Unmanned aerial vehicle intelligent scheduling and management method based on task planning

Abstract

The invention discloses an unmanned aerial vehicle intelligent scheduling and management method based on task planning, and relates to the technical field of dynamic path planning, and the method comprises the steps: collecting flight data, and carrying out the preprocessing; based on the preprocessed data, constructing an intelligent patrol dynamic convolutional network model to analyze the health state of the unmanned aerial vehicle, and generating a health state result; performing dynamic path planning through a recurrent neural network model by using the generated health state result and combining environment perception data; making a task and executing a plan according to dynamic path planning and an intelligent scheduling algorithm; according to the method, hierarchical feature extraction and fusion are performed on the preprocessed flight data through the intelligent patrol dynamic convolutional network model, the health state feature vector is generated, and accurate health state evaluation and anomaly detection are realized in combination with the health state value and the One-Class SVM.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic path planning, and in particular to an intelligent scheduling and management method for unmanned aerial vehicles based on mission planning. Background Art

[0002] With the rapid development of drone technology, its application scope has expanded from traditional aerial photography and entertainment to logistics and transportation, agricultural monitoring, environmental protection, emergency rescue and other fields; however, in multi-task and high-complexity application environments, the intelligent scheduling and management of drones face many challenges; existing technologies usually rely on static path planning algorithms and simple task allocation mechanisms, which lack the ability to respond to dynamic environmental changes in real time; in addition, for drone health status monitoring, traditional methods are mostly based on rule thresholds or simple statistical analysis, which makes it difficult to fully capture the health status characteristics of complex systems, resulting in low reliability of task execution under abnormal conditions; with the increasing complexity of task planning and the uncertainty of the drone operating environment, the limitations of existing technologies become more and more obvious.

[0003] In response to the above problems, the existing technologies have the following main deficiencies: First, there is a lack of intelligent analysis of the health status of the drone, and potential faults or abnormal conditions cannot be identified in a timely manner, which may lead to mission failure or safety hazards; second, the path planning method does not make sufficient use of environmental perception data, and fails to fully consider the real-time changes of dynamic interference factors (such as obstacles and weather conditions); third, the drone task scheduling and execution strategy lacks global optimization capabilities and fails to effectively integrate the drone's health status, mission requirements and environmental conditions. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a UAV intelligent scheduling and management method based on task planning to solve the health status detection problem.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, the present invention provides an intelligent scheduling and management method for drones based on task planning, which includes collecting flight data and performing preprocessing; based on the preprocessed data, constructing a smart patrol dynamic convolutional network model to analyze the health status of the drone and generate health status results; using the generated health status results, combined with environmental perception data, to perform dynamic path planning through a recurrent neural network model; based on the dynamic path planning, an intelligent scheduling algorithm formulates tasks and executes plans.

[0008] As a preferred solution of the UAV intelligent scheduling and management method based on mission planning described in the present invention, the flight data includes GPS location data, speed data, acceleration data, battery power data and environmental perception data.

[0009] As a preferred solution of the method for intelligent scheduling and management of unmanned aerial vehicles based on task planning described in the present invention, the preprocessing includes data cleaning, removing outliers and noise and normalizing the data.

[0010] As a preferred solution of the method for intelligent scheduling and management of unmanned aerial vehicles based on task planning described in the present invention, the specific steps of constructing the intelligent patrol dynamic convolutional network model are as follows:

[0011] Define the input layer of the Zhixun dynamic convolutional network model as preprocessed data;

[0012] Divide the preprocessed data into training set, validation set and test set;

[0013] Extract local features of data through dynamic convolutional layers;

[0014] Add a pooling layer to extract the main features of the data;

[0015] The local features and main features are combined into a one-dimensional feature vector, and the high-order feature relationship is learned through the fully connected layer to complete feature fusion and generate a health status feature vector;

[0016] Define the output layer of the Zhixun dynamic convolutional network model, and calculate the health status value based on the health status feature vector. The expression is:

[0017]

[0018] Among them, F(X) is the health status value, d is the dimension of the health status feature vector, and x i is the actual value of the i-th feature, μ is the average of all features, i is the index of the feature, and ∏ is the product symbol. is the factor that adjusts the scale of the results;

[0019] Set judgment thresholds;

[0020] If F(X)≤T, the health status is judged to be abnormal;

[0021] If F(X)>T, the health status is judged to be normal;

[0022] One-Class SVM is introduced to detect the health status feature vector and output the health status result;

[0023] Use the training set to train the Zhixun dynamic convolutional network model and optimize the classification task;

[0024] Use the validation set to evaluate the performance of the Zhixun dynamic convolutional network model and optimize hyperparameters;

[0025] The test set is used to comprehensively evaluate the Zhixun dynamic convolutional network model and One-Class SVM.

[0026] As a preferred solution of the intelligent scheduling and management method of unmanned aerial vehicles based on task planning described in the present invention, wherein: the introduction of One-Class SVM to detect the health status feature vector and output the health status result, the specific steps are:

[0027] According to the generated health status value and the set judgment threshold, the normal state feature vector samples are screened out as the training data of One-Class SVM;

[0028] Use RBF radial basis kernel as the kernel function of One-Class SVM and train One-Class SVM to learn the distribution of normal state feature vector samples;

[0029] In the testing phase, the health status feature vector is input into the trained One-Class SVM;

[0030] One-Class SVM determines whether the test sample belongs to the normal state according to the distribution of the normal state feature vector samples and outputs the healthy state result.

[0031] As a preferred solution of the intelligent scheduling and management method of unmanned aerial vehicles based on task planning described in the present invention, wherein: the generated health status results are combined with environmental perception data to perform dynamic path planning through a recurrent neural network model, and the specific steps are as follows:

[0032] According to the flight mission, the environmental perception data and health status value are combined as the input of path planning;

[0033] Calculate the priority of path adjustment, the expression is:

[0034]

[0035] Among them, A(P) is the path adjustment priority, P is the evaluation target of the current path planning, e is the total number of environmental factors, j is the index of the environmental factor, and E j is the emergency weight of the jth environmental factor, D j is the distance between the UAV and the jth environmental factor, k is the distance adjustment factor, and m is the result adjustment factor;

[0036] If the drone’s health status is abnormal, the return path to the base will be given priority directly;

[0037] If the drone is in normal health, it will prioritize avoiding high-risk areas;

[0038] The recurrent neural network model is used to process the time series data of environmental perception data and, combined with the health status value, to generate a dynamic optimization path node sequence as a dynamic path planning result;

[0039] Send the dynamic path planning results as navigation instructions to the UAV navigation center;

[0040] The UAV navigation center receives navigation instructions and parses them to generate flight parameters;

[0041] The drone executes navigation instructions and adjusts flight parameters in real time to avoid obstacles;

[0042] If the environmental conditions in the environmental perception data change, the path is recalculated and the navigation instructions are updated.

[0043] As a preferred solution of the intelligent scheduling and management method of unmanned aerial vehicles based on task planning described in the present invention, the time series data of environmental perception data is processed by a recurrent neural network model, and a dynamic optimization path node sequence is generated as a dynamic path planning result in combination with a health status value. The specific steps are as follows:

[0044] Define the first layer of the recurrent neural network model, input environmental perception data and health status values, and extract time series features;

[0045] Define the middle layer of the recurrent neural network model, introduce the LSTM long short-term memory network to capture the long-term dependencies in the time series, and extract global features;

[0046] Define the last layer of the recurrent neural network model, combine the extracted time series features with the global features, and output a dynamically optimized path node sequence;

[0047] Define loss functions to avoid high-risk areas, optimize path length, and constrain health status;

[0048] The dynamically optimized path node sequence generated by the recurrent neural network model definition is used as the dynamic path planning result.

[0049] As a preferred solution of the intelligent scheduling and management method of unmanned aerial vehicles based on task planning described in the present invention, wherein: the intelligent scheduling algorithm formulates tasks and executes plans based on dynamic path planning, and the specific steps are:

[0050] Input flight mission, dynamic path planning results and drone health status to initialize the intelligent scheduling algorithm;

[0051] Generate task allocation scheme through iterative optimization and output task allocation plan;

[0052] Send the mission allocation plan to the UAV navigation center to start the flight mission;

[0053] Through drone communication, real-time monitoring of flight status during flight missions;

[0054] After the drone completes the mission, it uploads the flight mission execution results to the central processing center and updates the mission status.

[0055] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the method for intelligent scheduling and management of unmanned aerial vehicles based on task planning as described in the first aspect of the present invention is implemented.

[0056] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the method for intelligent scheduling and management of unmanned aerial vehicles based on task planning as described in the first aspect of the present invention.

[0057] The beneficial effects of the present invention are as follows: the present invention uses the intelligent patrol dynamic convolutional network model to perform hierarchical feature extraction and fusion on the pre-processed flight data, generates a health status feature vector, and combines the health status value and One-Class SVM to achieve accurate health status assessment and anomaly detection; the model uses a dynamic convolution layer to extract local features, a pooling layer to extract main features, and a fully connected layer to learn high-order feature relationships to ensure the comprehensiveness and accuracy of health status analysis; the health status value is quantified by a formula, and combined with One-Class SVM to efficiently handle scenarios with insufficient abnormal samples, the reliability of UAV health monitoring is significantly improved; the intelligent patrol dynamic convolutional network model provides an accurate health status basis for dynamic path planning and intelligent scheduling, effectively ensuring the safety and stability of task execution. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0059] Figure 1 This is a flow chart of the intelligent scheduling and management method of unmanned aerial vehicles based on task planning in Example 1;

[0060] Figure 2 This is a schematic diagram of formulating a flight mission in Example 1. DETAILED DESCRIPTION

[0061] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0062] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0063] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0064] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a method for intelligent scheduling and management of unmanned aerial vehicles based on task planning, comprising the following steps:

[0065] S1. Collect flight data and perform preprocessing.

[0066] Furthermore, flight data, including GPS location data, speed data, acceleration data, battery power data and environmental perception data;

[0067] GPS location data is collected by the GPS receiver equipped with the drone, which is used to receive satellite signals and determine the geographic location of the drone, including longitude, latitude and altitude;

[0068] Speed ​​data is collected by an airspeed meter, which measures the speed of the drone relative to the air. It can be calculated from GPS data and can provide ground speed (the speed of the drone relative to the ground).

[0069] Acceleration data is collected by an accelerometer, which can measure the acceleration of the drone on three axes (X, Y, and Z). This data can be used to calculate the dynamic changes of the drone;

[0070] Battery power data is collected by the battery management center, which monitors the battery voltage, current and temperature to ensure battery safety and provide power information;

[0071] Environmental perception data includes weather conditions, obstacle information, and mission area information;

[0072] Weather conditions obtain temperature data, wind speed and direction data, and humidity data through meteorological sensors;

[0073] Temperature data is used to determine whether extreme temperatures affect drone performance;

[0074] Wind speed and direction data are used to assess flight stability and the feasibility of path planning;

[0075] Humidity data determines the impact of humidity on flight equipment (such as rain and fog);

[0076] Obstacle information: Obstacle location and size are acquired in real time through onboard sensors (such as lidar and ultrasonic sensors);

[0077] The distance and angle of the obstacle’s position relative to the drone;

[0078] The size of the obstacle helps determine the avoidance priority;

[0079] Mission area information is obtained by combining mission planning objectives with terrain features and no-fly zones through terrain maps, predefined mission areas, or visual sensors;

[0080] Topographic features such as undulating terrain, steep slopes, etc.

[0081] No-fly zones such as legally restricted areas or high-risk areas;

[0082] All collected data are transmitted to the central processing center in real time through the wireless communication of the drone as subsequent input.

[0083] S1.1. Preprocessing, specifically,

[0084] Preprocessing includes data cleaning, removing outliers, noise and normalizing the data;

[0085] Specifically, Kalman filtering is used to clean the data, remove outliers and noise, and normalize the data;

[0086] Set the initial state estimate and initial error covariance matrix;

[0087] Predict the next state based on dynamics;

[0088] Calculate the error covariance of the predicted state;

[0089] receiving new measurements;

[0090] Calculate the Kalman gain, which is a weight that weighs the predicted state against the measured value;

[0091] Update the state estimate, combining the predicted state and the measured value;

[0092] Update the error covariance matrix;

[0093] Identify and address outliers that deviate significantly from model predictions or historical data, which may be caused by sensor failures, external interference, or data transmission errors;

[0094] For each feature (such as speed, acceleration, etc.), determine its minimum and maximum value in the entire data set;

[0095] For each data point, normalization is performed using the formula, expressed as:

[0096]

[0097] Among them, x norm is the normalized data, x is the original data, and x max is the maximum value of the data, x min is the minimum value of the data;

[0098] When new data points are added, they are normalized using the same minimum and maximum values ​​to ensure data consistency;

[0099] It should be noted that by collecting GPS position, speed, acceleration, battery power and environmental perception data, the operating status and environmental information of the UAV are fully obtained to ensure that the data dimensions are rich and cover key flight factors; through the preprocessing step, Kalman filtering is used to remove noise and outliers, which significantly improves the accuracy and reliability of the data. At the same time, the data scale is unified through normalization processing, which enhances the comparability of data with different characteristics and the adaptability of the model; this process provides high-quality input data for subsequent health status analysis, path planning and task scheduling, ensuring the stability of the system and the accuracy of intelligent decision-making.

[0100] S2. Based on the preprocessed data, a dynamic convolutional network model is constructed to analyze the health status of the drone and generate health status results.

[0101] Furthermore, the input layer of the Zhixun dynamic convolutional network model is defined as the preprocessed data;

[0102] The preprocessed data is divided into training set, validation set and test set (70% for training set, 15% for validation set and 15% for test set);

[0103] The training set is used to train the Zhixun dynamic convolutional network model;

[0104] The validation set is used to evaluate model performance and tune hyperparameters;

[0105] The test set is used for comprehensive evaluation of the final model;

[0106] Extract local features of data through dynamic convolutional layers;

[0107] Adding pooling layers (such as max pooling or average pooling) to reduce dimensionality, extract the main features of the data, and reduce unnecessary noise;

[0108] The role of the pooling layer is to reduce the feature dimension while retaining the most important information;

[0109] The local features and main features are combined into a one-dimensional feature vector, and the high-order feature relationship is further learned through the fully connected layer to complete feature fusion and generate a health status feature vector;

[0110] Define the output layer of the Zhixun dynamic convolutional network model, and calculate the health status value (normal state or abnormal state) according to the health status feature vector to quantify the health status of the drone. The expression is:

[0111]

[0112] Among them, F(X) is the health status value, the abnormality score indicates the abnormality of data point X. The larger the value, the higher the degree of deviation of the data point from the normal mode. d is the dimension of the health status feature vector, that is, the number of features, and x i is the actual value of the ith feature. In drone health management applications, this could be a sample of a sensor reading (such as velocity, acceleration, etc.), and μ is the average of all features (x i The mean of the values ​​in the normal sample) is a statistic used to measure the central tendency of the data set. i is the index of the feature, which is used to traverse each feature in the health status feature vector. ∏ is the product symbol, which means multiplying the squares of the deviations of all features. is a factor that adjusts the scale of the results and is used to normalize the range of health status values;

[0113] The formula quantifies the health status based on the square product of the feature deviation from the mean. It comprehensively evaluates the degree of abnormality by combining the degree of deviation of each feature, avoiding the limitation of single feature abnormality. At the same time, it uses the product operation to enhance the sensitivity to abnormal correlation between multi-dimensional features. It introduces a normalization factor to ensure the normalization of the result range and facilitate the determination of abnormalities. It improves the comprehensiveness, accuracy and robustness of health status assessment.

[0114] Set the judgment threshold, T = 0.75 (the specific threshold can be set according to actual conditions);

[0115] If F(X)≤T, the health status value is too low, indicating that the sample deviates from the normal state, and the health status is judged to be abnormal;

[0116] If F(X)>T, the health status value is high, indicating that the sample is within the normal range, and the health status is determined to be normal, and its health status feature vector is input into the trained One-Class SVM for further confirmation;

[0117] For example, assuming that the health status value F(X) of a sample is 0.6, which means that 0.6≤0.75, then the health status of the sample is judged to be abnormal;

[0118] Assuming that the health status value of another sample F(X) = 1.2, which means 1.2≤0.75, the health status of the sample is determined to be normal, and its health status feature vector is input into the One-Class SVM for further confirmation;

[0119] One-Class SVM is introduced to detect the health status feature vector and output the health status result;

[0120] One-Class SVM is used as an anomaly detector to detect abnormal samples that deviate from the normal distribution by learning the distribution of normal samples. It complements the classification ability of the dynamic convolutional network to further improve the accuracy of health status detection;

[0121] One-Class SVM is an unsupervised or one-class classification algorithm based on support vector machine (SVM). It is mainly used in anomaly detection or novelty detection scenarios. Unlike traditional vector machines (SVM), One-Class SVM focuses on the data distribution of a single category and tries to identify which data points belong to this category (normal class) and which data points are outliers (abnormal class).

[0122] Use the training set to train the Zhixun dynamic convolutional network model, optimize the classification task (normal / abnormal state), including the dynamic convolutional network and One-Class SVM parts, monitor the changes in the loss function, and ensure model convergence;

[0123] Use the validation set to evaluate the performance of the Zhixun dynamic convolutional network model and optimize hyperparameters (such as convolution kernel size, learning rate, SVM kernel function, etc.);

[0124] Regularly adjust the parameters of the Zhixun dynamic convolutional network model (such as learning rate, batch size, etc.) to improve the classification accuracy and anomaly detection capabilities of the model. This step helps prevent overfitting and find the best Zhixun dynamic convolutional network model configuration;

[0125] Use grid search and random search to find the best hyperparameter combination;

[0126] The dynamic convolution network tuning parameters include convolution kernel size, number of layers, learning rate, Dropout ratio, etc.

[0127] The tuning parameters of One-Class SVM are kernel function type (such as RBF kernel), regularization parameter, etc.

[0128] Use the test set to fully evaluate the Zhixun dynamic convolutional network model and One-Class SVM to check their performance on unseen data. This step can provide a true estimate of the model's generalization ability.

[0129] Record the classification performance of the Zhixun dynamic convolutional network model on the test set, including key indicators such as accuracy, recall rate, and F1 score;

[0130] Dynamic convolutional networks are used to classify health status and output normal or abnormal status;

[0131] The results of the health status assessment formula are used to further assist in determining the degree of abnormality;

[0132] One-Class SVM is used to detect abnormal samples with non-normal distribution and improve the accuracy of anomaly detection.

[0133] S2.1. Introduce One-Class SVM to detect the health status feature vector and output the health status result, specifically,

[0134] According to the generated health status value and the set judgment threshold, the normal state feature vector samples are screened out as the training data of One-Class SVM; ensure that the training data does not contain abnormal samples so that One-Class SVM can accurately learn the distribution characteristics of "normal" data;

[0135] Use RBF radial basis kernel as the kernel function of One-Class SVM to train One-Class SVM to learn the distribution of normal state feature vector samples. RBF kernel function is suitable for high-dimensional data with nonlinear distribution and can effectively capture complex feature patterns.

[0136] In the test phase, the health status feature vector is input into the trained One-Class SVM. The One-Class SVM determines whether the test sample is in a normal state based on the distribution of the normal state feature vector sample and outputs a binary health status result (normal or abnormal).

[0137] It should be noted that by combining the Zhixun dynamic convolutional network model and One-Class SVM, accurate classification and anomaly detection of the drone's health status are achieved; the dynamic convolutional network layer extracts local and high-order features, generates health status feature vectors, and quantifies the degree of abnormality through the health status value formula; One-Class SVM further learns the normal sample distribution, effectively detects abnormal samples that deviate from the normal distribution, and makes up for the limitation of insufficient abnormal samples; it significantly improves the accuracy and robustness of health status assessment, provides a reliable basis for the dynamic path planning and task scheduling of drones, and enhances the safety and stability of task execution.

[0138] S3. Using the generated health status results and combining them with environmental perception data, dynamic path planning is performed through a recurrent neural network model.

[0139] Furthermore, according to the flight mission, the weather conditions, obstacle information and mission area information of the environmental perception data, as well as the normal and abnormal states of the health status value are combined as inputs for path planning;

[0140] Calculate the priority of path adjustment, the expression is:

[0141]

[0142] Among them, A(P) is the path adjustment priority, which is used to determine the urgency of path optimization, P is the evaluation target of the current path planning, e is the total number of environmental factors, j is the index of environmental factors, E j is the emergency weight of the jth environmental factor, D j is the distance between the UAV and the jth environmental factor, k is the distance adjustment factor, and m is the result adjustment factor, which is used to adjust the nonlinearity of the priority calculation result;

[0143] The formula comprehensively evaluates the priority of path adjustment through the weighted sum of the emergency weight of environmental factors and the distance of the drone. The weight reflects the importance of risk, and the distance factor enhances the sensitivity to close-range high-risk factors. The nonlinear adjustment factor is introduced and the flexibility of the formula is improved to adapt to different mission requirements, improve the accuracy of path optimization as a whole, and ensure the safety and efficiency of drone flights.

[0144] Setting the preferred adjustment threshold;

[0145] A(P)≤A, the path adjustment priority is high, and the path needs to be adjusted immediately or instantly (such as avoiding obstacles or returning to the base);

[0146] A(P)>A, the path adjustment priority is low, and the current path can be continued to execute the task;

[0147] For example (specific settings are made according to actual conditions), if A = 0.3, it is a high-risk task and the path needs to be adjusted immediately to ensure the safety of the drone (such as avoiding obstacles or returning to the base);

[0148] If A=0.5, it is a medium-risk task and the path needs to be adjusted in time, but the sensitivity to path adjustment is appropriately reduced;

[0149] If A=0.7, it is a low-risk task and you can continue on the current path and execute the task;

[0150] If the health status is abnormal or there are high-risk factors in the environment (such as obstacles too close or wind speeds exceeding the standard), path adjustment will be triggered first;

[0151] If the drone's health status is abnormal, the return path will be given the highest priority (stop the current mission, plan the shortest time and lowest risk return path, avoid high-risk areas (such as obstacles, bad weather, no-fly zones, etc.), and ensure that the path meets constraints such as power and safety margin);

[0152] If the health status is normal, the path priority is calculated based on the environmental perception data, and priority is given to avoiding areas with higher risks (such as environmental obstacles, such as buildings, electric poles, trees, etc., which may pose a threat to the flight of the drone, severe weather areas such as strong winds, rain, sandstorms, etc., which will affect the stability and navigation accuracy of the drone, areas with weak wireless signals, where the communication signal between the drone and the navigation center is weak or even interrupted, which may cause the mission to be out of control or navigation failure, malicious interference areas, where there is electronic interference (such as jamming signals) or GPS signal failure, which may cause the drone to be unable to navigate normally, and no-fly areas, where laws and regulations clearly stipulate that flying is prohibited, such as near airports and over military bases, etc.);

[0153] The criterion for determining environmental obstacles is the distance D between the obstacle and the drone. j Smaller (i.e., dangerous distance), and the emergency weight E j Higher;

[0154] The criteria for determining severe weather areas are weather data detected by the environmental sensing system, such as wind speed exceeding a set threshold (e.g., 10 m / s), precipitation or humidity reaching a dangerous level, etc.

[0155] The criteria for determining a weak wireless signal area is that the communication signal strength is lower than the set threshold (e.g. RSSI < -90dBm);

[0156] The criteria for determining a severe interference area are that the GPS signal is lost or the perceived interference signal strength exceeds a set threshold (for example, the GPS signal loss threshold C / N0 < 25dB-Hz or no signal for 5 consecutive seconds, the interference signal strength threshold J / S > 15dB or the interference signal power > -100dBm);

[0157] The criteria for determining a no-fly zone is to combine electronic map data and the coordinates of the no-fly zone to determine whether the drone is close to the no-fly zone;

[0158] The recurrent neural network model is used to process the time series data of environmental perception data and, combined with the health status value, to generate a dynamic optimization path node sequence as a dynamic path planning result;

[0159] Send the dynamic path planning results as navigation instructions to the UAV navigation center;

[0160] The UAV navigation center receives navigation instructions and parses them to generate flight parameters;

[0161] The drone executes navigation instructions and adjusts flight parameters in real time to avoid obstacles or high-risk areas;

[0162] Static obstacles include buildings, high-rise buildings, bridges, trees, electric poles, street lights, and terrain factors (such as mountains and cliffs);

[0163] Dynamic obstacles are other drones or aircraft, ground vehicles, and birds or other flying or moving animals;

[0164] High-risk areas include areas with severe weather, areas with GPS signal failure, areas with electronic interference, no-fly zones, and environmental obstacles;

[0165] If the environmental conditions in the environmental perception data change, the path planning will recalculate the path and update the navigation instructions in combination with the health status value and the new environmental data, and update the navigation instructions through the UAV navigation center. The UAV will adjust the flight parameters in real time according to the updated path, including speed, heading and flight altitude, to ensure the safety and efficiency of mission execution.

[0166] S3.1. Use the recurrent neural network model to process the time series data of environmental perception data, and combine it with the health status value to generate a dynamic optimization path node sequence as the dynamic path planning result, specifically,

[0167] Define the first layer of the recurrent neural network model, input environmental perception data and health status values, extract time series features, and analyze short-term change trends;

[0168] Define the middle layer of the recurrent neural network model, introduce the LSTM long short-term memory network to capture the long-term dependencies in the time series (for example, identify the trend of wind speed changes or the dynamic laws of obstacle movement), predict future environmental change trends (such as dynamic obstacle movement, the possibility of health status deterioration), and extract global features;

[0169] Define the last layer of the recurrent neural network model, combine the extracted time series features with the global features, and output the dynamically optimized path node sequence (including Node i GPS coordinates, flight altitude and flight direction);

[0170] Define loss functions to avoid high-risk areas, optimize path length, and constrain health status;

[0171] Avoiding high-risk areas means that the path nodes are as far away from obstacles or dangerous areas as possible;

[0172] Optimizing path length can reduce the distance between path nodes (saving energy);

[0173] Constraining the health status ensures that the path planning is consistent with the health status of the drone (e.g., returning first when an abnormality occurs);

[0174] Through the recurrent neural network model processing, a set of dynamic optimization path node sequences are generated as the dynamic path planning results;

[0175] Specifically including: GPS coordinates of path nodes, flight altitude and flight direction;

[0176] The GPS coordinates of its path nodes are the predicted geographical locations (latitude and longitude coordinates) where the drone needs to go next;

[0177] The flight altitude is dynamically adjusted according to the location of obstacles and the health status of the drone;

[0178] The flight direction is to determine the flight direction between each path node to avoid unnecessary detours;

[0179] It should be noted that by combining the health status value with environmental perception data (such as weather, obstacles, and mission area information), the recurrent neural network model is used to dynamically plan the UAV flight path, and the optimal generation of path nodes is achieved; LSTM captures time series characteristics, identifies environmental change trends, and ensures the flexibility and accuracy of path planning; prioritizes avoiding high-risk areas and adjusting the path according to the health status, returns to the base in abnormal situations, and optimizes the flight altitude and direction in normal situations to reduce mission risks; real-time updates of path planning to respond to environmental changes, improve the safety, energy efficiency, and mission execution success rate of UAV flights, and adapt to complex dynamic environments.

[0180] S4. Based on dynamic path planning, the intelligent scheduling algorithm formulates tasks and executes plans.

[0181] Furthermore, the flight mission, dynamic path planning results and health status of the UAV are input to initialize the intelligent scheduling algorithm (such as genetic algorithm or particle swarm optimization algorithm);

[0182] Genetic Algorithm:

[0183] Chromosome encoding, each chromosome represents a task allocation scheme (mapping relationship between drones and tasks);

[0184] The fitness function comprehensively considers task completion rate, path length, power limit and task priority;

[0185] Optimize task allocation schemes through selection, crossover, and mutation operations;

[0186] Particle Swarm Optimization Algorithm:

[0187] Each particle represents a task allocation scheme;

[0188] Consistent with genetic algorithms, it is used to evaluate task allocation schemes;

[0189] Optimize task allocation through dynamic adjustment of particle positions and velocities;

[0190] Generate the optimal task allocation plan through iterative optimization;

[0191] Output the task allocation plan, including the mapping relationship between drones and tasks;

[0192] Send the mission allocation plan to the UAV navigation center to start the flight mission (including the mission objectives, dynamic path sequence, navigation instructions and flight parameters of each UAV);

[0193] The mission objectives of each drone, such as the specific execution location;

[0194] Flight parameters, such as flight altitude and speed;

[0195] Through UAV communication, real-time monitoring of flight status during flight missions (including flight status monitoring, mission progress feedback, and exception handling);

[0196] Flight status monitoring is to update the drone’s flight position, battery power, speed and other status information in real time;

[0197] Task progress feedback is to report the completion progress of the task (such as the number of completed path nodes);

[0198] Exception handling means immediately adjusting the mission plan or terminating the mission if an abnormal drone health status is detected (such as insufficient battery or failure);

[0199] After the drone completes the mission, it uploads the flight mission execution results (such as inspection photos, data collection records, etc.) to the central processing center and updates the mission status ("Completed");

[0200] It should be noted that the intelligent scheduling algorithm optimizes the UAV task allocation plan (such as genetic algorithm or particle swarm optimization algorithm) through dynamic path planning results, health status and task requirements, comprehensively considers the task completion rate, path length, power limit and task priority to ensure the allocation of optimal resources; monitors the task execution status in real time, dynamically adjusts the task plan to deal with abnormal conditions (such as insufficient power or health abnormalities), and improves the flexibility and safety of task execution; uploads the results and updates the status after the task is completed to achieve a closed loop of task management.

[0201] This embodiment also provides a computer device, which is suitable for the case of a drone intelligent scheduling and management method based on mission planning, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the drone intelligent scheduling and management method based on mission planning as proposed in the above embodiment.

[0202] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0203] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for realizing intelligent scheduling and management of unmanned aerial vehicles based on task planning as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, disk or optical disk.

[0204] In summary, the present invention uses the intelligent patrol dynamic convolutional network model to perform hierarchical feature extraction and fusion on the pre-processed flight data, generates a health status feature vector, and combines the health status value and One-Class SVM to achieve accurate health status assessment and anomaly detection; the model uses a dynamic convolution layer to extract local features, a pooling layer to extract main features, and a fully connected layer to learn high-order feature relationships to ensure the comprehensiveness and accuracy of health status analysis; the health status value is quantified by a formula, and combined with One-Class SVM to efficiently handle scenarios with insufficient abnormal samples, the reliability of UAV health monitoring is significantly improved; the intelligent patrol dynamic convolutional network model provides an accurate health status basis for dynamic path planning and intelligent scheduling, effectively ensuring the safety and stability of task execution.

[0205] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent dispatching and management method for UAVs based on mission planning, characterized by: include, Collect flight data and pre-process it; Based on the pre-processed flight data, a dynamic convolutional network model is constructed to analyze the health status of the drone and generate health status results; Using the generated health status results and combining them with environmental perception data, dynamic path planning is performed through a recurrent neural network model; Based on dynamic path planning, intelligent scheduling algorithms formulate tasks and execute plans.

2. The method for intelligent dispatching and management of unmanned aerial vehicles based on mission planning according to claim 1 is characterized in that: The flight data includes GPS location data, speed data, acceleration data, battery power data and environmental perception data.

3. The method for intelligent dispatching and management of unmanned aerial vehicles based on mission planning according to claim 2 is characterized in that: The preprocessing includes data cleaning, removing outliers and noise and normalizing the data.

4. The method for intelligent dispatching and management of unmanned aerial vehicles based on mission planning according to claim 3 is characterized in that: The specific steps of constructing the intelligent patrol dynamic convolutional network model are as follows: Define the input layer of the Zhixun dynamic convolutional network model as preprocessed data; Divide the preprocessed data into training set, validation set and test set; Extract local features of data through dynamic convolutional layers; Add a pooling layer to extract the main features of the data; The local features and main features are combined into a one-dimensional feature vector, and the high-order feature relationship is learned through the fully connected layer to complete feature fusion and generate a health status feature vector; Define the output layer of the Zhixun dynamic convolutional network model, and calculate the health status value based on the health status feature vector. The expression is: Among them, F(X) is the health status value, d is the dimension of the health status feature vector, and x i is the actual value of the i-th feature, μ is the average of all features, i is the index of the feature, and ∏ is the product symbol. is the factor that adjusts the scale of the results; Set judgment thresholds; If F(X)≤T, the health status is judged to be abnormal; If F(X)>T, the health status is judged to be normal; One-Class SVM is introduced to detect the health status feature vector and output the health status result; Use the training set to train the Zhixun dynamic convolutional network model and optimize the classification task; Use the validation set to evaluate the performance of the Zhixun dynamic convolutional network model and optimize hyperparameters; The test set is used to comprehensively evaluate the Zhixun dynamic convolutional network model and One-Class SVM.

5. The method for intelligent dispatching and management of unmanned aerial vehicles based on mission planning according to claim 4 is characterized in that: The One-Class SVM is introduced to detect the health status feature vector and output the health status result. The specific steps are: According to the generated health status value and the set judgment threshold, the normal state feature vector samples are screened out as the training data of One-Class SVM; Use RBF radial basis kernel as the kernel function of One-Class SVM and train One-Class SVM to learn the distribution of normal state feature vector samples; In the testing phase, the health status feature vector is input into the trained One-Class SVM; One-Class SVM determines whether the test sample belongs to the normal state according to the distribution of the normal state feature vector samples and outputs the healthy state result.

6. The method for intelligent dispatching and management of unmanned aerial vehicles based on mission planning according to claim 5 is characterized in that: The generated health status results are combined with environmental perception data to perform dynamic path planning through a recurrent neural network model. The specific steps are: According to the flight mission, the environmental perception data and health status value are combined as the input of path planning; Calculate the priority of path adjustment, the expression is: Among them, A(P) is the path adjustment priority, P is the evaluation target of the current path planning, e is the total number of environmental factors, j is the index of the environmental factor, and E j is the emergency weight of the jth environmental factor, D j is the distance between the UAV and the jth environmental factor, k is the distance adjustment factor, and m is the result adjustment factor; If the drone’s health status is abnormal, the return path to the base will be given priority directly; If the drone is in normal health, it will prioritize avoiding high-risk areas; The recurrent neural network model is used to process the time series data of environmental perception data and, combined with the health status value, to generate a dynamic optimization path node sequence as a dynamic path planning result; Send the dynamic path planning results as navigation instructions to the UAV navigation center; The UAV navigation center receives navigation instructions and parses them to generate flight parameters; The drone executes navigation instructions and adjusts flight parameters in real time to avoid obstacles; If the environmental conditions in the environmental perception data change, the path is recalculated and the navigation instructions are updated.

7. The method for intelligent dispatching and management of unmanned aerial vehicles based on mission planning according to claim 6 is characterized in that: The method of using a recurrent neural network model to process the time series data of environmental perception data and combining it with the health status value to generate a dynamic optimization path node sequence as a dynamic path planning result is specifically as follows: Define the first layer of the recurrent neural network model, input environmental perception data and health status values, and extract time series features; Define the middle layer of the recurrent neural network model, introduce the LSTM long short-term memory network to capture the long-term dependencies in the time series, and extract global features; Define the last layer of the recurrent neural network model, combine the extracted time series features with the global features, and output a dynamically optimized path node sequence; Define loss functions to avoid high-risk areas, optimize path length, and constrain health status; The dynamically optimized path node sequence generated by the recurrent neural network model definition is used as the dynamic path planning result.

8. The method for intelligent dispatching and management of unmanned aerial vehicles based on mission planning according to claim 7 is characterized in that: According to the dynamic path planning, the intelligent scheduling algorithm formulates tasks and executes the plan. The specific steps are: Input flight mission, dynamic path planning results and drone health status to initialize the intelligent scheduling algorithm; Generate task allocation scheme through iterative optimization and output task allocation plan; Send the mission allocation plan to the UAV navigation center to start the flight mission; Through drone communication, real-time monitoring of flight status during flight missions; After the drone completes the mission, it uploads the flight mission execution results to the central processing center and updates the mission status.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for intelligent scheduling and management of unmanned aerial vehicles based on mission planning as described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for intelligent scheduling and management of unmanned aerial vehicles based on mission planning described in any one of claims 1 to 8 are implemented.

Images