduces the total energy consumption of the system in scenarios with limited resources and poor channel conditions, and achieves efficient resource utilization and task processing efficiency in a dynamic network environment.

[0026] Figure 1 A flow chart of an optimization method for STAR-RIS-assisted computation offloading and content caching based on an airborne wireless sensor network according to an exemplary embodiment of the present invention is shown.

[0027] S102: Construct a wireless sensor system model, wherein the system model includes: constructing a wireless sensor network model, a communication model, a computing model, and a cache model according to the distribution of sensor devices, UAVs with wireless rechargeable batteries, and edge clouds in the sensor network, as well as the allocation of cache, computing, and communication resources. The present invention first analyzes the distribution of sensor devices, UAVs with wireless rechargeable batteries, and edge clouds in the airborne STAR-RIS-assisted WSN, as well as the allocation of cache, computing, and communication resources, and then determines the wireless sensor network model, communication model, computing model, and cache model. The details are as follows:

[0028] Figure 2 A flowchart of constructing each sub-model of a wireless sensor system model according to an exemplary embodiment of the present invention is shown.

[0029] S202: Construct a wireless sensor network model, wherein the network model includes sensor devices, UAVs with wireless rechargeable batteries and edge cloud distribution, and STAR-RIS, wherein the sensors are connected to the UAVs and edge cloud via wireless links, and signals are transmitted between the sensors and the edge cloud via STAR-RIS as a relay node.

[0030] like Figure 3 As shown in the figure, the system model includes sensor devices, UAVs with wireless rechargeable batteries, edge clouds, and STAR-RIS. The sensors are connected to the UAVs and edge clouds via wireless links. Since the direct communication link between the sensors and the edge cloud is blocked, STAR-RIS is required as a relay node for signal transmission. Among them, only the edge cloud and UAV have both caching and computing capabilities, and the sensors only have computing capabilities. The caching and computing capabilities of the edge cloud, UAV, and sensors are assumed to be limited, and only part of the content in the network can be cached and calculated. For the mobility problem of sensors, all sensors are assumed to change randomly between each time slot, and for the number of requests sent by the sensors, the number of request packets actually sent by each sensor in each time slot is assumed to be fixed.

[0031] S204: Build a communication model, set the communication channel between the UAV, STAR-RIS and the edge cloud as a LoS link, and build a channel gain model based on the characteristics of the LoS link to describe the energy consumption of signal transmission during the communication process.

[0032] A three-dimensional Cartesian coordinate system is established to describe the locations of sensor devices, UAV, STAR-RIS, and edge cloud. The locations of sensor devices and edge cloud are represented by vector r m =(x m ,y m ,0) and vector r c =(x c ,y c ,0) description. It is assumed that the UAV hovers at a fixed position in each time slot to provide computing services for the sensor. The position of the UAV is r u =(x u ,y u ,h u ), the first (m c ,m r ) The positions of the STAR-RIS elements are

[0033] Between the sensor device and the UAV, between the sensor device and the (m c ,m r ) STAR-RIS elements and between the (m c ,m rThe communication channels between each STAR-RIS element and the edge cloud are set as LoS links, which all follow the free space path loss model. Therefore, the channel gain between node n and another node n′ can be expressed as follows:

[0034]

[0035] Where g0 is the power received at a distance of 1m and the transmission power is 1W. Therefore, the transmission rate from node n to node n′ can be expressed as:

[0036]

[0037] Among them, P n is the transmit power of node n. Assume that the noise power has a constant variance Among them, B n,n′ is the transmission bandwidth of the wireless link from node n to node n′.

[0038] S206: Constructing a calculation model, where the calculation model is at least used to describe the uplink communication energy consumption and the calculation energy consumption of each network node when the task request is not cached by the wireless sensor system.

[0039] Defining integer cache decision variables Indicates whether task k is cached on node j′. Therefore, the task caching strategy can be expressed as follows: For the task k offloading problem of sensor m, define the decision variables is the task offloading ratio of sensor m at node j.

[0040] Will Indicates that task k is not cached, and the offloading decision variable is Therefore, when task k is not cached, the total uplink communication energy consumption of sensor m to offload task k can be calculated as

[0041]

[0042] Will Indicates that task k is not cached, and the offloading decision variable is Therefore, the total computational energy consumption of sensor m for computing task request k can be expressed as

[0043]

[0044] Among them, k j is the effective capacitance coefficient of node j, which depends on the chip architecture of the processor.

[0045] S208: Build a cache model and design a task cache strategy to store completed partial tasks and data in the UAV or edge cloud. When the sensor sends the same computing task request again, if the UAV or edge cloud has cached the task and data, the calculation result is directly returned to the sensor from the cache, thereby reducing task processing delay and energy consumption.

[0046] Task caching involves storing completed tasks and related data in the UAV or edge cloud. Specifically, an independent resource container is maintained on the UAV or edge cloud. The caching process operates as follows: First, the sensor device sends a computing task request. If the UAV or edge cloud has already cached this task, the corresponding node notifies the sensor device of its availability on the cache server. Therefore, the sensor device can avoid offloading the same task to the UAV or edge cloud. Finally, after the UAV or edge cloud has processed the task, the result will be sent back to the sensor device. This caching mechanism reduces the need for redundant task offloading, thereby reducing sensor energy consumption and minimizing offloading delays.

[0047] Introducing an integer cache decision variable to indicate whether task k is cached on node j′. Therefore, the computation cache strategy can be defined as In this paper, we evaluate the task duration and energy consumption on UAV or edge cloud with and without task cache. Task duration is simplified to processing delay, using It can be expressed as

[0048]

[0049] The main cache energy consumption occurs in the UAV or edge cloud, and the sensor device does not incur any cache energy cost. Therefore, the energy consumption associated with task caching can be expressed as

[0050]

[0051] Among them, k j′ is the effective capacitance coefficient of node j′, which depends on the chip architecture of the processor.

[0052] S104: Designing a UAV energy collection mode, the edge cloud has a radio frequency energy transmitter for broadcasting wireless energy to the UAV, and the UAV has a rechargeable battery, an energy collection circuit component and a computing server, the rechargeable battery is used to store the collected energy to power its operation.

[0053] In the UAV energy harvesting mode, following the "collect first, then send" protocol, in each time slot, the UAV first collects energy from the RF signal sent by the edge cloud, and then uses the collected energy to process the offload task, thereby avoiding consuming the energy in its battery and extending the operation time of the UAV; in order to avoid consuming battery power and thus extending the operation time of the UAV, it is assumed that the energy consumed by the UAV in processing the offload task is less than or equal to the energy collected from the transmitted RF signal. represents the energy harvesting duration of the UAV during the EH phase, and P c represents the transmit power of the edge cloud. Therefore, the harvested energy of UAV u during the EH phase can be calculated as:

[0054]

[0055] where h c,u is the channel power gain between the edge cloud c and UAV u, and η (0≤η≤1) represents the energy conversion efficiency.

[0056] S106: Design a path-finding algorithm for intelligent task requests. After the UAV has collected energy, when the UAV caches the task, the corresponding node directly returns the result of the task to the sensor; when the UAV does not cache the computing task, the task is calculated in parallel on the sensor, UAV and edge cloud, and finally the result of the computing task is returned to the sensor device.

[0057] By analyzing the pathfinding process of task requests, as mentioned above, the routing strategy of this model is designed as follows: First, the sensor device sends a computing task request. If the UAV or edge cloud has already cached this task, the corresponding node will notify the sensor of its availability on the cache server. Therefore, the sensor can avoid offloading the same task to the UAV or edge cloud. Finally, after the UAV or edge cloud has processed the task, the result will be sent back to the sensor. When the computing task is not cached in the system, a partial offloading scheme is considered for latency-sensitive computing tasks in the STAR-RIS-assisted UAV system. This computing offloading model allows tasks to be computed in parallel on sensors, UAVs, and edge clouds. Tasks processed on sensor devices are called local tasks, tasks processed on UAVs are called UAV offload tasks, and tasks offloaded to edge clouds are called edge cloud offload tasks. In the local execution phase, the sensor processes its tasks through the local computing server. In the UAV task offloading phase, some of the remaining tasks are uploaded to the UAV and processed by the UAV's computing server. In the edge cloud task offloading phase, some of the remaining tasks are forwarded to the edge cloud for processing through STAR-RIS elements. When the task is completed, the calculation results of the UAV and the edge cloud are returned to the sensor. In downlink communication, since the UAV and the edge cloud often have higher transmission power and the calculation results are usually small, the download time is relatively negligible during the UAV task offloading stage and the edge cloud task offloading stage.

[0058] S108: Based on the wireless sensor system model, the UAV energy collection mode and the path-finding algorithm of the task request, a total energy consumption equation of the wireless sensor system is constructed, and the total energy consumption of the system is minimized by jointly optimizing the energy collection time variable, the cache variable, the calculation offloading variable, the UAV hovering position variable and the STAR-RIS passive beamforming variable, and ensuring that the UAV maintains continuous charging and endurance after each mission.

[0059] In the present invention, the total energy consumption model is divided into two parts for energy consumption analysis. The main energy consumption occurs in the UAV or edge cloud, and the sensor does not consume any energy. The main energy consumption is simplified to processing energy consumption, which is For task cache The main energy consumption includes the total uplink communication energy consumption of sensor m to offload task k and the total computational energy consumption of computing task request k.

[0060] Construct a total energy consumption model for satisfying a task request

[0061] Based on the above task communication, calculation, and caching process, the overall energy consumption of sensor device m from sending task request k to obtaining the calculation result is expressed as:

[0062]

[0063] The optimization objectives of optimal caching, unloading, UAV hovering position and transmission and reflection coefficient matrix decisions of STAR-RIS are given.

[0064] Based on the above analysis, the present invention formulates the system energy consumption minimization problem by jointly optimizing the cache decision, offloading decision, UAV hovering position and passive beamforming in the STAR-RIS assisted wireless sensor system. According to formula (8), the optimization problem of minimizing the total energy consumption of the STAR-RIS assisted wireless sensor system can be expressed as:

[0065]

[0066]

[0067] Among the above constraints, constraint (9b) ensures that the task offloading rate of each sensor must be between 0 (indicating no offloading) and 1 (indicating complete offloading), thereby ensuring that the offloading tasks are properly distributed. Constraint (9c) ensures that the sum of the offloading rates of task k on all offloading targets (sensor m, UAV u, and edge cloud c) is equal to 1. Constraint (9d) ensures that it controls the offloading rate of task k on the mth sensor. c ,m r ) STAR-RIS element, constraining it to the range [0,1]. Constraint (9e) defines the magnitude response of the (m c ,m r ) STAR-RIS elements, ensuring that it is in the range [0,2π). This is because the phase shift is a periodic quantity and its valid range should be between 0 and 2π. Constraint (9f) restricts the coordinates x of the UAV to u and u Do not exceed the maximum allowed and Ensure that the UAV operates within the defined area. Constraint (9g) ensures that the total cache content on each node does not exceed its maximum cache capacity O j′ Constraint (9h) ensures that the total computing resources allocated to processing tasks on each node does not exceed its maximum computing capacity f j Constraint (9i) limits the caching of binary decision variables The value is 0 or 1. This is usually used to indicate whether the task is cached to a specific node. Constraint (9j) ensures that the completion time of task k does not exceed the maximum tolerable time of the task This ensures that all tasks are completed within the allowed time window, thus ensuring that the latency requirement is met. Constraints (9k) and (9l) ensure that the energy consumed for caching and computation at the node UAV is assumed to be less than or equal to the energy it harvests during the EH phase.

[0068] Analysis and comparison

[0069] 1. Analyze the impact of energy harvesting on system performance

[0070] It is recognized that in UAV-assisted wireless sensor networks, energy harvesting plays a vital role in optimizing system performance. Through wireless energy transmission, the edge cloud can provide continuous energy supply to UAV nodes to avoid performance degradation and data loss caused by battery exhaustion. Therefore, designing an efficient energy harvesting strategy is crucial to optimize system energy consumption and task execution efficiency. Therefore, this study proposes an intelligent energy harvesting strategy based on SCA. Different from the traditional static energy transmission strategy, this strategy dynamically adjusts the energy transmission plan according to the energy demand of UAV in the network topology and the energy consumption of the task, thereby adapting to changes in network status.

[0071] In the simulation, the impact of energy harvesting strategy and no energy harvesting strategy on system performance is analyzed:

[0072] 1) Energy collection strategy optimization: A dynamic energy collection strategy based on node energy demand and task energy consumption is adopted to transmit energy to UAV nodes first, ensuring that the task is completed without depleting the UAV node battery, thereby improving the task execution rate and stability of the system.

[0073] 2) No energy harvesting strategy: All tasks are processed through traditional energy supply methods without using energy harvesting strategies, which increases the energy consumption of nodes and task delays.

[0074] Experimental results show that the solution based on intelligent energy harvesting strategy can significantly improve system performance. Considering energy harvesting:

[0075] 1) Reduction of system energy consumption: The WPT strategy significantly reduces the overall energy consumption of the system by reducing energy waste and frequent charging of sensor nodes.

[0076] 2) Delay reduction: By providing continuous and stable energy support, task interruption and reprocessing caused by energy exhaustion are reduced, thereby reducing the delay in task completion.

[0077] 3) Improved resource utilization: By dynamically adjusting the energy transmission strategy, the resources of UAVs and nodes in wireless sensor networks are used more efficiently, avoiding energy surplus or shortage.

[0078] Simulations further verify that this energy harvesting strategy has significant advantages when network resources are limited or channel conditions are poor, such as optimized energy consumption and reduced task delays.

[0079] 2. Analyze the impact of content caching on system performance

[0080] It is recognized that content caching plays a key role in optimizing system performance in UAV-assisted wireless sensor networks. By caching popular content to UAVs and edge cloud nodes, the transmission delay and energy consumption during task offloading can be significantly reduced. Therefore, designing an efficient content caching strategy is crucial for system energy consumption optimization.

[0081] The present invention provides a DRL-based intelligent caching strategy. Different from the traditional static caching strategy, this solution dynamically adjusts the cache content according to the task popularity, historical request frequency and network topology, so as to adapt to the changes in network status. Based on the Proximal Policy Optimization (PPO) algorithm, we optimize the caching decision through Markov Decision Process (MDP) modeling.

[0082] In the simulation, the impact of cache strategy and no cache strategy on system performance is discussed:

[0083] 1) Cache strategy optimization: A collaborative cache strategy based on content popularity and historical request frequency is adopted to cache popular content to edge cloud and UAV nodes first, thus improving the cache hit rate of task requests.

[0084] 2) No-cache strategy: All tasks are offloaded to UAV or edge cloud for processing through wireless transmission, and no cache strategy is used, which increases the transmission energy consumption of the system and the delay in task completion.

[0085] Experimental results show that the solution based on intelligent caching strategy can significantly improve system performance. Considering task caching:

[0086] 1) Reduced system energy consumption: The caching strategy reduces the frequency of task offloading between sensors and UAVs, reducing transmission energy consumption.

[0087] 2) Reduced latency: The improvement in cache hit rate reduces the transmission time of tasks in the network.

[0088] 3) Improved resource utilization: By dynamically adjusting the cache content distribution, the resources of UAV and edge cloud nodes are utilized more efficiently.

[0089] Simulations further verify the advantages of this caching strategy when network resources are limited or channel conditions are poor, which is manifested in significant optimization of energy consumption.

[0090] 3. Analyze the impact of UAV intelligent hovering on system performance

[0091] It is recognized that in wireless sensor networks, the intelligent hovering position of UAV is crucial to optimizing system performance. The traditional fixed hovering position is difficult to adapt to the dynamically changing task requirements and channel environment in the network, which may lead to a decrease in signal transmission performance and an increase in energy consumption. Therefore, the present invention proposes a UAV intelligent hovering position optimization strategy based on the SCA algorithm, which dynamically adjusts the hovering position of the UAV with the goal of minimizing system energy consumption.

[0092] In the simulation, the impact of whether the UAV is intelligently hovering on system performance is explored:

[0093] 1) Fixed hovering position: The UAV hovers at a preset fixed position, without considering the dynamic changes in sensor distribution and mission requirements in the network. Although this method is simple to implement, it may lead to uneven signal coverage and reduced transmission efficiency.

[0094] 2) Intelligent hovering position: The UAV dynamically adjusts its hovering position to adapt to sensor distribution and mission traffic, thereby optimizing the signal transmission path and reducing transmission energy consumption.

[0095] The experimental results show that the UAV's intelligent hovering strategy can significantly optimize system performance, which is specifically reflected in the following aspects:

[0096] 1) Reduced transmission energy consumption: Intelligent hovering enables the UAV to approach mission-intensive areas, reducing the transmission distance between the sensor and the UAV, thereby reducing transmission energy consumption.

[0097] 2) Channel quality improvement: Dynamically adjusting the hovering position can effectively avoid obstacles and improve channel gain, thereby enhancing the reliability of task transmission.

[0098] 3) Improved task processing efficiency: Intelligent hovering position optimization enables the UAV to use its computing and cache resources more efficiently, reduce task transmission delays, and reduce system energy consumption.

[0099] In addition, we further verified the effectiveness of the strategy under poor channel conditions or high task load. Simulation results show that compared with fixed hovering positions, the intelligent hovering strategy has significant advantages under resource-constrained conditions, which not only reduces system energy consumption but also improves the network's service quality.

[0100] 4. Analyze the impact of partial uninstallation on system performance

[0101] It is recognized that in wireless sensor networks, task offloading strategy plays a vital role in system performance. The two extreme strategies of full offloading and full local processing often cannot achieve a good balance between energy consumption and task processing efficiency. Therefore, the present invention proposes a partial offloading optimization strategy, which achieves overall optimization of system performance by dynamically adjusting the distribution ratio of tasks among sensors, UAVs and edge clouds.

[0102] In the simulation, the influence of partial unloading strategy on system performance is discussed:

[0103] 1) All offloading strategy: All tasks are completed locally at the sensor node or all tasks are offloaded to the UAV or edge cloud for processing. This will increase processing delays due to the limited computing power of a single node, and will also significantly increase system energy consumption due to high-frequency data transmission.

[0104] 2) Partial offloading strategy: According to task requirements and network status, the offloading ratio of tasks between local, UAV and edge cloud is dynamically determined to achieve a balance between energy consumption and processing efficiency.

[0105] Experimental results show that the partial offloading strategy has significant performance advantages in different network environments:

[0106] 1) Reduced system energy consumption: By rationally allocating computing tasks, unnecessary data transmission is reduced, and the energy consumption of sensors and UAVs is reduced.

[0107] 2) Improved processing efficiency: The partial offloading strategy selects the best computing node according to the task complexity, which significantly shortens the task completion time and reduces system energy consumption.

[0108] 3) Improved resource utilization: Dynamic task allocation enables more efficient utilization of computing resources of sensors, UAVs, and edge clouds.

[0109] Further simulation studies show that the partial offloading strategy is particularly effective in scenarios with dense sensors and high task loads. Compared with the full offloading strategy, this strategy not only significantly reduces the total energy consumption of the system, but also improves the reliability of task processing and service quality.

[0110] 5. Analyze the impact of STAR-RIS resource allocation on system performance

[0111] It is recognized that in wireless sensor networks, the resource allocation of STAR-RIS plays an important role in optimizing system performance. STAR-RIS can significantly enhance signal transmission performance and coverage by dynamically adjusting its transmission and reflection coefficients. However, if the resource allocation of STAR-RIS is not optimized, its potential in a dynamic network environment may be difficult to give full play. Therefore, the present invention proposes a STAR-RIS resource allocation optimization strategy based on the SCA algorithm to achieve the optimization of system performance.

[0112] In the simulation, the influence of STAR-RIS resource allocation optimization on system performance is discussed:

[0113] 1) Optimized STAR-RIS resource allocation strategy: The transmission and reflection coefficients of STAR-RIS are dynamically adjusted through the SCA algorithm to adapt it to the location of sensor nodes, task offloading requirements and channel conditions, thereby improving transmission performance.

[0114] 2) STAR-RIS technology is not used: UAV acts as a relay node to forward tasks to the edge cloud, or the sensor directly transmits the task to the edge cloud for processing, which cannot reduce the forwarding energy consumption of UAV.

[0115] Experimental results show that STAR-RIS resource allocation optimization can significantly improve system performance, as shown below:

[0116] 1) Improved signal coverage: Dynamically adjusting the transmission and reflection coefficients of STAR-RIS can effectively enhance the full-space coverage capability and improve the signal quality of sensor nodes.

[0117] 2) Reduced system energy consumption: Optimized resource allocation reduces the transmission energy consumption between sensors and UAVs, effectively extending the system's operating time.

[0118] 3) Improved task processing efficiency: The optimization of STAR-RIS has significantly improved the channel conditions for task offloading and reduced the energy consumption of task transmission.

[0119] In addition, we further verified the effect of STAR-RIS resource allocation optimization in scenarios with poor channel conditions or high task load. The simulation results show that compared with the non-STAR-RIS technology, the optimized STAR-RIS resource allocation scheme shows significant performance advantages under resource-constrained conditions, especially in scenarios with large task offloading requirements, which greatly reduces system energy consumption and improves service quality.

[0120] The performance of the method involved in the present invention is analyzed and compared in combination with the simulation experiment results.

[0121] The reference positions of the ground edge cloud, UAV and STAR-RIS are set at (0,20,0) meters, Mihe In addition, four ground sensors are fixed at (-40, 40, 0) m, (-40, -40, 0) m, (40, 40, 0) m, and (40, -40, 0) m, respectively, representing a typical distributed sensor network configuration. The detailed simulation parameters are summarized in Table 1.

[0122] Table 1 Simulation parameter settings

[0123]

[0124] The embodiments of the present invention improve the endurance of unmanned aerial vehicles (UAVs) by introducing wireless power transmission technology; at the same time, the STAR-RIS technology is introduced to further reduce the UAV relay energy consumption, assist sensor nodes in forwarding tasks, and reasonably allocate UAV-based WSN resources. In addition, by installing STAR-RIS on the UAV, the system can dynamically adjust the relative position between STAR-RIS and the sensor to further improve the efficiency of task transmission and processing. The practicality and effectiveness of the embodiments of the present invention have been verified in wireless sensor networks, successfully solving the problem of minimizing WSN energy consumption, and can be widely used in smart cities, environmental monitoring, UAV network communications and other fields.

[0125] It is to be understood that the present invention is described by some embodiments, and it is known to those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the scope of protection of the present invention.

Claims

1. A method for optimizing computation offloading and content caching based on STAR-RIS assisted in airborne wireless sensor networks, characterized in that: include: Constructing a wireless sensor system model, wherein the system model is based on the distribution of sensor devices, UAVs with wireless rechargeable batteries and edge clouds in a sensor network, and the allocation of cache, computing, and communication resources, including: constructing a wireless sensor network model, a communication model, a computing model, and a cache model; Designing a UAV energy harvesting mode, the edge cloud has a radio frequency energy transmitter for broadcasting wireless energy to the UAV, and the UAV has a rechargeable battery, an energy harvesting circuit component, and a computing server, the rechargeable battery is used to store the harvested energy to power its operation; Design a pathfinding algorithm for intelligent task requests. After the UAV has collected energy, when the UAV caches the task, the corresponding node directly returns the result of the task to the sensor; when the UAV does not cache the computing task, the task is calculated in parallel on the sensor, UAV and edge cloud, and finally the result of the computing task is returned to the sensor device; Based on the wireless sensor system model, UAV energy collection mode and the path-finding algorithm of the task request, the total energy consumption equation of the wireless sensor system is constructed. By jointly optimizing the energy collection time variable, cache variable, computational offloading variable, UAV hovering position variable and STAR-RIS passive beamforming variable, the total energy consumption of the system is minimized, and it is determined that the UAV maintains continuous charging and endurance after each mission.

2. The optimization method for computing offloading and content caching based on STAR-RIS assisted in airborne wireless sensor networks as claimed in claim 1, characterized in that: Constructing a wireless sensor network model, the network model includes sensor devices, UAVs with wireless rechargeable batteries and edge cloud distribution, and STAR-RIS, the sensors are connected to the UAVs and edge cloud via wireless links, and signals are transmitted between the sensors and the edge cloud via STAR-RIS as a relay node; Build a communication model, set the communication channel between UAV, STAR-RIS and edge cloud as LoS link, and build a channel gain model based on the characteristics of LoS link to describe the energy consumption of signal transmission during communication; Constructing a computational model, the computational model being used to describe at least uplink communication energy consumption and computational energy consumption of each network node when the task request is not cached by the wireless sensor system; Build a cache model and design a task cache strategy to store completed partial tasks and data in the UAV or edge cloud. When the sensor sends the same computing task request again, if the UAV or edge cloud has cached the task and data, the calculation results will be returned to the sensor directly from the cache, thereby reducing task processing delays and energy consumption.

3. The optimization method for computing offloading and content caching based on STAR-RIS assisted in airborne wireless sensor networks as claimed in claim 2, characterized in that: The total energy consumption equation of the wireless sensor system satisfies the following formula: The total energy consumption of the STAR-RIS-assisted wireless sensor system can be minimized as follows: Among them, T h is the energy harvesting time variable, Y is the cache variable, α is the computation offloading variable, and r u is the UAV hovering position variable, and Θ is the STAR-RIS passive beamforming variable.

4. The optimization method for computing offloading and content caching based on STAR-RIS assisted in airborne wireless sensor networks as claimed in claim 3, characterized in that: The optimization of the cache variables includes: dynamically optimizing the cache variables according to the DRL algorithm, and dynamically adjusting the cache content according to content popularity, historical access frequency and network topology.

5. The optimization method for computing offloading and content caching based on STAR-RIS assisted in airborne wireless sensor networks as claimed in claim 4, characterized in that: The path-finding algorithm based on the wireless sensor system model, UAV energy collection mode and task request constructs the total energy consumption equation of the wireless sensor system, and determines the minimization of the total energy consumption of the system by jointly optimizing the energy collection time variable, cache variable, calculation offloading variable, UAV hovering position variable and STAR-RIS passive beamforming variable, and determines that the UAV maintains continuous charging and endurance after each mission, including: optimizing the relative position of the UAV and the sensor, and the transmission and reflection coefficient matrix of STAR-RIS through the SCA algorithm.

6. The optimization method for computing offloading and content caching based on STAR-RIS assisted in airborne wireless sensor networks as claimed in claim 5, characterized in that: The computational offloading variable dynamically adjusts the offloading ratio of computing tasks among sensor devices, UAVs with wireless rechargeable batteries, and edge clouds through the SCA algorithm, achieving a balance between energy consumption and processing efficiency, and keeping the UAV continuously charged and on-the-go.

Images