Distribution network unmanned aerial vehicle route adaptive generation system based on AI vision assistance

Abstract

The invention relates to a distribution network unmanned aerial vehicle route adaptive generation system based on AI vision assistance. The system comprises a server, a main unmanned aerial vehicle and an auxiliary unmanned aerial vehicle. The server is composed of an arithmetic unit and a control unit, and the arithmetic unit is used for executing an adaptive algorithm, receiving and analyzing state data of the unmanned aerial vehicle to generate a control strategy and send a control instruction; the control unit is responsible for transmitting the instructions to the unmanned aerial vehicle. The main unmanned aerial vehicle executes an electric power inspection task according to a preset route, collects image data of the electric power equipment and transmits the image data to the server; the auxiliary unmanned aerial vehicle flies around the main unmanned aerial vehicle, collects attitude and motion data of the main unmanned aerial vehicle, and transmits the data to the server. The main unmanned aerial vehicle and the auxiliary unmanned aerial vehicle generate a main record Sa and an auxiliary record Sb respectively, the server integrates the main record Sa and the auxiliary record Sb into a total record S, the calculation unit adopts an actor-commentator algorithm for analysis and optimization, and a self-adaptive strategy is generated to optimize the route of the unmanned aerial vehicle.

Description

Technical Field

[0001] The present invention belongs to the field of aircraft technology, and in particular, relates to an adaptive generation system for network distribution UAV routes based on AI vision assistance. Background Art

[0002] With the continuous expansion of the scale and complexity of power grids, the traditional manual inspection method has significant deficiencies in efficiency and safety. The application of drone technology provides an efficient and safe solution for power inspection. However, the existing drone power inspection system mainly relies on preset routes for flight, lacks adaptive capabilities, and cannot adjust the flight route according to real-time conditions, resulting in unsatisfactory inspection results. In addition, drones may encounter complex environmental conditions during flight, such as wind changes, obstacles, etc. These factors will affect the flight stability and inspection accuracy of drones. In order to improve the efficiency and reliability of drone power inspection, there is an urgent need for an intelligent system that can adjust the flight route in real time and monitor and optimize the flight status. By introducing AI visual assistance and adaptive control technology, accurate flight and efficient inspection of drones in complex environments can be achieved to ensure the safe operation of power equipment.

[0003] After consulting relevant public technologies, the technical solution with publication number US20210224512A1 proposes a drone inspection system using a dual attention network, which performs image recognition of floating objects on the coastline to determine whether the floating objects are hazardous waste or other threatening objects; the technical solution with publication number CN106774221A proposes a patrol system that cooperates with drones and unmanned vehicles, which realizes three-dimensional recognition of various objects on the ground through the cooperation of the two, and the unmanned vehicle can support the drone's endurance working time; the technical solution with announcement number KR102440819B1 proposes a security system that uses drones to identify people in the area, and through regular patrols of drones, verifies whether the people in the designated area are security personnel or other personnel.

[0004] The above technical solutions all propose drone application technical solutions that use image recognition to identify objects or people, but there is still little mention of how to continuously optimize the strategy of drones in identification and inspection.

[0005] The foregoing discussion of the background art is intended only to facilitate an understanding of the present invention. This discussion does not acknowledge or admit that any of the material referred to is part of the common general knowledge. Summary of the invention

[0006] The purpose of the present invention is to provide a distribution network drone route adaptive generation system based on AI visual assistance; the generation system includes a server, a main drone and a secondary drone; the server is composed of a computing unit and a control unit, the computing unit is used to execute an adaptive algorithm, receive and analyze the state data of the drone, to generate a control strategy and send a control instruction; the control unit is responsible for transmitting these instructions to the drone; the main drone performs power inspection tasks according to the preset route, collects image data of power equipment and transmits it to the server; the secondary drone flies around the main drone, collects its posture and motion data, and transmits the data to the server. The main drone and the secondary drone generate a main record Sa and a secondary record Sb respectively, the server integrates them into a total record S, and the computing unit uses the actor-critic algorithm to analyze and optimize, and generates an adaptive strategy to optimize the drone route.

[0007] The present invention adopts the following technical solution: an adaptive generation system for network distribution drone routes based on AI vision assistance, the generation system comprising: Server, at least one primary drone and at least one secondary drone; The server comprises: A computing unit configured to receive and analyze the status data of the drone, generate, save, and optimize a control strategy for controlling the flight of the drone, and generate a control instruction for controlling the flight of one or more drones according to the control strategy; and A control unit, configured to communicate with one or more drones and send the control instructions to the drones; The master drone is configured to perform a power inspection task on a preset route, including flying and photographing the power equipment on the route, and transmitting the photographed image data to the server; The secondary drone is configured to fly around the primary drone, collect posture and motion data of the primary drone, and transmit the captured image data to the server; The data collected by the master drone is stored as a master record Sa, which includes the state data of the master drone during flight and the image data of the power equipment taken by the master drone; the data collected by the slave drone is stored as a slave record Sb, which includes the state data and action data of the master drone taken by the slave drone during flight; Furthermore, the operation unit performs the following steps: Integrate the main record Sa and the sub-record Sb to form a total record S; Using an actor-critic algorithm, based on the total record S, the current control strategy is evaluated; Continuously optimize the current control strategy based on the evaluation results, and finally generate an adaptive strategy optimized for the current route; Preferably, the computing unit runs a strategy network and an evaluation network for continuously executing the generation and optimization of adaptive strategies based on the actor-critic algorithm; wherein, The strategy network is configured to output the action of the drone at the next time node based on the current control strategy and the execution effect provided by the evaluation network, taking the current state and / or action of the drone as input; The evaluation network is configured to evaluate the execution effect of the drone in executing the current control strategy and feed the execution effect back to the strategy network; Preferably, the control strategy includes multiple strategies, and the control strategy describes the specific actions that the drone should perform based on the current state and / or action of the drone; Preferably, the computing unit further comprises using an invariant set principle to verify the stability of the evaluation network; Preferably, the evaluation network evaluates the execution effect of the drone in executing the current control strategy, and the evaluation content includes: Is the drone's attitude stable? The distance between the drone and the target power equipment to be inspected and whether its own positioning during inspection is accurate; Image quality of images collected by the drone; Preferably, the image quality of the image captured by the drone includes one or more of the following evaluation factors: Whether the target part of the target power equipment is captured in the picture; The proportion of the target part of the target power equipment in the picture; Whether the target part of the target power equipment is clear in the picture; Preferably, the aircraft control of the secondary UAV is achieved by one of the following methods: manually operated by at least one operator via the control unit, being automatically executed by the control unit; automatically executed by at least one onboard control unit deployed on the corresponding drone; Preferably, the main drone and the slave drone are equipped with at least one image acquisition device, and the image acquisition device is composed of one or more of the following sensors or image devices: camera, video camera, thermal imaging camera, infrared camera, night vision sensor, depth camera, ranging sensor, laser ranging sensor, radio ranging sensor.

[0008] The beneficial effects achieved by the present invention are: 1. The generation system of this technical solution realizes the adaptive generation and optimization of UAV routes by adopting the actor-critic algorithm and AI visual assistance; the server can continuously evaluate and adjust the flight control strategy based on the real-time data collected by the main UAV and the auxiliary UAV, so that the UAV can dynamically adapt to the complex inspection environment, improving the flight flexibility and the accuracy of task completion; 2. The generation system of this technical solution is equipped with a main drone that flies autonomously according to a predetermined route, and a cooperative working mechanism of a secondary drone that can fly automatically or manually; in the process of generating an adaptive strategy, the main drone focuses on executing the preset power inspection task, while the secondary drone is responsible for monitoring the posture and movement of the main drone and providing data support from a third-person perspective; this multi-angle and multi-perspective data collection method not only improves the drone's ability to perceive its own state, but also effectively avoids the risk of collision in the process of generating an adaptive strategy, ensuring the safety and reliability of the inspection task; after the adaptive strategy is generated, the adaptive flight of the current route can be significantly optimized, and the inspection work can be completed by one drone; 3. In the generation system of this technical solution, the server integrates the main record and the secondary record to form a total record, and uses the actor-critic algorithm to perform real-time analysis and optimize the control strategy; the evaluation network evaluates and feedbacks the execution effect of the drone, ensuring the continuous improvement and optimization of the control strategy; this mechanism not only improves the intelligence level of drone inspections, but also ensures the efficient processing and utilization of inspection data, and improves the response speed and work efficiency of the overall system; 4. The software and hardware parts of the generation system of this technical solution adopt a modular design. The various working modules and components of the hardware part of the system, as well as the instructions, parameters, and algorithms of the software part can be easily replaced and / or upgraded later, thereby reducing the construction cost and maintenance cost of this system. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The present invention can be further understood from the following description in conjunction with the accompanying drawings. The components in the figures are not necessarily drawn to scale, but the emphasis is placed on illustrating the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0010] Description of the accompanying drawings: 10-server; 20-master drone; 30-slave drone; 40-network; 50-target power equipment; 301-image input layer; 302-sensor input layer; 311-convolution layer A; 312-pooling layer A; 321-convolution layer B; 322-pooling layer B; 330-flattening layer; 341-fully connected layer A; 350-merging layer; 351-fully connected layer B; 352-fully connected layer C; 360-fully connected layer D; 500-computer system; 502-bus; 504-processor; 506-main memory; 508-read-only memory; 510-storage device; 512-display; 514-input device; 516-cursor control device; 518-network device; Figure 1 It is a schematic diagram of the framework of the growth system according to an embodiment of the present invention; Figure 2 is a schematic diagram of a data acquisition part of the growth system described in an embodiment of the present invention; Figure 3 is a schematic diagram of an evaluation network described in an embodiment of the present invention; Figure 4 Schematic diagram of performance comparison of the evaluation network after verification using the invariant set principle in an embodiment of the present invention; Figure 5 It is a schematic diagram of the framework of the computer system used by the computing unit in the embodiment of the present invention. DETAILED DESCRIPTION

[0011] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with its embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. For those skilled in the art, after reviewing the following detailed description, other systems, methods and / or features of the present embodiment will become apparent. It is intended that all such additional systems, methods, features and advantages are included in this specification. Included within the scope of the present invention and protected by the appended claims. Additional features of the disclosed embodiments are described in the following detailed description, and these features will be apparent from the following detailed description.

[0012] The same or similar reference numerals in the drawings of the embodiments of the present invention correspond to the same or similar components; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right" and the like indicating the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, it is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or component referred to must have a specific orientation. The invention is constructed and operated in a specific orientation, so the terms describing the positional relationship in the drawings are only used for exemplary description and cannot be understood as a limitation of this patent. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0013] Embodiment 1: For example, a network distribution drone route adaptive generation system based on AI vision assistance is proposed, and the generation system includes: Server, at least one primary drone and at least one secondary drone; The server comprises: A computing unit configured to receive and analyze the status data of the drone, generate, save, and optimize a control strategy for controlling the flight of the drone, and generate a control instruction for controlling the flight of one or more drones according to the control strategy; and A control unit, configured to communicate with one or more drones and send the control instructions to the drones; The master drone is configured to perform a power inspection task on a preset route, including flying and photographing the power equipment on the route, and transmitting the photographed image data to the server; The secondary drone is configured to fly around the primary drone, collect posture and motion data of the primary drone, and transmit the captured image data to the server; The data collected by the master drone is stored as a master record Sa, which includes the state data of the master drone during flight and the image data of the power equipment taken by the master drone; the data collected by the slave drone is stored as a slave record Sb, which includes the state data and action data of the master drone taken by the slave drone during flight; Furthermore, the operation unit performs the following steps: Integrate the main record Sa and the sub-record Sb to form a total record S; Using an actor-critic algorithm, based on the total record S, the current control strategy is evaluated; Continuously optimize the current control strategy based on the evaluation results, and finally generate an adaptive strategy optimized for the current route; Preferably, the computing unit runs a strategy network and an evaluation network for continuously executing the generation and optimization of adaptive strategies based on the actor-critic algorithm; wherein, The strategy network is configured to output the action of the drone at the next time node based on the current control strategy and the execution effect provided by the evaluation network, taking the current state and / or action of the drone as input; The evaluation network is configured to evaluate the execution effect of the drone in executing the current control strategy and feed the execution effect back to the strategy network; Preferably, the control strategy includes multiple strategies, and the control strategy describes the specific actions that the drone should perform based on the current state and / or action of the drone; Preferably, the computing unit further comprises using the invariant set principle to verify the stability of the adaptive strategy; Preferably, the evaluation network evaluates the execution effect of the drone in executing the current control strategy, and the evaluation content includes: Is the drone's attitude stable? The distance between the drone and the target power equipment to be inspected and whether its own positioning during inspection is accurate; Image quality of images collected by the drone; Preferably, the image quality of the image captured by the drone includes one or more of the following evaluation factors: Whether the target part of the target power equipment is captured in the picture; The proportion of the target part of the target power equipment in the picture; Whether the target part of the target power equipment is clear in the picture; Preferably, the aircraft control of the secondary UAV is achieved by one of the following methods: manually operated by at least one operator via the control unit, being automatically executed by the control unit; automatically executed by at least one onboard control unit deployed on the corresponding drone; Preferably, the master drone and the slave drone are equipped with at least one image acquisition device, and the image acquisition device is composed of one or more of the following sensors or image devices: camera, video camera, thermal imaging camera, infrared camera, night vision sensor, depth camera, ranging sensor, laser ranging sensor, radio ranging sensor; As attached Figure 1As shown, it is a schematic diagram of an exemplary embodiment of the generation system; wherein the server 10 can be various types of computer devices, including but not limited to: traditional physical servers, virtual servers, cloud computing servers, edge computing devices or other electronic devices with data processing capabilities; the server 10 is connected to the drone through the network 40, receives data sent by the drone, and sends control instructions to the drone; the type of the network 40 may be WiFi, radio, cellular network (such as 4G, 5G), satellite communication, Bluetooth, ZigBee or other wireless communication technologies suitable for data transmission; the communication bandwidth and communication distance of the network 40 can ensure real-time data exchange between the server and the drone, and support remote control, status monitoring and image data transmission of the drone; Exemplarily, both the main drone 20 and the secondary drone 30 are connected to the server 10 for communication; the main drone 20 and the secondary drone 30 may be drones of the same type, or based on the requirements of the power inspection task, the main drone 20 may be a drone of a different type and specification from the secondary drone 30 because it is responsible for a relatively complex task; Furthermore, the operator can operate the drone by using the control unit on the server; preferably, the main drone 20 and the slave drone 30 can be controlled by a preset automated flight program pre-stored in the server 10, thereby realizing the function of the generation system; optionally, the operator can also intervene through the control unit to manually operate the main drone 20 and / or the slave drone 30 to actively take over the drone; In a more specific embodiment, the state s of the drone represents various state information of the current environment and the drone itself, and the state s includes but is not limited to: The position and speed of the UAV, where the position may include the degree of deviation of the UAV from the flight route, and the hovering position of the UAV when it is necessary to hover at a fixed position to collect images of the target power equipment; The attitude of the drone, such as pitch, roll, and yaw angles; Current environmental perception information, such as sensor data, camera images, etc.; preferably, the environmental perception information includes the clarity of the collected images; when the drone collects images of the target power equipment, the collection point marked by the preset route may not be the optimal position due to errors, for example, the distance may be too far or too close, so that the detection image of the appropriate image range cannot be obtained, or the observation angle selected is incorrect, resulting in a large number of irrelevant objects in the viewfinder, which makes it difficult to focus; Action a refers to the specific control instructions that the drone can execute in a specific state, such as adjusting the speed of the drone, changing the flight direction of the drone, and adjusting the altitude of the drone. For the acquisition device on the drone, it includes controlling the angle, focal length, aperture, shutter, sensitivity, anti-shake, etc. of the camera; In the case of using only one drone to perform the task of photographing power equipment, there are many limitations and potential risks in its perception of the state s. Possible situations include: the drone's field of view is limited, the camera and sensor are mainly facing the front and the target object, and it is difficult to cover all directions around it, resulting in blind spots, and the side and back cannot be monitored, posing a risk of collision. In addition, the drone cannot monitor the front and rear environments at the same time during complex flight operations, and ignores the sudden situation of obstacles in the rear. At the same time, if the drone has problems such as sensor failure or communication interruption, the entire system mission will be interrupted because there are no other drones to monitor the status and environment of the main drone. In addition, drones cannot fully monitor the surrounding environment in real time in the complex environment of power lines, which increases the difficulty of dynamic obstacle avoidance. The drone also needs to perform flight control, equipment photography and environmental monitoring at the same time, which increases the complexity of the task and the system load, and insufficient resource allocation, affecting the efficiency of task execution. Therefore, in a preferred exemplary embodiment, the above problems can be significantly solved by introducing a secondary drone; first, the secondary drone can fly around the main drone, shoot and monitor the main drone and its surroundings from different angles, provide an all-round field of view, eliminate the blind spots of the main drone, and provide real-time environmental feedback to ensure that the main drone flies in the correct predetermined route and posture, and collect the status data of the main drone during flight from a third-person perspective; secondly, as an auxiliary system, the secondary drone can take over part of the monitoring and navigation tasks when the main drone fails or the sensor fails, provide redundant perception information, and ensure mission continuity; the secondary drone can also enhance dynamic obstacle avoidance capabilities, monitor dynamic changes around the main drone in real time, and promptly discover and feedback Potential obstacles, improve reaction speed, and ensure the safe flight of the main drone in complex environments; finally, the secondary drone shares part of the environmental monitoring and data processing tasks, reducing the burden on the main drone, allowing it to focus more on shooting and inspection tasks, and improving overall task efficiency through division of labor, cooperation and optimized resource utilization; in addition, the secondary drone can shoot the environmental position of the main drone from multiple angles, and by shooting the posture and movement of the main drone, confirm that the posture detected by the main drone through its own sensors is consistent, thereby confirming that the main drone has no mistakes in its own status monitoring; this double verification mechanism can ensure the accuracy of the flight posture and position data of the main drone, and improve the stability and safety of the entire system; Based on the above configuration, in a preferred exemplary embodiment, the master record Sa formed by the master drone includes: Self-status data: including the flight parameters of the main UAV, such as current position, speed, altitude, attitude angle and other information; these data are obtained through the main UAV's own sensors (such as GPS, inertial measurement unit) and used for real-time flight control and navigation; Collected image data of power equipment: Image data of power equipment taken by the host drone during flight, which is used for visual inspection and status assessment of power equipment; these images can contain detailed visual information of equipment such as wires and transformers; In a preferred exemplary embodiment, the secondary record Sb formed by the secondary drone includes: Main UAV status data: The status data of the main UAV captured by the secondary UAV, including the position, speed, attitude, etc. of the main UAV, to back up and verify its own status data in the main record Sa; Data on the interaction between the main drone and the environment: including the distance between the main drone and objects in the environment, whether there are dangerous objects in the flight path of the main drone, etc. Main UAV motion data: The main UAV motion data recorded by the secondary UAV, such as control instructions, heading changes, etc. during flight; these data are used to verify the flight dynamics and behavior of the main UAV; Furthermore, the main record Sa and the sub-record Sb are integrated into a data set of state s; both the main record Sa and the sub-record Sb are data sequences based on time series; wherein, Master record Sa={(Sa at ,Ia at )},at=1,2,3……Ta; among them, Sa at is the state vector of the main UAV at time t, Ia at is the image data taken by the master drone at time t; Ta is the total length of the data sequence Sa; Sub-record Sb={(Sb bt ,Ib bt )}, bt=1, 2, 3...Tb; among them, Sb bt Ib is the state vector of the main UAV at time bt, bt It is the image data taken by the secondary drone at time bt; The above-mentioned at and Sb bt Both state vectors include multiple components, each of which represents the value of a state of the master drone; Furthermore, the main record Sa and the secondary record Sb are integrated, including merging and deduplication of the two records, so that the total record S is finally integrated; the data of the total record S is used as the sample data of state s This integrated status data not only provides the real-time status and environmental information of the main UAV, but also enhances the entire system's guarantee of flight safety and data accuracy through the data verification mechanism of the secondary UAV; this design not only improves the reliability and stability of the system, but also optimizes the efficiency and accuracy of power equipment inspections. Furthermore, the policy function C W (s) means that in strategy C, the parameter set W is used to determine the probability of selecting action a in state s; preferably, the strategy function C is implemented by a neural network, where the parameter set W is the set of weights and biases of each node in the hidden layer of the neural network; the strategy function C W The output of (s) can be a probability distribution (in discrete action space) or a specific action (in continuous action space); by training a large number of parameters in the parameter set W, the policy network is made to perform well in the policy function C W (s) Under the strategic model constructed, learn to choose appropriate control actions in different states; Furthermore, in a preferred embodiment, the evaluation network can collect simulation data as training data by running a simulation environment, or use historical operation data of the UAV in the past power inspection as training data for neural network training; optionally, the neural network can also be trained by using randomly generated limited range data as training data; the evaluation network aims to estimate the expected value of future rewards for given states and actions, and its goal is to learn the reward structure and state transition rules of the environment, which information can be learned through interaction or historical data before the policy network is trained; Preferably, in a simulation environment, the initial control strategy can be a random strategy, a manually designed strategy, or a predefined strategy; thereafter, the initial control strategy can be tested multiple times, and each state-action pair and its corresponding reward and next state can be recorded; in this way, a large amount of data for training can be generated; Preferably, historical data experienced by the drone under the initial control strategy is used, and state-action pairs and their corresponding rewards can be extracted from these data; the historical data can come from previous operation records, experimental data or simulation data; At the same time, in the process of reinforcement learning, preferably, an experience replay buffer is used to store the experience of the agent's interaction with the environment; these experiences include state, action, reward and next state; these data can be reused to train the value assessment network; More specifically, the evaluation network is based on a state-action value function Q of a policy C C (s, a; ψ) as the output model for evaluation; The state-action value function Q() can be expressed as: ; In the above formula, represents the expected cumulative return when taking action a from state s and then following strategy C based on strategy C; ψ is the parameter group of the state-action value function Q; Where η is the discount factor, which ranges from [0 to 1] and is set by the technician; η t represents the change with time step t, η t Decreasing is used to indicate the change in the weight value of future rewards, thereby reducing the impact of future returns; this method makes the algorithm pay more attention to recent returns rather than the distant future; E is the expected value r t It is the sequence of reward values ​​expected to be obtained when executing each subsequent [state-action] pair according to strategy C; it is also determined by the technician when formulating the initial strategy; Take the data in Table 1 below as an example to explain the significance of the state-action value function Q: Table 1 Finally, we can calculate: Q=r0+0.9*r1+0.81*r2+0.729*r3=4.707; When the state s changes, such as when the takeoff is unstable or the direction is incorrect, the corresponding reward value is reduced; and the reward value r can be negative; Based on the above settings, the training process of the evaluation network is: (1) Sampling data: randomly sample a batch of samples from the training set, where each sample includes: a pair (s, a) of the drone’s state s and the action a that needs to be performed according to the established strategy C, the reward r corresponding to (s, a), and the expected state s´ after performing action a corresponding to the current state s; (2) Preferably, define the target value target of each sample. For a sample (s, a, r, s´), we have: target=r+ηMax a´ Q(s´,a´); In the above formula, r is the reward value obtained by taking action a in the current state s, which is pre-set by the technician in strategy C; η is the discount factor, which is in the interval [0, 1]; Max a´ Q(s´,a´) represents the maximum Q value of the optimal action in the future, which means the maximum Q value that can be obtained from all possible actions a´ in the next state s´. In the next state s´, the drone can choose different actions a´ in the hope of maximizing the Q value and calculate the maximum Q value. The significance of the above formula is that it combines the current immediate reward with the maximum possible reward in the future to evaluate the long-term reward of the current state-action pair. By continuously updating the Q value, the algorithm can gradually optimize the strategy C so that the action selected in each state can maximize the long-term reward. (3) In a preferred embodiment, the mean square error is used as the loss function L() to calculate the value function ; In the above formula, L(ψ) is the evaluation network model using the parameter set ψ, E represents the expected value or approximate expected value; (s, a, r, s´) ~ D is to randomly sample a four-tuple (s, a, r, s´) from the experience replay buffer D; the loss function is minimized by training the parameter set ψ of the evaluation network model; in the above formula, the expected value is calculated for the total return of all input data (sa) pairs in the sample and the square of the difference between Q and the target value target; In an exemplary embodiment, after obtaining the trained evaluation network, the actor-critic algorithm is implemented through the following steps to generate the adaptive strategy: (1) Initialize network parameters: Initialize the policy function C in the policy network W (s) and the evaluation function Q in the evaluation network C The parameter set W and ψ in (s,a;ψ); (2) Sampling: Execute strategy C in the environment W (s), collect multiple state-action-reward triplets (s, a, r); (3) Training the evaluation network: Use the collected data to train the evaluation network Q C (s,a;ψ); (4) Calculate the gradient: For each (s, a), calculate ∇ W logC W (a|s)Q(s,a;ψ); (5) Update the policy network: Use the calculated gradient to update the policy network parameters W, that is: W←W+α∇ W logC W (a|s)Q(s,a;ψ), α is the learning rate, which is set by relevant technical personnel; (6) Repeat the process of sampling, training the evaluation network, and updating the policy network until the policy network converges; Through the above steps, the training of the strategy network is completed, and the strategy network can be applied to generate an adaptive strategy during the flight of the master drone, and the master drone can be controlled to perform the inspection task of the power equipment alone according to the adaptive strategy; In an exemplary embodiment, the structure of the evaluation network is as shown in the attached Figure 3 As shown; preferably, the evaluation network includes two input layers, which are used to process images and sensor input data respectively; wherein the input layer includes an image input layer 301 for receiving input RGB image data, the image size is (H, W, 3), which means the height is H pixels, the width is W pixels, and each pixel has three color channels (red, green, blue), and this layer provides the image data input interface required by the model; the input layer also includes a sensor input layer 302, which is used to receive data from other sensors of the drone, the input size is (n), and the specific sensor data may include physical parameters such as speed and angle; this layer provides the sensor data input interface required by the model; Furthermore, the evaluation network also includes a feature extraction layer. The image processing part includes: A convolution layer 311: This layer uses 32 8x8 convolution kernels to perform convolution operations on the input image with a step size of 4; through the ReLU activation function, the low-level features in the image are extracted to generate a feature map; Pooling layer A 312: This layer uses a 2x2 pooling window and performs a maximum pooling operation with a step size of 2, thereby reducing the size of the feature map while retaining important features and reducing computational complexity; B Convolutional layer 321: This layer uses 64 4x4 convolution kernels and performs convolution operations with a stride of 2 to further extract mid-level features in the image and process the feature map through the ReLU activation function.

[0014] Pooling layer 322: This layer again uses a 2x2 pooling window and performs maximum pooling with a stride of 2 to further reduce the size of the feature map while retaining important features; Flattening layer 330: This layer flattens the multi-dimensional feature map generated by the convolution operation into a one-dimensional vector for subsequent full connection layer processing; The sensor processing part includes an A-connection layer 341, which contains 128 neurons and uses the ReLU activation function. This layer is used to process and extract high-dimensional features of sensor data and provide input for subsequent feature merging; Furthermore, the evaluation network also includes a data merging part, including a merging layer 351: used to splice the flattened output of the image processing part and the output of the sensor data processing part; by merging the features of the image and the sensor data, comprehensive information is provided for further feature extraction and decision-making; B fully connected layer 352: The second fully connected layer contains 512 neurons and uses the ReLU activation function. This layer performs further nonlinear transformation on the merged features to extract higher-level features; C fully connected layer 353: The third fully connected layer contains 256 neurons and uses the ReLU activation function. This layer further processes features and provides input for the final value assessment; Furthermore, the evaluation network also includes a fully connected layer 360 in the output part, which includes 1 neuron and uses a linear activation function; this layer is used to output the state-action value Q(s,a), evaluate the value of the current state and action combination, and provide feedback for the optimization of the strategy network.

[0015] Embodiment 2: This embodiment should be understood to include at least all the features of any of the above embodiments, and further improve upon them; Furthermore, in a preferred embodiment, the invariant set principle is used to analyze and verify the stability of the evaluation network in the technical solution; The verification steps include: (1) Choose an appropriate Lyapunov function V(s) such that V(s)>0 when x≠0 and V(0)=0 when x=0; More specifically, we set V(s)=Q(s,a;ψ)-Q(s,a;ψ) * , where Q(s,a;ψ) * represents the value of Q(s,a;ψ) under the optimal strategy; (2) Set s´=f(s,W), where the function f() represents the change law of state s under the current state s and control strategy parameters W; (3) Using the chain rule, we can calculate the time derivative V´(s) of V(s), and we have: V´(s)=(∂V(s) / ∂s)·s´=(∂V(s) / ∂s)·f(s,W); (4) Verify that for all states of the multidimensional variable s, V´(s) ≤ 0, that is, Q(s, a; ψ) can be determined to be a stable state, that is, the evaluation network has sufficient stability; After comparison, as shown in the attached Figure 4 As shown, the reward values ​​of the evaluation network after 100 training times are compared; the reward value curve using the present technical solution is curve Rew.A, and the reward value curve of the evaluation network that is verified without using the invariant set principle is curve Rew.B; it can be seen that curve Rew.A can reach stability faster and, compared with curve Rew.B, can obtain a higher reward value under the same number of training times, that is, better performance.

[0016] Embodiment 3: This embodiment should be understood to include at least all the features of any of the above embodiments, and further improve upon them; For example, as shown in the attached Figure 5As shown, the embodiment of the computer system 500 used by the computing unit in the generating system is described; the computer system 500 can be applied to the data storage, computing and result output process of each working module in the recognition and judgment system; Illustratively, computer system 500 includes a bus 502 or other communication mechanism for communicating information, one or more processors 504 coupled to bus 502 for processing information; processor 504 may be, for example, one or more general-purpose microprocessors; The computer system 500 also includes a main memory 506, such as a random access memory (RAM), a cache, and / or other dynamic storage device, coupled to the bus 502 for storing information and instructions to be executed by the processor 504; the main memory 506 may also be used to store temporary variables or other intermediate information during the execution of instructions to be executed by the processor 504; these instructions, when stored in a storage medium accessible to the processor 504, present the computer system 500 as a special-purpose machine customized to perform the operations specified in the instructions; The computer system 500 may also include a read-only memory (ROM) 508 or other static storage device coupled to the bus 502 for storing static information and instructions for the processor 504; a storage device 510 such as a disk, an optical disk, or a USB drive (flash drive) will be coupled to the bus 502 for storing information and instructions; And further, coupled to the bus 502 may also include a display 122 for displaying various information, data, media, etc., an input device 514 for allowing a user of the computer system 500 to control, manipulate, and / or interact with the computer system 500; A preferred way to interact with the management system may be through a cursor control device 516, such as a computer mouse or similar control / navigation mechanism; Furthermore, the computer system 500 may also include a network device 518 coupled to the bus 502; wherein the network device 518 may include, for example, a wired network card, a wireless network card, a switching chip, a router, a switch, and other components; In general, the terms "engine", "component", "system", "database", etc., as used herein, may refer to logic embodied in hardware or firmware, or to a collection of software instructions, possibly with entry and exit points, written in a programming language such as Java, C, or C++; software components may be compiled and linked into executable programs, installed in a dynamic link library, or may be written in an interpreted programming language (e.g., BASIC, Perl, or Python); it should be understood that software components may be callable from other components or from themselves, and / or may be called in response to detected events or interrupts; Software components configured to execute on a computing device may be provided on a computer-readable medium, such as a compact disc, digital video disc, flash drive, diskette, or any other tangible medium, or as a digital download (and may be initially stored) in a compressed or installable format that requires installation, decompression, or decryption prior to execution); such software code may be stored in part or in whole on a memory device of the executing computing device for execution by the computing device; software instructions may be embedded in firmware, such as an EPROM; it is also understood that hardware components may be composed of connected logic units (such as gates and flip-flops), and / or may be composed of programmable units (such as programmable gate arrays or processors); Computer system 500 includes the techniques described herein that may be implemented using custom hardwired logic, one or more ASICs or FPGAs, firmware, and / or program logic that, in combination with the computer system, renders computer system 500 a special-purpose computing device; According to one or more embodiments, the techniques herein may be performed by computer system 500 in response to processor 504 executing one or more sequences of one or more instructions contained in main memory 506; such instructions may be read into main memory 506 from another storage medium, such as storage device 510; execution of the sequences of instructions contained in main memory 506 causes processor 504 to perform the process steps described herein; in alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions; As used herein, the term "non-transitory media" and similar terms refer to any media that store data and / or instructions that cause a machine to operate in a specific manner; such non-transitory media may include non-volatile media and / or volatile media; non-volatile media include, for example, optical or magnetic disks, such as storage device 510; volatile media include dynamic memory, such as main memory 506; Among them, common forms of non-transitory media include, for example, floppy disks, diskettes, hard disks, solid-state drives, magnetic tapes or any other magnetic data storage medium, CD-ROMs, any other optical data storage medium, any physical medium having a pattern of holes, RAM, PROM and EPROM, FLASH-EPROM, NVRAM, any other memory chip or cartridge, and network versions thereof; Non-transient media are distinct from transmission media but may be used in conjunction with transmission media; transmission media participate in the transmission of information between non-transient media; for example, transmission media include coaxial cables, copper wires, and optical fibers, including the wires that make up bus 502; transmission media may also take the form of sound waves or light waves, such as radio waves and infrared data communications.

[0017] Although the present invention has been described above with reference to various embodiments, it should be understood that many changes and modifications may be made without departing from the scope of the present invention. That is, the methods, systems and devices discussed above are examples. Various configurations may appropriately omit, replace or add various processes or components. For example, in alternative configurations, the method may be performed in an order different from the order described, and / or various components may be added, omitted and / or combined. Moreover, the features described with respect to certain configurations may be combined in various other configurations, such as different aspects and elements of the configurations may be combined in a similar manner. In addition, the elements therein may be updated as the technology develops, i.e., many elements are examples and do not limit the scope of the present disclosure or claims.

[0018] Specific details are given in the specification to provide a thorough understanding of the exemplary configurations including implementations. However, the configurations may be practiced without these specific details. For example, well-known circuits, processes, algorithms, structures, and techniques have been shown without unnecessary details to avoid obscuring the configurations. This description provides only example configurations and does not limit the scope, applicability, or configurations of the claims. On the contrary, the foregoing description of the configurations will provide those skilled in the art with an enabling description for implementing the described techniques. Various changes may be made to the functions and arrangements of the elements without departing from the spirit or scope of the present disclosure.

[0019] In summary, it is intended that the above detailed description is considered to be illustrative rather than restrictive, and it should be understood that the above embodiments should be understood to be only used to illustrate the present invention and not to limit the scope of protection of the present invention. After reading the contents of the present invention, the technician can make various changes or modifications to the present invention, and these equivalent changes and modifications also fall within the scope defined by the claims of the present invention.

Claims

1. The adaptive generation system of network distribution drone routes based on AI vision assistance is characterized by: include: server; The server comprises: A computing unit configured to receive and analyze the status data of the drone, generate, save, and optimize a control strategy for controlling the flight of the drone, and generate a control instruction for controlling the flight of one or more drones according to the control strategy; and A control unit, configured to communicate with one or more drones and send the control instructions to the drones; A master drone is configured to perform a power inspection task on a preset route, including flying and photographing the power equipment on the route, and transmitting the photographed image data to the server; A secondary drone is configured to fly around the primary drone, collect posture and motion data of the primary drone, and transmit the captured image data to the server; The data collected by the master drone is stored as a master record Sa, which includes the state data of the master drone during flight and the image data of the power equipment taken by the master drone; the data collected by the slave drone is stored as a slave record Sb, which includes the state data and action data of the master drone taken by the slave drone during flight; Furthermore, the operation unit performs the following steps: Integrate the main record Sa and the sub-record Sb to form a total record S; Using an actor-critic algorithm, based on the total record S, the current control strategy is evaluated; The current control strategy is continuously optimized based on the evaluation results, and ultimately an adaptive strategy optimized for the current route is generated.

2. The generation system according to claim 1, characterized in that: The operation unit runs a strategy network and an evaluation network for continuously executing the generation and optimization of adaptive strategies based on the actor-critic algorithm; wherein, The strategy network is configured to output the action of the drone at the next time node based on the current control strategy and the execution effect provided by the evaluation network, taking the current state and / or action of the drone as input; The evaluation network is configured to evaluate the execution effect of the drone in executing the current control strategy and feed back the execution effect to the strategy network.

3. The generation system according to claim 2, characterized in that: The control strategy includes multiple strategies, and the control strategy describes the specific actions that the drone should perform based on the current state and / or action of the drone.

4. The generation system according to claim 3, characterized in that: The operation unit also includes using an invariant set principle to verify the stability of the evaluation network.

5. The generation system according to claim 4, characterized in that: The evaluation network evaluates the execution effect of the UAV's current control strategy, and its evaluation content includes: Is the drone's attitude stable? The distance between the drone and the target power equipment to be inspected and whether its own positioning during inspection is accurate; Image quality of images collected by drones.

6. The generation system according to claim 5, characterized in that: The image quality of the images collected by the drone includes one or more of the following evaluation factors: Whether the target part of the target power equipment is captured in the picture; The proportion of the target part of the target power equipment in the picture; Whether the target part of the target power equipment is clear in the picture.

7. The generation system according to claim 6, characterized in that: The flight control of the secondary drone is achieved by one of the following methods: manually operated by at least one operator via the control unit, being automatically executed by the control unit; The process is automatically performed by at least one onboard control unit deployed on the corresponding drone.

8. The generation system according to claim 7, characterized in that: The main drone and the slave drone are equipped with at least one image acquisition device, and the image acquisition device is composed of one or more of the following sensors or image devices: camera, video camera, thermal imaging camera, infrared camera, night vision sensor, depth camera, ranging sensor, laser ranging sensor, radio ranging sensor.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the functions of the generation system as described in any one of claims 1 to 8 are performed.

Images