Wireless charging and computation offloading method for power failure image detection task

Abstract

The present application relates to a kind of wireless charging and computing unloading method for power failure image detection task, comprising the following steps: step S1: unmanned helicopter-nest power failure image detection system is constructed, and unmanned helicopter flight model, unmanned helicopter hover model, power failure image detection task model and local execution model, power failure image detection task is unloaded to nest remote execution model and nest is charged to unmanned helicopter model;Step S2: define the optimization problem of power failure image detection task completion quantity maximization, and solve the optimization problem using particle swarm algorithm.This method can ensure that the power of unmanned helicopter inspection is sufficient while improving the number of power failure image detection, thereby improving the efficiency of power failure inspection operation.

Description

Technical Field

[0001] This invention relates to the field of power transmission line inspection technology, specifically to a wireless charging and computational offloading method for power fault image detection tasks. Background Technology

[0002] The power system plays a vital role in modern economic and social development, providing essential energy security for sustainable economic and social development. Power shortages or instability not only affect national economic and social development but also cause environmental problems such as energy waste. Therefore, timely understanding of the power system's status and ensuring its safe and reliable operation are crucial. The power system comprises multiple complex components including power generation, transmission, distribution, and consumption. To ensure the safe and stable operation of the power system, regular inspections of various electrical components on transmission lines are essential for timely fault detection and repair.

[0003] Initially, power transmission line inspections primarily involved personnel observing the towers from below using telescopes or climbing them with specialized safety equipment. However, this method was inefficient and infrequent, unable to reliably inspect all electrical components in real time. Furthermore, the high altitude of power transmission lines posed a risk of falls or electric shock to personnel. To improve efficiency, helicopters were introduced, allowing for clearer aerial observation of the lines. While this method offered a more comprehensive inspection range, helicopters required significant manpower and resources, and pilot error could lead to serious injuries or fatalities. Additionally, this method still required manual assessment of electrical components, leaving room for subjective errors in component evaluation. With the rise of drones and unmanned helicopters and the development of deep learning-based object detection algorithms, using drones or unmanned helicopters for power transmission line inspection has become the preferred solution. Drone or unmanned helicopter power line inspections require equipping them with cameras to capture images of the power transmission lines along the route, followed by the use of object detection algorithms for fault identification. This method is safer and more efficient than other inspection methods. However, most drones and unmanned helicopters are powered by electricity. Even if they can offload the image detection task to nearby roadside units or cloud servers to reduce their own energy consumption for image detection, they still need a lot of energy to fly. Therefore, it is necessary to manually release and retrieve drones and unmanned helicopters and charge them, which still requires a certain amount of manpower. Summary of the Invention

[0004] The purpose of this invention is to provide a wireless charging and computational offloading method for power fault image detection tasks. This method can increase the number of power fault images detected while ensuring sufficient power for unmanned helicopter inspections, thereby improving the efficiency of power fault inspection operations.

[0005] To achieve the above objectives, the technical solution adopted by this invention is: a wireless charging and computational offloading method for power fault image detection tasks, comprising the following steps:

[0006] Step S1: Construct an unmanned helicopter-nest power fault image detection system, and define the unmanned helicopter flight model, unmanned helicopter hovering model, power fault image detection task model and local execution model, power fault image detection task local offloading to the nest for remote execution model, and nest charging model for unmanned helicopter.

[0007] Step S2: Define an optimization problem that maximizes the number of completed power fault image detection tasks, and solve the optimization problem using the particle swarm optimization algorithm.

[0008] Furthermore, step S1 specifically includes the following steps:

[0009] Step S1.1: Construct an image detection system for power faults in unmanned helicopters and their nests;

[0010] Step S1.2: Define the unmanned helicopter flight model and the unmanned helicopter hovering model;

[0011] Step S1.3: Define the power fault image detection task model and the local execution model;

[0012] Step S1.4: Define the model for the power fault image detection task to be offloaded to the machine nest for remote execution;

[0013] Step S1.5: Define the model for charging the unmanned helicopter from the avionics nest.

[0014] Furthermore, the specific method for constructing an image detection system for power faults in unmanned helicopters and their nests is as follows:

[0015] The unmanned helicopter-nest power fault image detection system includes one unmanned helicopter nest and N unmanned helicopters U = {U1, U2, ..., U...} N The unmanned helicopter nest is equipped with a high-performance server, which can analyze power failure images unloaded from different unmanned helicopters at the same frequency. ser The system performs inspections; the drone nest is also equipped with a wireless power transmission device that can wirelessly charge drones within its range; each drone is equipped with a camera to capture images of power failures and a processor to recognize the images; the drone U...i The processor supports a maximum frequency of f i,max Assume that the task of detecting a power fault image frame is an indivisible task, that is, it cannot be partially offloaded, and can only be executed entirely locally on the unmanned helicopter or entirely offloaded to the nest for remote execution.

[0016] Communication between the UAV nest and the UAV includes the UAV unloading power failure image data to the nest and the UAV nest wirelessly charging the UAV. However, wireless charging and power failure image detection data unloading are not performed simultaneously. Time-division multiplexing technology is used, that is, part of the time the UAV hovers near the nest is used for task data unloading and part for wireless power transmission. The wireless charging and power failure image detection tasks can be executed locally at the same time, that is, the local execution time of the power failure image detection task is regarded as the time the UAV hovers near the nest.

[0017] Furthermore, the specific methods for defining the unmanned helicopter flight model and the unmanned helicopter hovering model are as follows:

[0018] For uniform linear flight with a speed of V i U-shaped unmanned helicopter i The formula for calculating its propulsion power consumption is as follows:

[0019]

[0020] Among them, P i 0 P1 and P2 represent the blade profile power and inductive power in the hovering state, respectively; U i,tip This represents the tip velocity of the rotor blades; v i 0 d represents the average velocity induced by the rotor during hovering. i and s i ρ represents the fuselage drag ratio and rotor solidity of the unmanned helicopter, respectively. i and A i These represent air density and rotor disk area, respectively; when the unmanned helicopter hovers in the air, V... i Substituting 0 into formula (1), we obtain its hovering power as:

[0021] P i H =P i 0 +P i (2)

[0022] Let r represent the radius of the unmanned helicopter inspection. Then, the distance the unmanned helicopter needs to fly for one inspection is 2r; when the unmanned helicopter's inspection flight speed is V... iAt that time, the flight time for the inspection is:

[0023]

[0024] Therefore, the propulsion energy required for unmanned helicopter inspection flights is:

[0025] E i F =P i F ×t i F (4)

[0026] The propulsion energy required for hovering of an unmanned helicopter is:

[0027] E i H =P i H ×t i H (5)

[0028] Among them, t i H Let t be the hovering time of the unmanned helicopter. i F +t i H =T, where T is the total time for one round of inspection;

[0029] U-shaped unmanned helicopter i The sampling frame rate for capturing images of power failures via camera during cruise is f. i It captured a total of f images of power faults during the cruise. i ×t i F frame.

[0030] Furthermore, the specific methods for defining the power fault image detection task model and the local execution model are as follows:

[0031] U-shaped unmanned helicopter i The resolution of the captured power fault detection image is r i Then the unmanned helicopter U i Each captured frame has r i ×r i 100 pixels; σ represents the amount of data contained in each pixel, i.e., the number of pixels in each frame of an image. Bit data; φ is needed to process each unit of data. i The unmanned helicopter's processor uses dynamic voltage and frequency regulation technology to adjust the CPU frequency for processing power fault images, and its maximum CPU frequency can be f... i,maxThe time required for each frame of image to be executed locally on the unmanned helicopter is calculated using the following formula:

[0032]

[0033] Among them, f i This indicates the CPU frequency during detection; the energy required to detect each frame of power fault image locally is:

[0034]

[0035] Among them, κ i U-type unmanned helicopter i The energy efficiency ratio of the equipped processor;

[0036] Furthermore, the specific method for defining the model of offloading the power fault image detection task to remote execution at the nest is as follows:

[0037] U-shaped unmanned helicopter i The data transmission rate for unloading data into the unmanned helicopter nest is:

[0038]

[0039] Where B is the spectral bandwidth. U-type unmanned helicopter i The transmission power, h i U-type unmanned helicopter i Channel gain between the nest and the δ 2 Background noise; a frame of image from an unmanned helicopter U i The time required to unload to the hive is:

[0040]

[0041] A frame of image from an unmanned helicopter U i The energy required to unload to the hive is:

[0042]

[0043] After the power failure images are offloaded to the UAV hive, image detection tasks are performed on the hive's server; multiple detection tasks offloaded from multiple UAVs to the hive can be executed in parallel, and detection tasks offloaded from the same UAV are maintained in a queue; using f ser Let represent the frequency at which the unmanned helicopter nest server performs power failure image detection tasks. Then, the time required to execute the image detection task on the unmanned helicopter nest is:

[0044]

[0045] Furthermore, the specific method for defining the model of the avionics charging the unmanned helicopter is as follows:

[0046] The unmanned helicopter nest is equipped with a radio frequency energy transmitter, and the unmanned helicopter is equipped with a radio frequency energy harvester. The unmanned helicopter nest can wirelessly transfer power to the unmanned helicopter, i.e., wireless charging; during a round of inspection, the unmanned helicopter U... i The received energy is:

[0047]

[0048] Where, μ i ∈(0,1) represents the unmanned helicopter U i Energy receiving efficiency, Represents energy emission power, g i Indicates wireless channel gain. Indicates charging time.

[0049] Furthermore, in step S2, the specific method for defining the optimization problem of maximizing the number of completed power fault image detection tasks is as follows:

[0050] Use respectively and U-shaped unmanned helicopter i The number of captured power fault images detected locally and at the unmanned helicopter nest server during the current inspection process are calculated using the following formulas:

[0051]

[0052]

[0053] The available time for local detection is the time the unmanned helicopter hovers near the unmanned helicopter pod, i.e. In this system, power fault image detection is an indivisible task; that is, if the remaining time is insufficient to complete the detection of one frame of image, then the detection of that frame will not be performed. Due to the half-duplex transmission of the channel, the time used for unloading power fault image data is the difference between the total hovering time and the wireless charging time of the unmanned helicopter, i.e. Image detection tasks on the UAV nest server begin execution after all data transmission is complete; that is, the time spent executing detection tasks in the UAV nest is equal to the UAV charging time. The total number of power fault image detection tasks completed in the UAV nest is the smaller of the number of unloaded images and the number of tasks completed within the time limit. The total number of power fault image detections completed in one round of inspection is...

[0054] Therefore, the optimization problem of maximizing the number of completed power fault image detection tasks can be formalized as:

[0055]

[0056] The meanings of the above constraints are as follows:

[0057] C1: The frequency when performing power fault detection tasks locally must not exceed the highest frequency supported by the server equipped on the unmanned helicopter;

[0058] C2: The total time for the unmanned helicopter to take aerial photos and hover near the unmanned helicopter nest is the time for one round of inspection.

[0059] C3: The sum of the energy consumption of the unmanned helicopter during a round of inspection, the energy consumption of hovering near the nest, the energy consumption required to perform local power fault detection tasks, and the energy consumption of transmitting power fault images to the nest shall not exceed the energy obtained by the unmanned helicopter near the nest during a round of inspection, so as to ensure that the unmanned helicopter has sufficient energy for the next round of inspection.

[0060] C4: The total number of local and remote power fault image detections in a single inspection cycle shall not exceed the number of images acquired.

[0061] Furthermore, in step S2, the specific method for using the particle swarm optimization algorithm to solve the optimization problem of maximizing the number of completed power fault image detection tasks is as follows:

[0062] The flight speed V of each unmanned helicopter is determined using the particle swarm optimization algorithm. i The frequency f of performing power fault detection tasks locally on unmanned helicopters i loc And, and the time each unmanned helicopter spends charging at the nest. To solve the aforementioned optimization problem;

[0063] The position vectors of particles in a particle swarm are defined as 3N-dimensional vectors; the position vector of a particle is represented as... Where V i ι ∈(0,V max ) represents the unmanned helicopter U corresponding to the ι-th particle. i Flight speed, V max f is the maximum flight speed that an unmanned helicopter can support; i loc,ι ∈(0,f i,max ) represents the unmanned helicopter U corresponding to the ι-th particle. i The frequency at which local power fault detection tasks are performed; U represents the unmanned helicopter corresponding to the ι-th particle. i The duration of charging at the hive; correspondingly, the particle's velocity vector is expressed as...

[0064] K initial particles (p) are randomly generated. 1 ,p 2 ,…,p K ), and set it as the individual optimal particle (p 1* ,p 2* ,…,p K* For these K initial particles, the number of power fault image detection tasks they can complete is calculated using formulas (13) and (14), i.e., the particle fitness. The particle with the highest fitness is set as the globally optimal particle. The velocity vector of each particle is updated iteratively as follows:

[0065]

[0066]

[0067]

[0068] Where ω1, ω2, ω3 are inertia weights, and c 1,1 ,c 2,1 ,c 3,1 For individual cognitive weights, c 1,2 ,c 2,2 ,c 3,2 For social cognitive weight, r 1,1 ,r 1,2 ,r 2,1 ,r 2,2 ,r 3,1 ,r 3,2 They are all uniformly distributed between [0,1]. After obtaining the new particle velocity vector, the particle positions are updated. The particle position vector update method is as follows:

[0069] p ι =p ι +v ι (19)

[0070] After obtaining the new particle position vector, the fitness of each particle is calculated, and the optimal particles for all individuals and the global optimal particle are updated. After completing Y rounds of updates, the global optimal particle is the optimal solution to the problem of optimizing the number of completed tasks in the power fault image detection task.

[0071] Compared with the prior art, the present invention has the following beneficial effects:

[0072] 1) This invention constructs an unmanned helicopter-nest system. The nest can automatically control the takeoff and cruise route settings of the unmanned helicopter and wirelessly charge it. Power system maintenance personnel only need to view the results at the command center, determine the location of the fault, and notify the maintenance team for timely handling. This system offers high inspection efficiency, excellent fault detection, and ensures the personal safety of inspection personnel, saving labor costs and improving the automation level of power fault inspection operations.

[0073] 2) This invention uses a particle swarm optimization algorithm to maximize the number of power fault images detected locally and remotely, considering the limited inspection time and energy carried by the unmanned helicopter. It also ensures that the unmanned helicopter has enough energy to complete the next round of power inspection at the end of the current round. This algorithm is simple and fast, and can quickly make decisions when the inspection route changes. Furthermore, this invention considers half-duplex transmission of the channel, making it more practical and applicable. Attached Figure Description

[0074] Figure 1 This is a flowchart illustrating the method implementation of an embodiment of the present invention;

[0075] Figure 2 This is a schematic diagram of the time allocation for unmanned helicopter inspections in an embodiment of the present invention. Detailed Implementation

[0076] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0077] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0078] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0079] like Figure 1 As shown, this embodiment provides a wireless charging and computation offloading method for power fault image detection tasks, including the following steps:

[0080] Step S1: Construct an unmanned helicopter-nest power fault image detection system, and define the unmanned helicopter flight model, unmanned helicopter hovering model, power fault image detection task model and local execution model, power fault image detection task local offloading to the nest for remote execution model, and nest charging model for unmanned helicopter.

[0081] Step S1.1: Construct an image detection system for power failure in unmanned helicopters and their nests.

[0082] Construct an image detection system for power faults in unmanned helicopters and their nests. This system includes one unmanned helicopter nest and N unmanned helicopters U = {U1, U2, ..., U...} N The unmanned helicopter nest is equipped with a high-performance server, which can analyze power failure images unloaded from different unmanned helicopters at the same frequency. ser Testing is being conducted. The drone helicopter nest is also equipped with a wireless power transmission device that can wirelessly charge drones within its range. Each drone is equipped with a camera to capture images of power failures and a processor capable of image recognition. Drone U i The processor supports a maximum frequency of f i,max Since both the unmanned helicopter and the nest are equipped with processors, the power failure image detection task can be executed locally on the unmanned helicopter or offloaded to the nest for remote execution. This invention assumes that the detection task of a power failure image frame is an indivisible task, that is, it cannot be partially offloaded, and can only be executed entirely locally on the unmanned helicopter or entirely offloaded to the nest for remote execution.

[0083] Communication between the unmanned helicopter (UHV) hive and the UHV includes the UHV unloading power fault image data from the hive and the hive wirelessly charging the UHV. However, due to the half-duplex transmission of the channel, wireless charging and power fault image detection data unloading cannot occur simultaneously. This invention employs time-division multiplexing technology, using part of the time the UHV hovers near the hive for task data unloading and part for wireless power transmission. Wireless charging and local execution of the power fault image detection task can occur simultaneously; that is, the local execution time of the power fault image detection task can be considered as the time the UHV hovers near the hive. The time allocation in a round of inspection is as follows: Figure 2 As shown.

[0084] Step S1.2: Define the unmanned helicopter flight model and the unmanned helicopter hovering model.

[0085] Unmanned helicopters consume energy to move forward and maintain hovering in the air; this energy is called propulsion energy consumption. Propulsion energy consumption is related to the unmanned helicopter's speed and acceleration. In the inspection process of this invention, the unmanned helicopter travels at a constant speed; acceleration only occurs during the transition from hovering to constant-speed forward flight, which takes a very short time. Therefore, this invention ignores the additional propulsion energy consumption caused by acceleration. For uniform straight-line flight with a speed of V... i U-shaped unmanned helicopter i The formula for calculating its propulsion power consumption is as follows:

[0086]

[0087] Among them, P i 0 P1 and P2 represent the blade profile power and inductive power in the hovering state, respectively; U i,tip This represents the tip velocity of the rotor blades; v i 0 d represents the average velocity induced by the rotor during hovering. i and s i ρ represents the fuselage drag ratio and rotor solidity of the unmanned helicopter, respectively. i and A i These represent air density and rotor disk area, respectively; when the unmanned helicopter hovers in the air, V... i Substituting 0 into formula (1), we obtain its hovering power as:

[0088] P i H =P i 0 +P i (2)

[0089] Let r represent the radius of the unmanned helicopter inspection. Then, the distance the unmanned helicopter needs to fly for one inspection is 2r; when the unmanned helicopter's inspection flight speed is V... i At that time, the flight time for the inspection is:

[0090]

[0091] Therefore, the propulsion energy required for unmanned helicopter inspection flights is:

[0092] E i F =P i F ×t i F (4)

[0093] The propulsion energy required for hovering of an unmanned helicopter is:

[0094] E i H =P i H ×t i H (5)

[0095] Among them, t i H Let t be the hovering time of the unmanned helicopter. i F +t i H =T, where T is the total time for one round of inspection.

[0096] U-shaped unmanned helicopter i The sampling frame rate for capturing images of power failures via camera during cruise is f. i It captured a total of f images of power faults during the cruise. i ×t i F frame.

[0097] Step S1.3: Define the power fault image detection task model and the local execution model.

[0098] U-shaped unmanned helicopter i The resolution of the captured power fault detection image is r i Then the unmanned helicopter U i Each captured frame has r i ×r i In this invention, σ represents the amount of data contained in each pixel, that is, the number of pixels in each frame of an image. Bit data. Processing each unit of data requires φ i The processor of the unmanned helicopter can use dynamic voltage frequency regulation technology to adjust the CPU frequency for processing power fault images, and its maximum CPU frequency can be f... i,max The time required for each frame of image to be processed locally on the unmanned helicopter can be calculated using the following formula:

[0099]

[0100] Among them, f i This indicates the CPU frequency during detection; the energy required to detect each frame of power fault image locally is:

[0101]

[0102] Among them, κ i U-type unmanned helicopter iThe energy efficiency ratio of the equipped processor.

[0103] Step S1.4: Define the model for the power fault image detection task to be offloaded to the machine nest for remote execution.

[0104] Power fault images can be offloaded to the unmanned helicopter's nest processor for detection. According to the Shannon-Hartley theorem, the unmanned helicopter U... i The data transmission rate for unloading data into the unmanned helicopter nest is:

[0105]

[0106] Where B is the spectral bandwidth. U-type unmanned helicopter i The transmission power, h i U-type unmanned helicopter i Channel gain between the nest and the δ 2 Background noise. A frame of image from an unmanned helicopter... i The time required to unload to the hive is:

[0107]

[0108] A frame of image from an unmanned helicopter U i The energy required to unload to the hive is:

[0109]

[0110] After the power failure images are offloaded to the unmanned helicopter nest, image detection tasks are performed on the nest's server. Due to the high performance of the servers deployed in the unmanned helicopter nest, this invention proposes that detection tasks from multiple unmanned helicopters offloaded to the nest can be executed in parallel, with detection tasks from the same unmanned helicopter maintained in a single queue. Using f ser Let represent the frequency at which the unmanned helicopter nest server performs power failure image detection tasks. Then, the time required to execute the image detection task on the unmanned helicopter nest is:

[0111]

[0112] Step S1.5: Define the model for charging the unmanned helicopter from the avionics nest.

[0113] This invention proposes that an unmanned helicopter nest is equipped with a radio frequency (RF) power transmitter, and the unmanned helicopter is equipped with an RF power harvester. The unmanned helicopter nest can wirelessly transmit power to the unmanned helicopter, i.e., wireless charging. During a round of inspections, the unmanned helicopter U... i The received energy is:

[0114]

[0115] Where, μ i ∈(0,1) represents the unmanned helicopter U i Energy receiving efficiency, Represents energy emission power, g i Indicates wireless channel gain. Indicates charging time.

[0116] Step S2: Define an optimization problem that maximizes the number of completed power fault image detection tasks, and solve the optimization problem using the particle swarm optimization algorithm.

[0117] Step S2.1: Define the optimization problem of maximizing the number of completed power fault image detection tasks.

[0118] Use respectively and U-shaped unmanned helicopter i The number of captured power fault images detected locally and at the unmanned helicopter nest server during the current inspection process are calculated using the following formulas:

[0119]

[0120]

[0121] The available time for local detection is the time the unmanned helicopter hovers near the unmanned helicopter pod, i.e. In this invention, power fault image detection is an indivisible task; that is, if the remaining time is insufficient to complete the detection of one frame of image, then the detection of that frame of image is not performed. Due to the half-duplex transmission of the channel, the time used for unloading power fault image data in this invention is the difference between the total hovering time and the wireless charging time of the unmanned helicopter, i.e. The image detection task on the UAV nest server begins execution after all data transmission is completed; that is, the time spent executing the detection task in the UAV nest is the UAV charging time. The total number of power fault image detection tasks completed in the UAV nest is the smaller value between the number of unloaded images and the number of tasks completed within the limited time. Therefore, the total number of power fault image detections completed in one round of inspection in this invention is...

[0122] Therefore, the optimization problem of maximizing the number of completed power fault image detection tasks can be formalized as:

[0123]

[0124] The meanings of the above constraints are as follows:

[0125] C1: The frequency when performing power fault detection tasks locally must not exceed the highest frequency supported by the server equipped on the unmanned helicopter;

[0126] C2: The total time for the unmanned helicopter to take aerial photos and hover near the unmanned helicopter nest is the time for one round of inspection.

[0127] C3: The sum of the energy consumption of the unmanned helicopter during a round of inspection, the energy consumption of hovering near the nest, the energy consumption required to perform local power fault detection tasks, and the energy consumption of transmitting power fault images to the nest shall not exceed the energy obtained by the unmanned helicopter near the nest during a round of inspection, so as to ensure that the unmanned helicopter has sufficient energy for the next round of inspection.

[0128] C4: The total number of local and remote power fault image detections in a single inspection cycle shall not exceed the number of images acquired.

[0129] Step S2.2: The particle swarm optimization algorithm is used to solve the optimization problem of maximizing the number of completed power fault image detection tasks.

[0130] The flight speed V of each unmanned helicopter is determined using the particle swarm optimization algorithm. i The frequency f of performing power fault detection tasks locally on unmanned helicopters i loc And, and the time each unmanned helicopter spends charging at the nest. To solve the aforementioned optimization problem;

[0131] In this invention, the position vector and position vector of a particle in a particle swarm are defined as 3N-dimensional vectors; the position vector of a particle is represented as... Where V i ι ∈(0,V max ) represents the unmanned helicopter U corresponding to the ι-th particle. i Flight speed, V max f is the maximum flight speed that an unmanned helicopter can support; i loc,ι ∈(0,f i,max ) represents the unmanned helicopter U corresponding to the ι-th particle. i The frequency at which local power fault detection tasks are performed; U represents the unmanned helicopter corresponding to the ι-th particle. i The duration of charging at the hive; correspondingly, the particle's velocity vector is expressed as...

[0132]

[0133] First, randomly generate K initial particles (p 1 ,p 2 ,…,pK ), and set it as the individual optimal particle (p 1* ,p 2* ,…,p K* For these K initial particles, the number of power fault image detection tasks they can complete is calculated using formulas (13) and (14), i.e., the particle fitness. The particle with the highest fitness is set as the globally optimal particle. The velocity vector of each particle is updated iteratively as follows:

[0134]

[0135]

[0136]

[0137] Where ω1, ω2, ω3 are inertia weights, and c 1,1 ,c 2,1 ,c 3,1 For individual cognitive weights, c 1,2 ,c 2,2 ,c 3,2 For social cognitive weight, r 1,1 ,r 1,2 ,r 2,1 ,r 2,2 ,r 3,1 ,r 3,2 They are all uniformly distributed between [0,1]. After obtaining the new particle velocity vector, the particle positions are updated. The particle position vector update method is as follows:

[0138] p ι =p ι +v ι (19)

[0139] After obtaining the new particle position vector, the fitness of each particle is calculated, and the optimal particles for all individuals and the global optimal particle are updated. After completing Y rounds of updates, the global optimal particle is the optimal solution to the problem of optimizing the number of completed tasks in the power fault image detection task.

[0140] This invention provides a wireless charging and computational offloading method for power fault image detection tasks. In this method, an unmanned helicopter hive can control the unmanned helicopter to take off and pre-set inspection routes. After the inspection mission, the unmanned helicopter can return to its original hive or a nearby hive, offloading a portion of the power images to the hive for fault detection and wireless charging to ensure sufficient power for the next inspection. Power system maintenance personnel only need to view the results at the command center, determine the fault location, and inform the maintenance team for timely handling. The unmanned helicopter-hive system for power system inspection is not only highly efficient and effective but also ensures the personal safety of inspection personnel, saves labor costs, and improves the automation level of power fault inspection operations. In this invention, the unmanned helicopter can capture images during power inspection and perform local detection on a portion of the images. The unmanned helicopter hive can wirelessly charge unmanned helicopters within its range and remotely detect the power inspection images offloaded to it. However, due to the half-duplex transmission of the channel, wireless charging and image data offloading cannot occur simultaneously. This invention uses a particle swarm optimization algorithm to determine the flight speed of the unmanned helicopter during inspections, the amount of image data the unmanned helicopter unloads onto the hive, and the charging time of the unmanned helicopter. This ensures that the unmanned helicopter has enough energy for the next inspection and can complete as many power fault image detections as possible.

[0141] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application 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.

[0142] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. 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... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0143] 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.

[0144] 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.

[0145] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A wireless charging and computational offloading method for power fault image detection tasks, characterized in that, Includes the following steps: Step S1: Construct an unmanned helicopter-nest power fault image detection system, and define the unmanned helicopter flight model, unmanned helicopter hovering model, power fault image detection task model and local execution model, power fault image detection task offloading to the nest for remote execution model, and nest charging unmanned helicopter model. Step S2: Define an optimization problem that maximizes the number of completed power fault image detection tasks, and solve the optimization problem using the particle swarm optimization algorithm; In step S2, the specific method for defining the optimization problem of maximizing the number of completed power fault image detection tasks is as follows: Use respectively and Indicates unmanned helicopter The number of captured power fault images detected locally and at the unmanned helicopter nest server during the current inspection process are calculated using the following formulas: （13） （14） The available time for local detection is the time the unmanned helicopter hovers near the unmanned helicopter pod, i.e. In this system, power fault image detection is an indivisible task; that is, if the remaining time is insufficient to complete the detection of one frame of image, then the detection of that frame will not be performed. Due to the half-duplex transmission of the channel, the time used for unloading power fault image data is the difference between the total hovering time and the wireless charging time of the unmanned helicopter, i.e. The image detection task on the UAV nest server begins execution after all data transmission is complete; that is, the time spent executing the detection task in the UAV nest is the UAV charging time. The total number of power fault image detection tasks completed in the UAV nest is the smaller value between the number of unloaded images and the number of tasks completed within the time limit. The total number of power fault image detections completed in one round of inspection is... ; Therefore, the optimization problem of maximizing the number of completed power fault image detection tasks can be formalized as follows: （15） The meanings of the above constraints are as follows: The frequency during local power fault detection tasks must not exceed the maximum frequency supported by the server equipped on the unmanned helicopter. The total time spent by an unmanned helicopter taking aerial photos and hovering near its nest constitutes one round of inspection time. The sum of the energy consumption of the unmanned helicopter during a round of inspection, the energy consumption of hovering near the nest, the energy consumption required to perform local power fault detection tasks, and the energy consumption of transmitting power fault images to the nest must not exceed the energy obtained by the unmanned helicopter near the nest during a round of inspection, so as to ensure that the unmanned helicopter has sufficient energy for the next round of inspection. The total number of local and remote power fault images detected in a single inspection cycle shall not exceed the number of images acquired.

2. The wireless charging and computational offloading method for power fault image detection tasks according to claim 1, characterized in that, Step S1 specifically includes the following steps: Step S1.1: Construct an image detection system for power faults in unmanned helicopters and their nests; Step S1.2: Define the unmanned helicopter flight model and the unmanned helicopter hovering model; Step S1.3: Define the power fault image detection task model and the local execution model; Step S1.4: Define the model for offloading the power fault image detection task to remote execution in the nest; Step S1.5: Define the model for charging the unmanned helicopter from the avionics nest.

3. The wireless charging and computational offloading method for power fault image detection tasks according to claim 2, characterized in that, The specific method for constructing an image detection system for power faults in unmanned helicopters and their nests is as follows: The unmanned helicopter-nest power fault image detection system includes one unmanned helicopter nest. One unmanned helicopter ; The drone helicopter nest is equipped with a high-performance server that displays power failure images unloaded from different drone helicopters at the same frequency. The system performs inspections; the drone helicopter nest is also equipped with a wireless power transmission device to wirelessly charge drones within its range; each drone helicopter is equipped with a camera to capture images of power failures and a processor to recognize the images; drone helicopters The processor supports a maximum frequency of ; Assuming that the detection task of a power fault image frame is an indivisible task, that is, it cannot be partially offloaded, it can only be executed entirely locally on the unmanned helicopter or entirely offloaded to the nest for remote execution. Communication between the UAV nest and the UAV includes the UAV unloading power failure image data to the nest and the UAV nest wirelessly charging the UAV. However, wireless charging and power failure image detection data unloading are not performed simultaneously. Time-division multiplexing technology is used, that is, part of the time the UAV hovers near the nest is used for task data unloading and part for wireless power transmission. Meanwhile, wireless charging and power failure image detection tasks are executed locally simultaneously, that is, the local execution time of the power failure image detection task is regarded as the time the UAV hovers near the nest.

4. The wireless charging and computational offloading method for power fault image detection tasks according to claim 2, characterized in that, The specific methods for defining the unmanned helicopter flight model and the unmanned helicopter hovering model are as follows: For uniform linear flight with a flight speed of unmanned helicopters The formula for calculating its propulsion power consumption is as follows: （1） in, and These represent the blade profile power and inductive power in the hovering state, respectively; This indicates the tip velocity of the rotor blades; This represents the average velocity induced by the rotor during hovering. and These represent the fuselage drag ratio and rotor solidity of the unmanned helicopter, respectively. and These represent air density and rotor disk area, respectively; when the unmanned helicopter hovers in the air, Substituting into formula (1), we obtain its hovering power as: （2） use Let represent the radius of the unmanned helicopter's inspection range. Then, the distance the unmanned helicopter needs to fly for one inspection is... When the unmanned helicopter's inspection flight speed is At that time, the flight time for the inspection is: （3） Therefore, the propulsion energy required for unmanned helicopter inspection flights is: （4） The propulsion energy required for hovering of an unmanned helicopter is: （5） in, The hovering time of the unmanned helicopter, and ,in This refers to the total time for one round of inspections; Unmanned helicopter The sampling frame rate for capturing power failure images via camera during cruise is [value missing]. It captured a total of power failure images during the cruise. frame.

5. The wireless charging and computational offloading method for power fault image detection tasks according to claim 2, characterized in that, The specific methods for defining the power fault image detection task model and the local execution model are as follows: Unmanned helicopter The resolution of the captured power fault detection images is Then unmanned helicopter Each captured frame has 1 pixel; using This indicates the amount of data contained in each pixel, that is, the amount of data contained in each frame of an image. Bit data; processing each unit of data requires The unmanned helicopter's processor uses dynamic voltage and frequency regulation technology to adjust the CPU frequency for processing power fault images, with its maximum CPU frequency being [number] CPU cycles; The time required for each frame of image to be executed locally on the unmanned helicopter is calculated using the following formula: （6） in, This indicates the CPU frequency during detection; the energy required to detect each frame of power fault image locally is: （7） in, For unmanned helicopters The energy efficiency ratio of the equipped processor.

6. The wireless charging and computational offloading method for power fault image detection tasks according to claim 5, characterized in that, The specific method for defining a model that offloads the power fault image detection task to remote execution at the data center is as follows: Unmanned helicopter The data transmission rate for unloading data into the unmanned helicopter nest is: （8） in, For spectrum bandwidth, For unmanned helicopters Transmission power, For unmanned helicopters Channel gain between the nest and the host. Background noise; transferring a frame of image from an unmanned helicopter The time required to unload to the hive is: （9） A frame of image from an unmanned helicopter The energy required to unload to the hive is: （10） After power failure images are offloaded to the drone nest, image detection tasks are performed on the nest's server; multiple drone nest unloaded detection tasks are executed in parallel, with detection tasks from the same drone nest maintained in a queue; using Let represent the frequency at which the unmanned helicopter nest server performs power failure image detection tasks. Then, the time required to execute the image detection task on the unmanned helicopter nest is: （11）。 7. The wireless charging and computational offloading method for power fault image detection tasks according to claim 2, characterized in that, The specific method for defining the model of charging unmanned helicopters from a nacelle is as follows: The drone pod is equipped with a radio frequency (RF) power transmitter, and the drone is equipped with an RF power harvester. The drone pod wirelessly transmits power to the drone, effectively wirelessly charging it. During a round of inspections, the drone... The received energy is: （12） in, Indicates unmanned helicopter Energy receiving efficiency, Indicates energy emission power. Indicates wireless channel gain. Indicates charging time.

8. The wireless charging and computational offloading method for power fault image detection tasks according to claim 1, characterized in that, In step S2, the specific method for solving the optimization problem of maximizing the number of completed tasks in the power fault image detection task using the particle swarm optimization algorithm is as follows: The flight speed of each unmanned helicopter is determined using a particle swarm optimization algorithm. The frequency of performing power fault detection tasks locally on unmanned helicopters And, and the time each unmanned helicopter spends charging at the nest. To solve the aforementioned optimization problem; Define the position vector and position vector of a particle in a particle swarm as follows: 1D vector; the position vector of a particle is represented as ,in Indicates the first Each particle corresponds to an unmanned helicopter Flight speed, This is the maximum flight speed that an unmanned helicopter can support; Indicates the first Each particle corresponds to an unmanned helicopter The frequency at which local power fault detection tasks are performed; Indicates the first Each particle corresponds to an unmanned helicopter The duration of charging at the hive; correspondingly, the particle's velocity vector is expressed as... ; Randomly generated The initial particles ( ), and set it as the individual optimal particle ( ); Regarding this For each initial particle, the number of power fault image detection tasks it can complete is calculated using formulas (13) and (14), i.e., the particle fitness. The particle with the highest fitness is set as the globally optimal particle. Iterative updates are performed on each particle; the particle velocity vector is updated as follows: （16） （17） （18） in, For inertial weights, For individual cognitive weight, As a measure of social perception weight, Evenly distributed in Between these steps, after obtaining the new particle velocity vector, the particle positions are updated. The particle position vector update method is as follows: （19） After obtaining the new particle position vectors, the fitness of each particle is calculated, and the fitness of all individual optimal particles and the global optimal particle are updated; after completion... After each update, the globally optimal particle obtained is the optimal solution to the problem of optimizing the number of completed tasks in the power fault image detection task.

Images