A drone nest control method and system

Abstract

This application discloses a drone nest control method and system, which includes obtaining first configuration data of a flying drone and obtaining local configuration data of a drone nest that matches the flying drone; determining second configuration data of the flying drone based on a target mission request, comparing and analyzing the first configuration data and the second configuration data to obtain a configuration indicator to be replaced; screening the local configuration data based on the configuration indicator to be replaced, and controlling the drone nest to execute an internal configuration component mobilization signal based on the screening result to bring the drone nest to an initial switching state; obtaining first flight posture data of the flying drone; predicting the landing point of the flying drone based on the first flight posture data to obtain predicted landing point coordinates; and controlling the drone nest to execute a target switching state signal generated by the initial switching state and the predicted landing point coordinates. This method provides strong support for multi-tasking applications of drones.

Description

Technical Field

[0001] The present application relates to the field of UAV technology, and in particular to a method and system for controlling a UAV nest. Background Art

[0002] With the widespread application of drone technology, drones need to switch tasks and dock with the drone nest efficiently and accurately when performing diverse tasks.

[0003] Traditionally, drones only begin deployment for their next mission after returning to their home base. This model works well in simple scenarios with relatively few mission types and infrequent switching. However, in today's complex and ever-changing real-world environments with highly diverse mission requirements, especially when switching missions frequently, this can lead to mission delays and waste resources.

[0004] It can be seen that how to optimize the control method of the drone nest to improve the response speed of the drone nest has become a technical problem that needs to be solved urgently by those skilled in the art. Summary of the Invention

[0005] The present application provides a drone nest control method and system to solve the technical problem of how to optimize the drone nest control method to achieve the effect of improving the response speed of the drone nest.

[0006] In order to solve the above technical problems, an embodiment of the present application provides a drone nest control method, which is characterized by comprising:

[0007] In response to the target task request, obtaining first configuration data of a flying drone and obtaining local configuration data of a drone nest matching the flying drone;

[0008] Determining second configuration data of the flying drone according to the target task request, and performing a comparison and analysis of the first configuration data and the second configuration data based on natural language technology to obtain a configuration indicator of the flying drone to be replaced;

[0009] Filtering the local configuration data based on the configuration indicator to be replaced, and controlling the drone nest to execute an internal configuration component mobilization signal according to the filtering result, so that the drone nest reaches an initial switching state;

[0010] When it is detected that the position information of the flying drone meets a preset range, first flight posture data captured by a first sensor provided on the flying drone is acquired;

[0011] Predicting the landing point of the flying UAV based on the first flight posture data to obtain predicted landing point coordinates;

[0012] The drone nest is controlled to execute a target switching state signal generated by the initial switching state and the predicted landing point coordinates.

[0013] As one preferred solution, after obtaining the first configuration data of the flying drone, the method further includes:

[0014] Obtaining raw data of the priority level, remaining time, and completion progress of the current mission of the flying UAV, and normalizing the raw data to obtain a current mission feature vector;

[0015] Matching the current task feature vector with a pre-built decision rule table to calculate the switching matching degree of the current task;

[0016] If the switching matching degree is greater than a preset task switching threshold, it is determined that task switching is required.

[0017] As one preferred solution, determining the second configuration data of the flying UAV according to the target mission request includes:

[0018] Inputting the target task request into the pre-trained large language model for parsing to obtain a parsing result;

[0019] Matching the parsing result with a pre-built knowledge graph, and extracting flight data, sensor data, communication data, and payload data corresponding to the target mission request from the knowledge graph based on the matching result;

[0020] The second configuration data is generated according to the flight data, the sensor data, the communication data and the mounting data.

[0021] As one preferred solution, predicting the landing position of the flying UAV based on the first flight posture data to obtain the predicted landing point coordinates includes:

[0022] Acquiring second flight posture data captured by a second sensor provided on the drone nest;

[0023] Performing weighted fusion on the first flight posture data and the second flight posture data to calculate fused posture data of the flying UAV;

[0024] The fused pose data is input into a pre-trained UAV pose model for prediction to obtain the predicted landing point coordinates of the flying UAV.

[0025] As one preferred solution, the step of obtaining the second flight posture data captured by the second sensor provided at the drone nest includes:

[0026] Acquire image data of the flying drone captured by visual sensors arranged at various positions of the drone nest, and perform image enhancement and denoising on the image data to obtain a preprocessed image;

[0027] Analyzing the pre-processed image based on image processing technology to obtain a six-degree-of-freedom posture of the flying UAV;

[0028] constructing an image sequence of the flying drone based on the preprocessed images, parsing the image sequence based on an optical flow method to obtain a motion trajectory of the flying drone;

[0029] The second flight posture data is generated according to the six-degree-of-freedom posture and the motion trajectory.

[0030] As one preferred solution, performing weighted fusion on the first flight posture data and the second flight posture data to calculate the fused posture data of the flying UAV includes:

[0031] Construct a distance weight formula from the flying drone to the drone nest, expressed as:

[0032]

[0033]

[0034] in, About distance A monotonically decreasing function of is the rate of weight change, is the distance threshold, is the weight of the first flight posture data, is the weight of the second flight posture data;

[0035] The first flight posture data and the second flight posture data are fused according to the distance weight formula to obtain fused posture data of the flying UAV.

[0036] As one preferred solution, before inputting the fused pose data into a pre-trained UAV pose model for prediction, the method further includes:

[0037] Acquire historical landing data of the flying UAV, wherein the historical landing data includes historical flight posture data and historical landing position data;

[0038] Performing weighted processing on the historical flight posture data based on the distance between the flying UAV and the UAV nest to obtain a first sample data set;

[0039] Constructing an initial UAV posture model based on an LSTM network, inputting the first sample data set into the initial UAV posture model, and constructing a loss function based on the obtained output results and the historical landing position data;

[0040] The initial UAV attitude model is optimized based on the loss function to obtain a trained UAV attitude model.

[0041] As one preferred solution, the controlling the drone nest to execute a target switching state signal generated by the initial switching state and the predicted landing point coordinates includes:

[0042] Constructing an objective function for minimizing the landing deviation of the flying UAV based on the predicted landing point coordinates;

[0043] Solving the objective function based on a dynamic programming algorithm and optimizing the solution process through a pruning strategy and an approximate dynamic programming method to obtain an optimal decision for the drone nest;

[0044] A control scheme for the drone nest is generated according to the optimal decision, and the drone nest is controlled to execute a target switching state signal generated by the control scheme.

[0045] As one preferred solution, after generating the control solution of the drone nest according to the optimal decision, the method further includes:

[0046] Calculating a predicted landing time of the flying UAV to the predicted landing point coordinates based on the first flight posture data;

[0047] The predicted landing time is compared with the obtained adjustment time of the control scheme, and the speed of the flying UAV is controlled according to the comparison result so that the predicted landing time matches the adjustment time.

[0048] Another embodiment of the present application provides a drone nest control system, comprising:

[0049] an acquisition module, configured to acquire first configuration data of a flying drone in response to a target task request, and acquire local configuration data of a drone nest matching the flying drone;

[0050] an analysis module, configured to determine second configuration data of the flying drone according to the target task request, and compare and analyze the first configuration data and the second configuration data based on natural language technology to obtain a configuration indicator of the flying drone to be replaced;

[0051] A mobilization module, configured to filter the local configuration data based on the configuration indicator to be replaced, and control the drone nest to execute an internal configuration component mobilization signal according to the screening result, so that the drone nest reaches an initial switching state;

[0052] a detection module, configured to obtain first flight posture data captured by a first sensor provided on the flying drone when detecting that the position information of the flying drone meets a preset range;

[0053] A prediction module, configured to predict the landing position of the flying UAV based on the first flight posture data to obtain predicted landing point coordinates;

[0054] A control module is used to control the drone nest to execute a target switching state signal generated by the initial switching state and the predicted landing point coordinates.

[0055] Compared with the prior art, the embodiments of the present application have the following advantages:

[0056] 1) This application predicts the landing point of a flying drone and generates a target switching state signal in advance, allowing the drone nest to complete corresponding adjustments and preparations when the drone is about to land, further shortening the time interval from the drone landing to the execution of the next mission, and overall improving the drone system's response speed to tasks.

[0057] 2) This application does not require excessive human intervention throughout the entire process, from acquiring data, analyzing and comparing, controlling the mobilization of machine nest components to predicting the landing point and generating target switching status signals. The machine nest can autonomously complete a series of operations, realizing the automation of drone mission switching and machine nest docking, reducing the impact of human factors, and improving the stability and reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 This is a flow chart of a method for controlling a drone nest in one embodiment of the present application;

[0059] Figure 2 It is a structural diagram of a drone nest control system in one embodiment of the present application. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. The purpose of providing these embodiments is to make the disclosure of this application more thorough and comprehensive. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0061] In the description of this application, the terms "first," "second," "third," etc. are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first," "second," "third," etc. may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0062] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be a communication between the two elements. The terms "vertical", "horizontal", "left", "right", "up", "down" and similar expressions used herein are for illustrative purposes only, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application. The term "and / or" used herein includes any and all combinations of one or more related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0063] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this application have the same meanings as those commonly understood by those skilled in the art. The terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood by those skilled in the art in specific circumstances.

[0064] An embodiment of the present application provides a method for controlling a drone nest. For details, see Figure 1 , Figure 1 The figure shows a flow chart of a method for controlling a drone nest in one embodiment of the present application, which includes steps S1-S6:

[0065] S1: In response to a target task request, obtain first configuration data of a flying drone and obtain local configuration data of a drone nest that matches the flying drone;

[0066] In this step, the first configuration data includes information such as the drone's current flight parameters, sensor status, communication settings, and mounted equipment. This data reflects the drone's state before receiving a new mission request. The drone's associated drone nest also has its own local configuration data, including its location, charging capacity, maintenance equipment status, and storage capacity. Obtaining this data helps the system comprehensively consider the drone's mission execution environment and supporting conditions.

[0067] Preferably, in one embodiment of the present application, after obtaining the first configuration data of the flying drone, the method further includes:

[0068] Obtain the raw data of the priority level, remaining time and completion progress of the current UAV mission, and normalize the raw data to obtain the current mission feature vector;

[0069] Match the current task feature vector with the pre-built decision rule table and calculate the switching matching degree of the current task;

[0070] If the switching matching degree is greater than a preset task switching threshold, it is determined that task switching is required.

[0071] Priority indicates the importance of the task, remaining time reflects the time urgency of the task, and completion progress reflects the degree of task execution. Because different types of data (such as priority, remaining time, and completion progress) may have different dimensions and value ranges, these raw data are normalized to facilitate unified analysis and processing. This normalized data is then integrated into a single feature vector for the current task, which comprehensively reflects the status characteristics of the current task.

[0072] The current task feature vector obtained above is matched against a pre-built decision rule table. This decision rule table is pre-built based on specific business logic and experience, and it specifies the relationship between different task feature vectors and the likelihood of task switching. Through this matching process, the current task's switching match is calculated. This match quantifies the likelihood of the current task switching. For example, if the decision rule table specifies that when the task feature vector is within a certain range, the task switching match is 0.8, then when the current task feature vector matches this range, the result is a switching match of 0.8. The higher the match, the more likely the current task will require switching.

[0073] The calculated switching match is compared with the preset task switching threshold. If the switching match exceeds the preset threshold, it indicates that the current task meets the pre-set conditions for task switching, and therefore a task switch is determined to be necessary. Otherwise, the current task is maintained. For example, in a specific drone application scenario, the business logic may be to allow a new high-priority task to preempt the current task if the current task has a lower priority, a longer remaining time, and a higher completion progress. This is known as a task switch.

[0074] S2: Determine second configuration data of the flying UAV according to the target mission request, compare and analyze the first configuration data and the second configuration data based on natural language technology, and obtain a configuration indicator to be replaced of the flying UAV;

[0075] Preferably, in one embodiment of the present application, determining the second configuration data of the flying drone according to the target mission request includes:

[0076] Input the target task request into the pre-trained large language model for parsing to obtain the parsing result;

[0077] Match the parsing results with the pre-built knowledge graph, and extract the flight data, sensor data, communication data, and payload data corresponding to the target mission request from the knowledge graph based on the matching results;

[0078] Second configuration data is generated based on the flight data, sensor data, communication data, and mounting data.

[0079] The large language model has powerful natural language understanding capabilities, enabling it to perform semantic analysis and keyword extraction on the text in task requests, generating parsing results that provide a preliminary understanding of the user's requirements for the target task.

[0080] The parsing results from the large language model are matched against a pre-built knowledge graph. A knowledge graph is a structured semantic knowledge base that organizes and stores various knowledge in a graphical structure. It contains a wealth of knowledge about drone flight, sensors, communications, and payloads, as well as the relationships between them. Based on the matching results, the knowledge graph is used to extract flight data (such as altitude, speed, and route), sensor data (such as sensor type and parameter settings), communication data (such as communication frequency and bandwidth), and payload data (such as payload type and function) corresponding to the target mission request.

[0081] Based on the extracted flight data, sensor data, communication data, and payload data, a second configuration data set for the target mission is generated. This set of data will be used to guide the drone to adjust its state to meet the requirements of the new target mission.

[0082] The first and second configuration data are then converted into natural language text. Using semantic understanding techniques from natural language processing, the textualized first and second configuration data are analyzed in depth to extract key features and semantic information. For example, descriptions of flight parameters, sensor performance, communication settings, and other aspects of the data can be identified and their specific meanings understood.

[0083] The first configuration data and the second configuration data are compared using natural language processing techniques such as similarity calculation and entity matching. For example, the two data sets can be compared to see if the flight altitude values are the same or if the sensor types are the same. For similar concepts in text descriptions, semantic similarity can also be used to determine the degree of difference.

[0084] Based on the comparison and analysis, the differences between the first and second configuration data are identified. These differences represent configuration indicators that need to be adjusted or replaced. For example, if the sensor type in the first configuration data is a standard camera, but the second configuration data requires a lidar, then the sensor type is a configuration indicator to be replaced.

[0085] Identified differences are further organized and summarized to determine specific configuration indicators for replacement. These indicators can be presented as lists or structured data, clearly indicating which configurations need to be replaced or adjusted, along with the direction and target value. For example, the configuration indicators for replacement could be expressed as: "Sensor type: change from standard camera to lidar; flight altitude: adjust from 100 meters to 150 meters."

[0086] S3: Filter the local configuration data based on the configuration indicators to be replaced, and control the drone nest to execute the internal configuration component mobilization signal according to the screening results, so that the drone nest reaches the initial switching state;

[0087] Specifically, the system compares the configuration indicators to be replaced with the local configuration data one by one. For each configuration indicator to be replaced, the system searches the local configuration data for matching configuration component information. For example, if the configuration indicator requires replacing the drone's sensor from a standard camera to a lidar, the system will search the local configuration data for available lidar sensors, their quantity, and their status (e.g., whether they are operational or require maintenance).

[0088] Based on the screening results, if the local configuration data contains a configuration component that meets the requirements for the replacement, the system generates a corresponding internal configuration component mobilization signal. These signals contain clear operational instructions, such as which storage location on the drone nest to retrieve a specific sensor or mount device from and how to install it on the drone. These mobilization signals are then transmitted to the drone nest's control unit via a communication link between the nest and the control system. This communication link can be wired (e.g., Ethernet) or wireless (e.g., Wi-Fi, Bluetooth, etc.), ensuring accurate and timely transmission of the signals to the drone nest.

[0089] Upon receiving the activation signal, the drone's control unit controls the internal mechanical and electrical systems to perform the corresponding operations. For example, it controls the robotic arm to remove the required sensor from the storage rack and move it to the drone's installation location, or adjusts the parameters of the charging equipment to meet the drone's new power requirements.

[0090] S4: When it is detected that the position information of the flying drone meets a preset range, first flight posture data captured by a first sensor provided on the flying drone is obtained;

[0091] The preset range is a geographical area or location interval pre-set based on a specific application scenario. The first sensor installed on a flying drone is a sensor used to measure the drone's flight posture. Common sensors include an inertial measurement unit (IMU), a magnetometer, and a barometer. The IMU can measure the drone's acceleration and angular velocity, and through integration operations, it can obtain the drone's attitude (such as pitch, roll, and yaw) and speed information. The magnetometer can measure the direction of the Earth's magnetic field to determine the drone's heading. The barometer can measure atmospheric pressure and calculate the drone's altitude through the relationship between pressure and altitude.

[0092] When the system detects that the drone's position information falls within a preset range, it triggers data acquisition. Using the drone's internal communication bus (such as I2C or SPI), the system reads the captured first flight pose data from the first sensor. This data is typically stored in binary or text format in the sensor's registers. The system must read and parse the data according to the sensor's communication protocol, converting it into actual physical quantities (such as angle, velocity, and altitude).

[0093] S5: Predicting the landing point of the flying UAV based on the first flight posture data to obtain predicted landing point coordinates;

[0094] Preferably, in one embodiment of the present application, predicting the landing position of the flying drone based on the first flight posture data to obtain the predicted landing point coordinates includes:

[0095] Acquiring second flight posture data captured by a second sensor provided on the drone's nest;

[0096] Performing weighted fusion on the first flight pose data and the second flight pose data to calculate fused pose data of the flying UAV;

[0097] The fused pose data is input into the pre-trained UAV pose model for prediction to obtain the predicted landing point coordinates of the flying UAV.

[0098] Specifically, to improve prediction accuracy, the system utilizes not only the first flight pose data captured by the first sensor on the flying drone, but also the second flight pose data captured by the second sensor on the drone's cradle. By weightedly fusing these two sets of data and then feeding them into a pre-trained drone attitude model, the system generates more reliable predicted landing point coordinates.

[0099] Preferably, in one embodiment of the present application, obtaining second flight posture data captured by a second sensor provided at a drone nest includes:

[0100] Obtain image data of the flying UAV captured by visual sensors installed at various positions of the UAV nest, perform image enhancement and denoising on the image data to obtain a pre-processed image;

[0101] The pre-processed image is analyzed based on image processing technology to obtain the six-degree-of-freedom posture of the flying UAV;

[0102] Construct an image sequence of a flying UAV based on the preprocessed images, and analyze the image sequence based on the optical flow method to obtain the motion trajectory of the flying UAV;

[0103] Second flight posture data is generated according to the six-degree-of-freedom posture and motion trajectory.

[0104] Vision sensors installed at various locations on the drone's nest capture images of the drone in flight and acquire image data. These sensors capture drone images from different angles, providing more comprehensive information. Because actual shooting environments may present issues such as uneven lighting and noise, image enhancement and denoising are necessary to improve image quality. Image enhancement can utilize methods such as histogram equalization and contrast adjustment to highlight key information within the image. Denoising can utilize algorithms such as Gaussian filtering and median filtering to remove noise points from the image, resulting in a pre-processed image.

[0105] The pre-processed image is analyzed using image processing techniques. For example, feature extraction algorithms (such as SIFT and SURF) are used to extract the drone's characteristic points in the image. By analyzing the position and relationship of these characteristic points and combining them with the camera's calibration parameters, the six degrees of freedom (DOF) of the flying drone are calculated. These are the three translational degrees of freedom (along the X, Y, and Z axes) and the three rotational degrees of freedom (rotations around the X, Y, and Z axes).

[0106] An image sequence of the flying drone is constructed from the preprocessed images. These images are arranged in chronological order, recording the drone's status at different moments. The image sequence is then parsed using the optical flow method. This method calculates the speed and direction of an object by analyzing the movement of pixels in adjacent images. By analyzing the optical flow information of the drone in the image sequence, the drone's trajectory can be determined, including its position changes at each moment.

[0107] Combining the six-degree-of-freedom attitude and motion trajectory information obtained above, the second flight pose data is generated. This data contains key information such as the attitude and motion state of the drone relative to the aircraft nest, providing an important basis for subsequent data fusion and landing point prediction.

[0108] A distance weighting formula for the distance between a flying drone and the drone nest is constructed. This formula is a monotonically decreasing function of distance. The core idea is that when the drone is close to the nest, the data obtained by the second sensor on the nest is relatively more accurate, so the second flight pose data is given a higher weight. When the drone is far away from the nest, the data obtained by the first sensor on the flying drone is relatively more reliable, so the first flight pose data is given a higher weight.

[0109] Preferably, in one embodiment of the present application, weighted fusion is performed on the first flight posture data and the second flight posture data to calculate the fused posture data of the flying UAV, including:

[0110] Construct a distance weight formula from the flying drone to the drone nest, expressed as:

[0111]

[0112]

[0113] in, About distance A monotonically decreasing function of is the rate of weight change, is the distance threshold, is the weight of the first flight posture data, is the weight of the second flight posture data;

[0114] The first flight posture data and the second flight posture data are fused according to a distance weight formula to obtain fused posture data of the flying UAV.

[0115] Among them, the rate of weight change controls how fast the weight changes with distance; the distance threshold is a key parameter used to define the range of the impact of distance on the weight.

[0116] Based on the distance-weighted formula, the first and second flight pose data are weightedly fused. For each pose parameter (such as position, velocity, and attitude), a corresponding weight is calculated using the weighting formula. The weighted sum of the pose parameters in the first and second pose data is then applied to generate the fused pose data for the UAV. This fusion approach fully accounts for the varying reliability of sensor data at different distances, improving the accuracy and reliability of the fused data.

[0117] Preferably, in one embodiment of the present application, before inputting the fused pose data into the pre-trained UAV pose model for prediction, the method further includes:

[0118] Obtain the historical landing data of the flying UAV, which includes historical flight posture data and historical landing position data;

[0119] Performing weighted processing on historical flight posture data based on the distance between the flying UAV and the UAV nest to obtain a first sample data set;

[0120] Build an initial UAV posture model based on the LSTM network, input the first sample data set into the initial UAV posture model, and construct a loss function based on the obtained output results and historical landing position data;

[0121] The initial UAV attitude model is optimized based on the loss function to obtain a trained UAV attitude model.

[0122] Specifically, before inputting the fused pose data into the pre-trained UAV attitude model for prediction, the model needs to be trained and optimized to ensure that it can accurately predict the landing position of the UAV based on the flight pose data.

[0123] The distance between the drone and the nest affects data reliability at different flight stages. Generally, when the drone is close to the nest, sensors around the nest can more accurately capture drone information, resulting in higher data reliability. However, when the drone is farther away, it may be subject to more external interference during flight, resulting in lower data reliability. Therefore, historical flight posture data needs to be weighted based on the drone's distance from the nest.

[0124] Each historical flight pose data sample is assigned a distance-related weight. Samples closer to the aircraft nest receive a higher weight, while samples farther from the aircraft nest receive a lower weight. This approach emphasizes the influence of highly reliable data on model training, improving model accuracy. The weighted historical flight pose data forms the first sample dataset.

[0125] The Long Short-Term Memory (LSTM) network is a specialized recurrent neural network (RNN) that effectively processes sequential data and addresses the vanishing or exploding gradient issues that traditional RNNs often encounter when processing long sequences. In drone attitude prediction tasks, flight pose data is a chronological sequence. LSTM networks are well-suited for this task because they can capture temporal dependencies within the data. Based on the task requirements and data characteristics, an initial drone attitude model based on the LSTM network is constructed. The model typically consists of an input layer, an LSTM layer, and an output layer. The input layer receives historical flight pose data from the first sample dataset; the LSTM layer processes and memorizes the input sequence data using internal memory cells and gating mechanisms; and the output layer outputs the predicted landing position data.

[0126] The loss function measures the difference between the model's predictions and actual historical landing location data. By minimizing the loss function, the model continuously adjusts its parameters to improve prediction accuracy. The first sample dataset is fed into the initial drone attitude model, which outputs predicted landing location data. These predictions are then compared with the actual historical landing location data. An appropriate loss function (such as mean squared error or cross entropy) is used to calculate the difference between the two. The smaller the loss function value, the closer the model's predictions are to the actual situation.

[0127] Select an appropriate optimization algorithm (such as stochastic gradient descent or the Adam optimizer) to minimize the loss function. The optimization algorithm continuously adjusts the model parameters based on the gradient of the loss function, ensuring that the model's predictions gradually approach the actual historical landing location data. During training, the first sample dataset is fed into the model multiple times for iterative training. With each iteration, the optimization algorithm updates the model parameters based on the gradient of the loss function until the loss function converges to a small range or the preset number of training rounds is reached. After training and optimization, a trained drone attitude model is obtained. This model can be used to predict the fused pose data and obtain the predicted landing coordinates of the flying drone.

[0128] S6: Control the drone nest to execute the target switching state signal generated by the initial switching state and the predicted landing point coordinates.

[0129] After obtaining the initial switching state and predicted landing point coordinates of the flying UAV, it is necessary to generate a target switching state signal to control the UAV nest so that the nest is ready to welcome the UAV landing.

[0130] Preferably, in one embodiment of the present application, controlling the drone nest to execute a target switching state signal generated by the initial switching state and the predicted landing point coordinates includes:

[0131] Constructing an objective function to minimize the landing deviation of the flying UAV based on the predicted landing point coordinates;

[0132] The objective function is solved based on the dynamic programming algorithm, and the solution process is optimized through pruning strategy and approximate dynamic programming method to obtain the optimal decision of the drone nest;

[0133] A control scheme for the UAV nest is generated according to the optimal decision, and the UAV nest is controlled to execute the target switching state signal generated by the control scheme.

[0134] The objective function is constructed to minimize the landing deviation of a flying drone. Landing deviation can be understood as the difference between the actual landing position of the drone and the predicted landing point coordinates. By considering the predicted landing point coordinates as well as factors such as the drone's flight characteristics and the location of the aircraft nest, a mathematical function is constructed to describe this deviation. For example, the objective function can be defined as the sum of the squares of the Euclidean distances between the drone's landing point and the predicted landing point in plane coordinates, or a more complex function can be constructed by combining multi-dimensional factors such as altitude deviation. The purpose of this objective function is to provide a quantitative evaluation standard for the subsequent optimal decision-making, so that the solution process can proceed in the direction of minimizing landing deviation.

[0135] Dynamic programming is a method for solving optimization problems in multi-stage decision-making processes. In this scenario, the decision-making process for the drone nest is considered a multi-stage process, with different decision options at each stage. The goal is to find a series of decisions that ultimately minimize the objective function value. Dynamic programming algorithms can decompose complex global optimization problems into multiple subproblems, solving these subproblems sequentially to ultimately obtain the global optimal solution.

[0136] During the problem-solving process, pruning strategies are used to reduce unnecessary computation. By analyzing the problem's characteristics and constraints, decision branches that clearly do not lead to the optimal solution are preemptively eliminated, significantly reducing the search space and improving solution efficiency. For some complex problems, exact solutions can be computationally prohibitive or time-consuming. Approximate dynamic programming methods use approximate methods to further improve solution speed while maintaining a certain level of solution accuracy. For example, value function approximation and strategy approximation can be used to simplify the problem-solving process.

[0137] The above solution and optimization process yields the optimal decision for the drone nest. This optimal decision includes the specific actions the nest should take at each stage, such as adjusting the position of internal equipment, preparing charging facilities, and opening the hatch. Based on this optimal decision, a detailed control scheme is generated, which specifies the action instructions and time sequence for each actuator in the nest. Finally, the drone nest is controlled to execute the target state switching signals generated by this control scheme, ensuring that the nest switches states according to the predetermined plan and prepares for the drone's landing.

[0138] Preferably, in one embodiment of the present application, after generating a control plan for the drone nest according to the optimal decision, the method further includes:

[0139] Calculating a predicted landing time of the flying UAV to the predicted landing point coordinates based on the first flight posture data;

[0140] The predicted landing time is compared with the obtained adjustment time of the control scheme, and the speed of the flying drone is controlled according to the comparison result so that the predicted landing time matches the adjustment time.

[0141] Based on the first flight pose data, including the drone's current position, speed, acceleration, and other information, combined with the predicted landing point coordinates, the predicted landing time required for the drone to reach the predicted landing point coordinates can be calculated using kinematic formulas or more complex flight models. For example, without considering complex factors such as air resistance, a simple calculation can be performed using the formulas for uniform linear motion or uniformly accelerated linear motion. For more precise calculations, multiple factors such as the drone's flight trajectory, wind direction, and air density must be considered.

[0142] Get the control scheme's adjustment time, which represents the time required for the drone's nest to adjust from its current state to a state ready for landing. Compare the predicted landing time with the adjustment time to determine whether they match. If the predicted landing time is less than the adjustment time, the drone may arrive at the nest too early and not be ready. If the predicted landing time is greater than the adjustment time, the nest may have completed its adjustment early, waiting too long for the drone.

[0143] Based on the comparison results, the drone's speed is controlled to ensure the predicted landing time matches the adjusted time. If the predicted landing time is less than the adjusted time, the drone's flight speed can be appropriately reduced to extend its arrival time at the predicted landing point. If the predicted landing time is greater than the adjusted time, the drone's flight speed can be appropriately increased to shorten its arrival time. This ensures that the drone lands accurately after the nest completes its state transition, improving the overall system's operational efficiency and reliability.

[0144] An embodiment of the present application provides a drone nest control system. For details, see Figure 2 , Figure 2 The figure shows a schematic diagram of a drone nest control system in one embodiment of the present application, which includes:

[0145] An acquisition module 11 is configured to acquire first configuration data of a flying UAV in response to a target task request, and acquire local configuration data of a UAV nest that matches the flying UAV;

[0146] An analysis module 12 is configured to determine second configuration data of the flying UAV according to the target mission request, and compare and analyze the first configuration data and the second configuration data based on natural language technology to obtain a configuration indicator of the flying UAV to be replaced;

[0147] The mobilization module 13 is used to filter the local configuration data based on the configuration index to be replaced, and control the drone nest to execute the internal configuration component mobilization signal according to the screening result, so that the drone nest reaches the initial switching state;

[0148] The detection module 14 is configured to obtain first flight posture data captured by a first sensor provided on the flying drone when detecting that the position information of the flying drone meets a preset range;

[0149] The prediction module 15 is used to predict the landing position of the flying UAV based on the first flight posture data to obtain the predicted landing point coordinates;

[0150] The control module 16 is used to control the drone nest to execute the target switching state signal generated by the initial switching state and the predicted landing point coordinates.

[0151] Compared with the prior art, the embodiments of the present application have the following advantages:

[0152] 1) This application predicts the landing point of a flying drone and generates a target switching state signal in advance, allowing the drone nest to complete corresponding adjustments and preparations when the drone is about to land, further shortening the time interval from the drone landing to the execution of the next mission, and overall improving the drone system's response speed to tasks.

[0153] 2) This application does not require excessive human intervention throughout the entire process, from acquiring data, analyzing and comparing, controlling the mobilization of machine nest components to predicting the landing point and generating target switching status signals. The machine nest can autonomously complete a series of operations, realizing the automation of drone mission switching and machine nest docking, reducing the impact of human factors, and improving the stability and reliability of the system.

[0154] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A method for controlling a drone nest, characterized in that: include: In response to the target task request, obtaining first configuration data of a flying drone and obtaining local configuration data of a drone nest matching the flying drone; Determining second configuration data of the flying drone according to the target task request, and performing a comparison and analysis of the first configuration data and the second configuration data based on natural language technology to obtain a configuration indicator of the flying drone to be replaced; Filtering the local configuration data based on the configuration indicator to be replaced, and controlling the drone nest to execute an internal configuration component mobilization signal according to the filtering result, so that the drone nest reaches an initial switching state; When it is detected that the position information of the flying drone meets a preset range, first flight posture data captured by a first sensor provided on the flying drone is acquired; Predicting the landing point of the flying UAV based on the first flight posture data to obtain predicted landing point coordinates; The drone nest is controlled to execute a target switching state signal generated by the initial switching state and the predicted landing point coordinates.

2. The drone nest control method according to claim 1, wherein: After obtaining the first configuration data of the flying drone, the method further includes: Obtaining raw data of the priority level, remaining time, and completion progress of the current mission of the flying UAV, and normalizing the raw data to obtain a current mission feature vector; Matching the current task feature vector with a pre-built decision rule table to calculate the switching matching degree of the current task; If the switching matching degree is greater than a preset task switching threshold, it is determined that task switching is required.

3. The drone nest control method according to claim 2, wherein: Determining the second configuration data of the flying drone according to the target mission request includes: Inputting the target task request into the pre-trained large language model for parsing to obtain a parsing result; Matching the parsing result with a pre-built knowledge graph, and extracting flight data, sensor data, communication data, and payload data corresponding to the target mission request from the knowledge graph based on the matching result; The second configuration data is generated according to the flight data, the sensor data, the communication data and the mounting data.

4. The drone nest control method according to claim 1, wherein: The step of predicting the landing position of the flying UAV based on the first flight posture data to obtain predicted landing point coordinates includes: Acquiring second flight posture data captured by a second sensor provided on the drone nest; Performing weighted fusion on the first flight posture data and the second flight posture data to calculate fused posture data of the flying UAV; The fused pose data is input into a pre-trained UAV pose model for prediction to obtain the predicted landing point coordinates of the flying UAV.

5. The drone nest control method according to claim 4, wherein: The obtaining of second flight posture data captured by a second sensor provided at the drone nest includes: Acquire image data of the flying drone captured by visual sensors arranged at various positions of the drone nest, and perform image enhancement and denoising on the image data to obtain a preprocessed image; Analyzing the pre-processed image based on image processing technology to obtain a six-degree-of-freedom posture of the flying UAV; constructing an image sequence of the flying drone based on the preprocessed images, parsing the image sequence based on an optical flow method to obtain a motion trajectory of the flying drone; The second flight posture data is generated according to the six-degree-of-freedom posture and the motion trajectory.

6. The drone nest control method according to claim 4, wherein: The weighted fusion of the first flight posture data and the second flight posture data to calculate the fused posture data of the flying UAV includes: Construct a distance weight formula from the flying drone to the drone nest, expressed as: in, About distance A monotonically decreasing function of is the rate of weight change, is the distance threshold, is the weight of the first flight posture data, is the weight of the second flight posture data; The first flight posture data and the second flight posture data are fused according to the distance weight formula to obtain fused posture data of the flying UAV.

7. The drone nest control method according to claim 4, wherein: Before inputting the fused pose data into the pre-trained UAV pose model for prediction, the method further includes: Acquire historical landing data of the flying UAV, wherein the historical landing data includes historical flight posture data and historical landing position data; Performing weighted processing on the historical flight posture data based on the distance between the flying UAV and the UAV nest to obtain a first sample data set; Constructing an initial UAV posture model based on an LSTM network, inputting the first sample data set into the initial UAV posture model, and constructing a loss function based on the obtained output results and the historical landing position data; The initial UAV attitude model is optimized based on the loss function to obtain a trained UAV attitude model.

8. The drone nest control method according to claim 1, wherein: The controlling the drone nest to execute a target switching state signal generated by the initial switching state and the predicted landing point coordinates includes: Constructing an objective function for minimizing the landing deviation of the flying UAV based on the predicted landing point coordinates; Solving the objective function based on a dynamic programming algorithm and optimizing the solution process through a pruning strategy and an approximate dynamic programming method to obtain an optimal decision for the drone nest; A control scheme for the drone nest is generated according to the optimal decision, and the drone nest is controlled to execute a target switching state signal generated by the control scheme.

9. The drone nest control method according to claim 8, wherein: After generating the control scheme of the drone nest according to the optimal decision, the method further includes: Calculating a predicted landing time of the flying UAV to the predicted landing point coordinates based on the first flight posture data; The predicted landing time is compared with the obtained adjustment time of the control scheme, and the speed of the flying UAV is controlled according to the comparison result so that the predicted landing time matches the adjustment time.

10. A drone nest control system, characterized in that: include: an acquisition module, configured to acquire first configuration data of a flying drone in response to a target task request, and acquire local configuration data of a drone nest matching the flying drone; an analysis module, configured to determine second configuration data of the flying drone according to the target task request, and compare and analyze the first configuration data and the second configuration data based on natural language technology to obtain a configuration indicator of the flying drone to be replaced; A mobilization module, configured to filter the local configuration data based on the configuration indicator to be replaced, and control the drone nest to execute an internal configuration component mobilization signal according to the screening result, so that the drone nest reaches an initial switching state; a detection module, configured to obtain first flight posture data captured by a first sensor provided on the flying drone when detecting that the position information of the flying drone meets a preset range; A prediction module, configured to predict the landing position of the flying UAV based on the first flight posture data to obtain predicted landing point coordinates; A control module is used to control the drone nest to execute a target switching state signal generated by the initial switching state and the predicted landing point coordinates.

Images