Unmanned aerial vehicle low-altitude flight record management system

Abstract

The invention relates to the technical field of unmanned aerial vehicle management, in particular to an unmanned aerial vehicle low-altitude flight record management system. The invention aims to solve the problems that unmanned aerial vehicle low-altitude flight records are mainly used for flight data storage and simple playback in an existing unmanned aerial vehicle management system, and deep analysis and judgment of the unmanned aerial vehicle low-altitude flight records are lacked. According to the invention, the task evaluation module evaluates the feasibility of the low-altitude flight task of the unmanned aerial vehicle and generates an initial flight scheme, the path optimization module optimizes the initial flight scheme to generate a flight path adapted to a low-altitude environment, and the real-time monitoring module evaluates the flight state of the unmanned aerial vehicle in real time. The recording and judging module evaluates the flight state of the unmanned aerial vehicle after the low-altitude flight task of the unmanned aerial vehicle is completed; and the recording and playback module provides whole-course playback of the low-altitude flight process of the unmanned aerial vehicle according to low-altitude flight records of the unmanned aerial vehicle. Through mutual cooperation among the modules, comprehensive data support and application support are provided for low-altitude flight management of the unmanned aerial vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle management, and in particular to a low-altitude flight record management system for unmanned aerial vehicles. Background Art

[0002] With the rapid development of drone technology, low-altitude drones are increasingly used in urban management, environmental monitoring, emergency rescue and other fields. Currently, drones used for low-altitude flight are usually equipped with miniaturized multi-sensor systems, including high-resolution cameras, lidar, ultrasonic sensors, etc., which can capture surrounding environment information in real time. At the same time, the flight control algorithm based on artificial intelligence can enable drones to achieve autonomous navigation and obstacle avoidance in complex low-altitude environments.

[0003] However, in the current drone management system, the records of drone low-altitude flights generally include basic information such as flight paths, flight time and flight altitude. These records are mainly used for storage and simple playback of flight data, and lack in-depth analysis and evaluation of drone low-altitude flight records. At the same time, most systems only focus on flight data from a single source of drones, lack of effective integration of multi-source data, resulting in insufficient comprehensive grasp of low-altitude flight conditions.

[0004] Therefore, there is an urgent need to develop a UAV low-altitude flight record management system to overcome the shortcomings of the existing technology. Summary of the invention

[0005] In order to solve the problem that the existing UAV management system mainly uses the UAV low-altitude flight records for the storage and simple playback of flight data, but lacks in-depth analysis and evaluation of the UAV low-altitude flight records, the present invention provides such a UAV low-altitude flight record management system, including a task evaluation module, a path optimization module, a real-time monitoring module, a record evaluation module and a record playback module;

[0006] The mission evaluation module is used to receive a low-altitude flight mission request, evaluate the feasibility of the low-altitude flight mission based on multi-source data, and generate an initial flight plan;

[0007] The path optimization module is used to optimize the initial flight plan according to the preliminary flight plan and in combination with low-altitude environmental factors, and generate a flight trajectory adapted to the low-altitude environment;

[0008] The real-time monitoring module is used to collect and integrate the drone status data, environmental perception data and mission execution data during the drone's low-altitude flight, and evaluate the drone's flight status;

[0009] The record evaluation module is used to evaluate the flight performance after the UAV low-altitude flight mission is completed;

[0010] The recording and playback module is used to provide a full playback of the UAV's low-altitude flight process based on the UAV's low-altitude flight record.

[0011] Furthermore, the specific execution steps of the mission evaluation module to evaluate the feasibility of the UAV low-altitude flight mission include:

[0012] S101: receiving a low-altitude flight mission request generated by a user, wherein the mission request includes a target location, a flight starting point, a mission time, an estimated trajectory, and mission requirements;

[0013] S102: Acquire multi-source data related to the low-altitude flight mission, wherein the multi-source data includes environmental data, terrain and obstacle data, airspace restrictions and no-fly zone information, historical low-altitude flight data, and the performance and health status of the UAV itself;

[0014] S103: Evaluate based on the acquired environmental data to determine the impact of environmental variables on the UAV low-altitude flight mission; if the environment does not meet the requirements, provide an alternative time or reject the application; if the environment meets the requirements, proceed to the next step;

[0015] S104: Use GIS to analyze the terrain in the mission area, identify potential obstacles, and further confirm the surrounding terrain status in combination with historical flight data and geographic information database; if there are obstacles that cannot be circumvented, reject the application; if there are no obstacles that cannot be circumvented, proceed to the next step;

[0016] S105: Check the airspace restrictions and regulatory requirements within the mission area, including no-fly zones, restricted-fly zones and flight altitude restrictions; if there are airspace restrictions or regulatory requirements, provide alternative routes or reject the application; if there are no airspace restrictions or regulatory requirements, proceed to the next step;

[0017] S106: Evaluate the performance and health status of the drone, including battery status, load capacity, flight time and communication quality; if the performance or health status of the drone does not meet the requirements, provide an alternative drone or reject the application; if the performance or health status of the drone meets the requirements, proceed to the next step;

[0018] S107: Based on the multi-source data collected, a risk assessment model is used to assess the feasibility of the UAV low-altitude flight mission. If the risk is too high, avoidance suggestions are provided or the application is rejected; the risk assessment model is:

[0019]

[0020] Among them, Risk is the overall risk assessment value; w i is the weight coefficient of each risk factor; f i (D i) is an evaluation function for each risk factor, which is used to represent the impact of multi-source data on the UAV low-altitude flight mission. The impact judgment is evaluated through different machine learning algorithms; D i is the data of the ith risk factor.

[0021] Furthermore, after evaluating the feasibility of the UAV low-altitude flight mission, the mission evaluation module performs three-dimensional modeling and terrain analysis on the target flight area in the UAV low-altitude flight mission based on GIS, and divides the target flight area into near-surface layer, middle layer and upper layer according to the flight performance and mission requirements of the UAV;

[0022] According to the characteristics of the near-ground layer, the middle layer and the upper layer, preliminary flight plans are generated respectively using a path planning algorithm; then the preliminary flight plans of the near-ground layer, the middle layer and the upper layer are fused, and during the fusion, the continuity of the plans at each level in time, space and action is ensured to generate an initial flight plan for the UAV's low-altitude flight mission.

[0023] Furthermore, the specific steps of the path optimization module generating a flight trajectory adapted to the low-altitude environment based on the initial flight plan are:

[0024] S201: The path optimization module divides the initial flight plan into multiple flight segments, and for each flight segment, obtains corresponding environmental data, including terrain, obstacle and meteorological data;

[0025] S202: for each flight segment, determining whether the UAV low-altitude flight trajectory in the initial flight plan is reasonable, if not reasonable, using a probability graph-based path planning algorithm to dynamically generate a feasible optimized UAV low-altitude flight trajectory;

[0026] S203: Fusing the optimized multiple flight segment trajectories in time and space to ensure the continuity and seamless connection between each flight segment, and generate the final UAV low-altitude flight trajectory;

[0027] S204: Send the generated low-altitude flight trajectory of the UAV to the real-time monitoring module, and monitor the execution of the low-altitude flight trajectory of the UAV in real time, and perform secondary optimization when an abnormal situation occurs.

[0028] Furthermore, the real-time monitoring module uses a sliding window to calculate the deviation between the preset low-altitude flight trajectory of the drone and the real-time low-altitude flight trajectory when evaluating the flight status of the drone, dynamically adjusts the size of the sliding window according to the flight speed and environmental changes, and calculates the position deviation, speed deviation and attitude deviation in each sliding window;

[0029] The real-time monitoring module generates a comprehensive deviation vector based on the position deviation, speed deviation and posture deviation, and adjusts the weights of the position deviation, speed deviation and posture deviation according to the mission stage and environmental factors to generate a weighted comprehensive deviation value;

[0030] The real-time monitoring module uses a random forest algorithm to identify deviation patterns of the drone in the deviation data;

[0031] The causes of the deviation are determined by combining the comprehensive deviation vector, weighted comprehensive deviation value and deviation pattern with environmental perception data, mission execution data and UAV status data, including environmental interference, mission changes and equipment failure.

[0032] Furthermore, the real-time monitoring module performs the following operations after determining the cause of the deviation:

[0033] When it is determined that the cause of the deviation is environmental interference, the real-time monitoring module sends out an environmental warning signal to remind the operator to pay attention to relevant environmental changes and adjust the low-altitude flight strategy of the UAV in time according to the real-time environmental monitoring data;

[0034] When it is determined that the cause of the deviation is a mission change, the real-time monitoring module feeds back abnormal condition information to the path optimization module, and the path optimization module replans the low-altitude flight path;

[0035] When it is determined that the cause of the deviation is equipment failure, the real-time monitoring module issues a fault warning message and, based on the fault level of the drone, dispatches a spare drone to replace the current drone or continues the operation of the current drone to perform the task.

[0036] Furthermore, the specific execution steps of the record evaluation module to evaluate the low-altitude flight performance of the drone are:

[0037] S301: Acquire the flight data of the UAV low-altitude flight mission and perform preprocessing operations on the UAV low-altitude flight data;

[0038] S302: Calculate multi-dimensional evaluation indicators for the UAV to perform the UAV low-altitude flight mission;

[0039] S303: extracting multi-dimensional features from the low-altitude flight data of the UAV, including time dimension, space dimension and correlation dimension;

[0040] S304: constructing a flight performance baseline model based on historical low-altitude flight data, comparing the real-time collected UAV low-altitude flight data with the baseline, and identifying abnormal conditions that deviate from the baseline;

[0041] S305: Based on the UAV low-altitude flight data collected by different sensors, a multi-modal fusion low-altitude flight evaluation model is constructed to conduct a comprehensive evaluation of the UAV low-altitude flight performance;

[0042] S306: Determine the low-altitude flight performance level of the UAV using a similarity sorting algorithm;

[0043] S307: Store the evaluation results and related data in the low-altitude flight performance database for reference in subsequent evaluations.

[0044] Furthermore, the low-altitude flight evaluation model is constructed based on the historical low-altitude flight data of the UAV and is used to comprehensively evaluate the low-altitude flight performance of the UAV;

[0045] First, extract feature vectors from the flight data collected by different sensors: GPS feature vector X GPS 、IMU feature vector X IMU , image feature vector X img ,

[0046] The feature vectors of different modes are concatenated to form a joint feature vector x = [X GPS , X IMU , X img ],

[0047] Define the multimodal kernel function K m (x i , x j ) to fuse data from different modalities:

[0048]

[0049] Among them, x i and x j Represents two multimodal input samples, m represents different modalities, w m is the weight of each mode, K m is a kernel function specific to each mode, Y is an adjustable parameter, is the multimodal correlation function,

[0050] Then, the multimodal kernel function is used for analysis to obtain the drone’s low-altitude flight performance level:

[0051]

[0052] Among them, α i is the Lagrange multiplier, y i is the label of the training sample, b is the bias term,

[0053] By calculating the distance between the joint feature vector and the baseline model, abnormal states are detected:

[0054]

[0055] Among them, X base is the feature vector of the baseline model;

[0056] Finally, the similarity score is calculated based on the drone's low-altitude flight performance level and anomaly detection results:

[0057] S(X) = αf(X) + βd(X, X) base )

[0058] Among them, α and β are weight coefficients used to balance the impact of classification results and anomaly detection results.

[0059] Furthermore, the multi-dimensional evaluation indicators of the record evaluation module when evaluating the low-altitude flight performance of the UAV specifically include safety indicators, efficiency indicators, compliance indicators, and task completion quality indicators;

[0060] The safety indicators include flight altitude stability and obstacle avoidance effect, the efficiency indicators include mission completion time and energy consumption rate, the compliance indicators include airspace restriction compliance and flight parameter compliance, and the mission completion quality indicators include data collection accuracy and target recognition accuracy;

[0061] The time dimension in the multi-dimensional extraction of UAV low-altitude flight data includes flight duration, mission execution speed and key event time points; the spatial dimension includes flight trajectory, coverage area and altitude change; and the associated dimension includes equipment status, environmental factors and operation command response.

[0062] Furthermore, the recording and playback module extracts key data of the UAV during low-altitude flight according to the low-altitude flight report generated by the recording and evaluation module, and uses 3D modeling to reproduce the low-altitude flight process of the UAV in a real environment, and displays the dynamic trajectory, flight posture and interaction of the UAV with the surrounding environment;

[0063] When reproducing the low-altitude flight trajectory of a drone, quickly locate the flight trajectory of the drone at the low-altitude flight stage through the timeline, and control the display of the drone's low-altitude flight trajectory through the playback speed and viewing angle, zoom in on local details or switch to different angles to observe the entire low-altitude flight process of the drone;

[0064] The recording and playback module also automatically indicates potential collision risk points between the drone and surrounding obstacles during the display of the drone's low-altitude flight trajectory.

[0065] Beneficial effects:

[0066] 1. The present invention can not only receive the UAV low-altitude flight mission request submitted by the user, but also obtain multi-source data such as real-time weather data, terrain information and airspace restrictions, comprehensively evaluate the feasibility of the UAV low-altitude flight mission, and optimize the initial flight plan to enable the UAV to better adapt to the complex low-altitude environment, effectively ensuring the safety and reliability of the UAV low-altitude flight mission;

[0067] 2. In the process of mission evaluation and path optimization, the present invention will optimize the flight trajectory of the UAV as much as possible, reduce unnecessary energy consumption and time consumption, and at the same time, the real-time monitoring module can timely discover and correct the deviation of the UAV during low-altitude flight, ensuring that the UAV low-altitude flight mission can be carried out in an orderly manner according to the established plan, further improving the execution efficiency of the mission;

[0068] 3. The present invention uses a recording and evaluation module to comprehensively evaluate and record the low-altitude flight performance of the UAV, including multiple dimensions such as safety indicators, efficiency indicators, and compliance indicators, providing rich data support for subsequent analysis and research. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 is a system flow chart of the present invention;

[0070] Figure 2 A flow chart for generating a flight trajectory adapted to a low-altitude environment for the present invention;

[0071] Figure 3 The flowchart of the present invention for evaluating the low-altitude flight performance of a UAV. DETAILED DESCRIPTION

[0072] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0073] A UAV low-altitude flight record management system, as shown in the attached Figure 1 As shown, it includes a task evaluation module, a path optimization module, a real-time monitoring module, a record evaluation module and a record playback module.

[0074] In this embodiment, the mission evaluation module is used to receive a low-altitude flight mission request, evaluate the feasibility of the UAV low-altitude flight mission based on multi-source data, and generate an initial flight plan.

[0075] Specifically, the specific execution steps of the mission assessment module to assess the feasibility of the UAV low-altitude flight mission include:

[0076] S101: receiving a UAV low-altitude flight mission request generated by a user, where the UAV low-altitude flight mission request includes a target location, a flight start point, a mission time, an estimated trajectory, and mission requirements;

[0077] S102: Acquire multi-source data related to the UAV low-altitude flight mission, including but not limited to real-time weather data, environmental information, terrain and obstacle data, airspace restrictions and no-fly zone information, historical flight data, and the performance and health status of the UAV itself;

[0078] S103: Based on the acquired real-time weather data and environmental information, an assessment is made to determine the impact of environmental variables on the UAV low-altitude flight mission; if the weather conditions or environment do not meet the requirements, an alternative time is provided or the application is rejected; if the weather conditions or environment meet the requirements, the next step is carried out;

[0079] S104: Use GIS to analyze the terrain in the mission area, identify potential obstacles, and further confirm the surrounding terrain status in combination with historical low-altitude flight data and geographic information database; if there are obstacles that cannot be circumvented, reject the application; if there are no airspace restrictions or regulatory requirements, proceed to the next step;

[0080] S105: Check the airspace restrictions and regulatory requirements within the mission area, including no-fly zones, restricted-fly zones and flight altitude restrictions; if the airspace restrictions or regulatory requirements within the mission area are not met, that is, when there are airspace restrictions, provide alternative routes or reject the application; if the airspace restrictions or regulatory requirements are met, proceed to the next step;

[0081] S106: Evaluate the performance indicators and health status of the drone, including battery status, load capacity, flight time and communication quality, etc. The battery status is to determine whether the battery power can support the entire low-altitude flight mission, the load capacity is to evaluate whether the drone's load capacity meets the load parameters and mission requirements, and the communication quality is to determine whether the communication quality is stable and reliable in the mission area. If the drone's performance or health status does not meet the requirements, provide an alternative drone or reject the application; if the drone's performance or health status meets the requirements, proceed to the next step;

[0082] S107: Based on the collected multi-source data, use the risk assessment model to evaluate the feasibility of the UAV low-altitude flight mission. If the risk is too high, provide avoidance suggestions or reject the application; the calculation formula of the risk assessment model is:

[0083]

[0084] Among them, Risk is the overall risk assessment value; w i is the weight coefficient of each risk factor; f i (D i ) is an evaluation function for each risk factor, which is used to represent the impact of multi-source data on the UAV low-altitude flight mission. The impact judgment is evaluated through different machine learning algorithms; D i is the data of the ith risk factor.

[0085] Specifically, after evaluating the feasibility of the UAV low-altitude flight mission, the mission assessment module conducts 3D modeling and terrain analysis of the target flight area in the UAV low-altitude flight mission based on GIS, and divides the target flight area into near-surface layer, middle layer and upper layer according to the flight performance and mission requirements of the UAV. Each layer has different environmental characteristics and operation requirements, and requires a targeted low-altitude flight plan for planning and optimization. For the near-ground layer, the focus is on close-range obstacle avoidance and fine control to ensure that the UAV can fly safely and smoothly in complex terrain and dynamic environments. At the same time, it is necessary to update it at a high frequency through perception technology and autonomous decision-making algorithms to adapt to the rapid changes in the low-altitude environment; for the middle layer, the focus is on mission coverage and efficient flight, and the UAV's route and flight time are reasonably planned to meet the needs of the operation mission. It is necessary to balance multiple goals, including flight time, energy consumption and safety, to generate the optimal solution that meets the mission requirements, and at the same time consider the influence of environmental factors such as wind fields and airflows to ensure flight stability and controllability; for the upper layer, the focus is on global planning and the collaborative operation of multiple UAVs, reasonably allocate tasks and resources, and achieve overall efficiency optimization. The upper layer needs to take into account the flight status and mission requirements of each UAV, dynamically adjust the flight plan according to real-time environmental information, ensure coordination and cooperation, and at the same time comply with the control requirements of the target flight area to ensure the legality and compliance of the flight.

[0086] According to the characteristics of the near-surface layer, middle layer and upper layer, the path planning algorithm is used to generate preliminary flight plans respectively; then the preliminary flight plans of the near-surface layer, middle layer and upper layer are integrated, and during the integration, the continuity of the plans at each level in time, space and action is ensured to generate the initial flight plan of the flight mission; after the initial flight plan is generated, the hierarchical planning in the initial flight plan can be improved through simulation testing and on-site verification to improve the reliability and robustness of the initial flight plan.

[0087] In this embodiment, the path optimization module is used to optimize the initial flight plan based on the preliminary flight plan and in combination with low-altitude environmental factors, and generate a flight trajectory that adapts to the low-altitude environment.

[0088] Specifically, the path optimization module generates a low-altitude flight trajectory of the UAV adapted to the low-altitude environment based on the initial flight plan. The specific steps are as follows: Figure 2 Shown are:

[0089] S201: The path optimization module divides the initial flight plan into multiple flight segments, and obtains corresponding environmental data for each flight segment, including terrain, obstacle, weather and other data.

[0090] The division of each flight segment mainly divides the flight mission into several main stages, including the take-off stage, the flight stage, the mission execution stage and the landing stage; in each main stage, the time period is divided into appropriate sub-time periods according to the mission duration and the UAV's endurance, and each sub-time period corresponds to a flight segment; then, in each flight segment, the space is divided into appropriate sub-areas according to the geographical environment characteristics of the flight area, and each sub-area corresponds to a sub-flight segment in the flight segment; in each sub-flight segment, according to the specific mission objectives and flight environment, the mission objectives and obstacle distribution are considered, and the boundaries of the sub-flight segments are improved to ensure that the UAV can fly safely.

[0091] S202: For each flight segment, determine whether the low-altitude flight trajectory of the UAV in the initial flight plan is reasonable. If it is not reasonable, use a probability graph-based path planning algorithm to dynamically generate a feasible optimized low-altitude flight trajectory.

[0092] Among them, when judging whether the low-altitude flight trajectory of the UAV in the initial flight plan is reasonable, the initial UAV low-altitude flight trajectory is detected to see whether there is a risk of colliding with obstacles based on the obtained environmental data including terrain, obstacles and weather; at the same time, combined with the performance parameters of the UAV including the maximum climb / descent speed and the maximum lateral speed, it is detected whether the initial UAV low-altitude flight trajectory meets the UAV's flight capability requirements; finally, considering the mission requirements including flight altitude, range and flight time, it is judged whether the initial UAV low-altitude flight trajectory can meet the UAV flight mission objectives.

[0093] When dynamically optimizing the low-altitude flight trajectory of a UAV, the path planning algorithm based on the probability graph first constructs a probability graph based on environmental data, in which the nodes in the graph represent flyable positions and the edges represent flyable sections. Then, according to the performance parameters and mission requirements of the UAV, a corresponding weight is assigned to each edge to identify the flight cost of the path segment. Finally, the path planning algorithm is used to search for the optimal low-altitude flight trajectory of the UAV in the probability graph.

[0094] S203: Fusing the optimized multiple flight segment trajectories in time and space to ensure the continuity and seamless connection between each flight segment, and generate the final UAV low-altitude flight trajectory.

[0095] When the flight segment trajectories are merged, the trajectory of each flight segment is synchronized in time. The timestamps of the trajectory points can be adjusted so that the end time of the adjacent segments and the start time of the next segment are completely consistent, ensuring the seamless connection of each segment in time; then, for the adjacent UAV low-altitude flight segment trajectories, interpolation, fitting and other methods are used at the intersection of the two segments to generate a smooth transition trajectory, eliminate mutation points, and ensure the continuity of the UAV low-altitude flight process. Finally, the optimized overall UAV low-altitude flight trajectory is smoothed to remove sharp changes in direction and speed, making the UAV low-altitude flight trajectory more continuous and stable, so as to reduce the operating load on the UAV.

[0096] S204: Send the generated UAV low-altitude flight trajectory to the real-time monitoring module, and monitor the execution of the UAV's low-altitude flight trajectory in real time. When an abnormal situation occurs, perform secondary optimization.

[0097] In this embodiment, the real-time monitoring module is used to collect and integrate the drone status data, environmental perception data and mission execution data during the drone's low-altitude flight, and evaluate the drone's flight status.

[0098] Specifically, when evaluating the flight status of the UAV, the real-time monitoring module uses an active window to calculate the deviation between the preset low-altitude flight trajectory of the UAV and the real-time low-altitude flight trajectory, and dynamically adjusts the size of the sliding window according to the flight speed and environmental changes. Within each sliding window, the position deviation, speed deviation and attitude deviation are calculated; wherein, the size of the sliding window can be dynamically adjusted according to the current flight speed of the UAV. The greater the flight speed, the larger the window is to capture more trajectory data; at the same time, the size of the sliding window can also be adjusted according to the current environmental complexity. For example, in complex situations with multiple obstacles or high wind speeds, a smaller window is used to more finely monitor the real-time low-altitude flight trajectory of the UAV.

[0099] The real-time monitoring module generates a comprehensive deviation vector based on position deviation, velocity deviation and attitude deviation. At the same time, due to the different degrees of influence of the deviations of each dimension of the comprehensive deviation vector on the overall flight safety in different mission stages and environmental conditions, different weights are given to the deviations of different dimensions to generate a weighted comprehensive deviation value, which is used to judge the severity of the current flight status, so as to more accurately reflect the comprehensive deviation degree of the current flight status.

[0100] The real-time monitoring module uses the random forest algorithm to identify the deviation mode of the drone in the deviation data; the deviation modes include high-frequency oscillation mode, gradual drift mode, sudden jump mode and compound mode.

[0101] The causes of the deviation are determined by combining the comprehensive deviation vector, weighted comprehensive deviation value and deviation pattern with environmental perception data, mission execution data and UAV status data, including environmental interference, mission changes and equipment failure.

[0102] Specifically, the real-time monitoring module performs the following operations after determining the cause of the deviation:

[0103] When it is determined that the cause of the deviation is environmental interference, the real-time monitoring module sends out an environmental warning signal to remind the operator to pay attention to relevant environmental changes, and adjust the UAV's low-altitude flight strategy in time according to the real-time environmental monitoring data, and automatically trigger the path optimization module to re-plan the UAV's low-altitude flight trajectory to avoid harsh environments; the real-time monitoring module can also dynamically reduce the size of the sliding window, improve the monitoring resolution, and track real-time flight status changes more accurately.

[0104] When it is determined that the cause of the deviation is a mission change, the real-time monitoring module will feedback the abnormal condition information to the path optimization module, and the path optimization module will re-plan the UAV's low-altitude flight path, or the real-time monitoring module will automatically modify the mission instructions to guide the UAV to perform a new flight mission; the real-time monitoring module can also dynamically adjust the sliding window to cover the new preset UAV low-altitude flight trajectory, and promptly capture the deviation of the UAV after switching waypoints.

[0105] When it is determined that the cause of the deviation is equipment failure, the real-time monitoring module issues a fault warning message and, based on the level of the UAV failure, dispatches a backup UAV to replace the current UAV or continues the task by the current UAV. The real-time monitoring module can also appropriately increase the sliding window and comprehensively consider flight data over a longer period of time to more reliably diagnose the cause of the failure.

[0106] In this embodiment, the recording and judging module is used to evaluate the flight performance after the UAV low-altitude flight mission is completed.

[0107] Specifically, the specific execution steps of the record evaluation module to evaluate the low-altitude flight performance of the UAV are as follows: Figure 3 Shown are:

[0108] S301: Acquire low-altitude flight data of a UAV low-altitude flight mission, and perform preprocessing operations on the UAV low-altitude flight data;

[0109] S302: Calculate the multi-dimensional evaluation indicators of the UAV performing low-altitude flight missions; the multi-dimensional evaluation indicators include safety indicators, efficiency indicators, compliance indicators and mission completion quality indicators, and the safety indicators include flight altitude stability and obstacle avoidance effect, the efficiency indicators include mission completion time and energy consumption rate, the compliance indicators include airspace restriction compliance and flight parameter compliance, and the mission completion quality indicators include data collection accuracy and target recognition accuracy;

[0110] S303: Extract multi-dimensional features from the drone low-altitude flight data, including time dimension, space dimension and association dimension; the time dimension includes flight duration, mission execution speed and key event time points; the space dimension includes the drone low-altitude flight trajectory, coverage area and altitude change; the association dimension includes device status, environmental factors and operation command response;

[0111] S304: Build a flight performance baseline model based on historical low-altitude flight data, compare the real-time collected flight data with the baseline, and identify abnormal conditions that deviate from the baseline by setting dynamic thresholds and adapting according to mission types and environmental conditions;

[0112] S305: Based on the UAV low-altitude flight data collected by different sensors, a multi-modal fusion low-altitude flight evaluation model is constructed to conduct a comprehensive evaluation of the UAV low-altitude flight performance;

[0113] S306: Determine the low-altitude flight performance level of the UAV using a similarity sorting algorithm;

[0114] S307: The evaluation results and related data are stored in the flight performance database for reference in subsequent evaluations.

[0115] Specifically, the low-altitude flight evaluation model is built based on the historical low-altitude flight data of the UAV and is used to comprehensively evaluate the flight performance of the UAV;

[0116] First, extract feature vectors from different modal data, i.e., the low-altitude flight data of drones collected by different sensors: GPS feature vector X GPS 、IMU feature vector X IMU , image feature vector X img ,

[0117] The feature vectors of different modes are concatenated to form a joint feature vector x = [X GPS , X IMU , X img ],

[0118] Define the multimodal kernel function K m (x i , x j ) to fuse data from different modalities:

[0119]

[0120] Among them, x i and x j Represents two multimodal input samples, m represents different modalities, w m is the weight of each mode, K mis a kernel function specific to each mode, γ is an adjustable parameter, is the multimodal correlation function,

[0121] Then use the multimodal kernel function to analyze and get the flight performance level:

[0122]

[0123] Among them, α i is the Lagrange multiplier, y i is the label of the training sample, b is the bias term,

[0124] By calculating the distance between the joint feature vector and the baseline model, abnormal states are detected:

[0125]

[0126] Among them, X base is the feature vector of the baseline model;

[0127] Finally, based on the drone low-altitude flight performance registration and anomaly detection results, the similarity score is calculated:

[0128] S(X)=αf(X)+βd(X,X base )

[0129] Among them, α and β are weight coefficients used to balance the impact of classification results and anomaly detection results.

[0130] In this embodiment, the recording and playback module is used to provide a full playback of the drone's low-altitude flight process based on the drone's low-altitude flight record.

[0131] Specifically, the recording and playback module extracts key data of the drone during low-altitude flight based on the drone low-altitude flight report generated by the recording and evaluation module, and uses 3D modeling to reproduce the drone's low-altitude flight process in a real environment, showing the drone's dynamic trajectory, flight posture, and interaction with the surrounding environment; at the same time, the recording and playback module simulates the scene when the fault occurs during the trajectory reproduction process based on the fault information recorded in the flight report, helping users analyze the cause of the fault;

[0132] When reproducing the flight trajectory, the recording and playback module provides a visual timeline and key event markers, allowing users to use the timeline to quickly locate the drone's low-altitude flight trajectory and view the drone's key flight stages; at the same time, users can also control the playback speed and viewing angle, and control the display of the drone's low-altitude flight trajectory by manually adjusting the speed ratio, automatically slowing down local details, and switching different viewing angles with one click, etc., to zoom in on local details or switch to different angles to observe the entire low-altitude flight process of the drone;

[0133] The recording and playback module also provides a more intuitive visual display during the display of the UAV's low-altitude flight trajectory, including the use of red highlight marks and three-dimensional model dynamic rendering to automatically identify potential collision risk points between the UAV and surrounding obstacles.

[0134] The above-mentioned embodiments only express the preferred implementation modes of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present invention. It should be pointed out that, for a person skilled in the art, several modifications, improvements and substitutions can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention patent shall be subject to the attached claims.

Claims

1. A UAV low-altitude flight record management system, characterized in that: It includes task evaluation module, path optimization module, real-time monitoring module, record evaluation module and record playback module; The mission evaluation module is used to receive a low-altitude flight mission request, evaluate the feasibility of the low-altitude flight mission based on multi-source data, and generate an initial flight plan; The path optimization module is used to optimize the initial flight plan according to the preliminary flight plan and in combination with low-altitude environmental factors, and generate a flight trajectory adapted to the low-altitude environment; The real-time monitoring module is used to collect and integrate the drone status data, environmental perception data and mission execution data during the drone's low-altitude flight, and evaluate the drone's flight status; The record evaluation module is used to evaluate the flight performance after the UAV low-altitude flight mission is completed; The recording and playback module is used to provide a full playback of the UAV's low-altitude flight process based on the UAV's low-altitude flight record.

2. The unmanned aerial vehicle low-altitude flight record management system according to claim 1 is characterized in that: The specific execution steps of the mission assessment module for assessing the feasibility of the UAV low-altitude flight mission include: S101: receiving a low-altitude flight mission request generated by a user, wherein the mission request includes a target location, a flight starting point, a mission time, an estimated trajectory, and mission requirements; S102: Acquire multi-source data related to the flight mission, wherein the multi-source data includes environmental data, terrain and obstacle data, airspace restrictions and no-fly zone information, historical low-altitude flight data, and the performance and health status of the UAV itself; S103: Evaluate based on the acquired environmental data to determine the impact of environmental variables on the UAV low-altitude flight mission; if the environment does not meet the requirements, provide an alternative time or reject the application; if the environment meets the requirements, proceed to the next step; S104: Use GIS to analyze the terrain in the mission area, identify potential obstacles, and further confirm the surrounding terrain status in combination with historical low-altitude flight data and geographic information database; if there are obstacles that cannot be circumvented, reject the application; if there are no obstacles that cannot be circumvented, proceed to the next step; S105: Check the airspace restrictions and regulatory requirements within the mission area, including no-fly zones, restricted-fly zones and flight altitude restrictions; if there are airspace restrictions or regulatory requirements, provide alternative routes or reject the application; if there are no airspace restrictions or regulatory requirements, proceed to the next step; S106: Evaluate the performance and health status of the drone, including battery status, load capacity, flight time and communication quality; if the performance or health status of the drone does not meet the requirements, provide an alternative drone or reject the application; if the performance or health status of the drone meets the requirements, proceed to the next step; S107: Based on the multi-source data collected, a risk assessment model is used to assess the feasibility of the UAV low-altitude flight mission. If the risk is too high, avoidance suggestions are provided or the application is rejected; the risk assessment model is: Among them, Risk is the overall risk assessment value; w i is the weight coefficient of each risk factor; f i (D i ) is an evaluation function for each risk factor, which is used to represent the impact of multi-source data on the UAV low-altitude flight mission. The impact judgment is evaluated through different machine learning algorithms; D i is the data of the ith risk factor.

3. The unmanned aerial vehicle low-altitude flight record management system according to claim 2 is characterized in that: After evaluating the feasibility of the UAV low-altitude flight mission, the mission evaluation module performs three-dimensional modeling and terrain analysis on the target flight area in the UAV low-altitude flight mission based on GIS, and divides the target flight area into near-surface layer, middle layer and upper layer according to the flight performance and mission requirements of the UAV; According to the characteristics of the near-ground layer, the middle layer and the upper layer, preliminary flight plans are generated respectively using a path planning algorithm; then the preliminary flight plans of the near-ground layer, the middle layer and the upper layer are fused, and during the fusion, the continuity of the plans at each level in time, space and action is ensured to generate an initial flight plan for the UAV's low-altitude flight mission.

4. The unmanned aerial vehicle low-altitude flight record management system according to claim 1 is characterized in that: The specific steps of the path optimization module generating a flight trajectory adapted to the low-altitude environment based on the initial flight plan are as follows: S201: The path optimization module divides the initial flight plan into multiple flight segments, and for each flight segment, obtains corresponding environmental data, including terrain, obstacle and meteorological data; S202: for each flight segment, determining whether the UAV low-altitude flight trajectory in the initial flight plan is reasonable, if not reasonable, using a probability graph-based path planning algorithm to dynamically generate a feasible optimized UAV low-altitude flight trajectory; S203: Fusing the optimized multiple flight segment trajectories in time and space to ensure the continuity and seamless connection between each flight segment, and generate the final UAV low-altitude flight trajectory; S204: Send the generated low-altitude flight trajectory of the UAV to the real-time monitoring module, and monitor the execution of the low-altitude flight trajectory of the UAV in real time, and perform secondary optimization when an abnormal situation occurs.

5. The unmanned aerial vehicle low-altitude flight record management system according to claim 1 is characterized in that: The real-time monitoring module uses a sliding window to calculate the deviation between the preset low-altitude flight trajectory of the drone and the real-time low-altitude flight trajectory when evaluating the flight status of the drone, dynamically adjusts the size of the sliding window according to the flight speed and environmental changes, and calculates the position deviation, speed deviation and attitude deviation in each sliding window; The real-time monitoring module generates a comprehensive deviation vector based on the position deviation, speed deviation and posture deviation, and adjusts the weights of the position deviation, speed deviation and posture deviation according to the mission stage and environmental factors to generate a weighted comprehensive deviation value; The real-time monitoring module uses a random forest algorithm to identify deviation patterns of the drone in the deviation data; The causes of the deviation are determined by combining the comprehensive deviation vector, weighted comprehensive deviation value and deviation pattern with environmental perception data, mission execution data and UAV status data, including environmental interference, mission changes and equipment failure.

6. The unmanned aerial vehicle low-altitude flight record management system according to claim 5 is characterized in that: The real-time monitoring module performs the following operations after determining the cause of the deviation: When it is determined that the cause of the deviation is environmental interference, the real-time monitoring module sends out an environmental warning signal to remind the operator to pay attention to relevant environmental changes and adjust the low-altitude flight strategy of the UAV in time according to the real-time environmental monitoring data; When it is determined that the cause of the deviation is a mission change, the real-time monitoring module feeds back abnormal condition information to the path optimization module, and the path optimization module replans the low-altitude flight path; When it is determined that the cause of the deviation is equipment failure, the real-time monitoring module issues a fault warning message and, based on the fault level of the drone, dispatches a spare drone to replace the current drone or continues the operation of the current drone to perform the task.

7. The unmanned aerial vehicle low-altitude flight record management system according to claim 1 is characterized in that: The specific execution steps of the record evaluation module to evaluate the low-altitude flight performance of the drone are: S301: Acquire the flight data of the UAV low-altitude flight mission and perform preprocessing operations on the UAV low-altitude flight data; S302: Calculate multi-dimensional evaluation indicators for the UAV to perform the UAV low-altitude flight mission; S303: extracting multi-dimensional features from the low-altitude flight data of the UAV, including time dimension, space dimension and correlation dimension; S304: constructing a flight performance baseline model based on historical low-altitude flight data, comparing the real-time collected UAV low-altitude flight data with the baseline, and identifying abnormal conditions that deviate from the baseline; S305: Based on the UAV low-altitude flight data collected by different sensors, a multi-modal fusion low-altitude flight evaluation model is constructed to conduct a comprehensive evaluation of the UAV low-altitude flight performance; S306: Determine the low-altitude flight performance level of the UAV using a similarity sorting algorithm; S307: Store the evaluation results and related data in the low-altitude flight performance database for reference in subsequent evaluations.

8. The unmanned aerial vehicle low-altitude flight record management system according to claim 7 is characterized in that: The low-altitude flight evaluation model is constructed based on the historical low-altitude flight data of the UAV and is used to comprehensively evaluate the low-altitude flight performance of the UAV; First, extract feature vectors from the flight data collected by different sensors: GPS feature vector X GPS 、IMU feature vector X IMU , image feature vector X img , The feature vectors of different modes are concatenated to form a joint feature vector x = [X GPS , X IMU , X img ], Define the multimodal kernel function K m (x i , x j ) to fuse data from different modalities: Among them, x i and x j Represents two multimodal input samples, m represents different modalities, w m is the weight of each mode, K m is a kernel function specific to each mode, γ is an adjustable parameter, is the multimodal correlation function, Then, the multimodal kernel function is used for analysis to obtain the performance level of the drone’s low-altitude flight: Among them, α i is the Lagrange multiplier, y i is the label of the training sample, b is the bias term, and the abnormal state is detected by calculating the distance between the joint feature vector and the baseline model: Among them, X base is the feature vector of the baseline model; Finally, the similarity score is calculated based on the drone's low-altitude flight performance level and anomaly detection results: S(X)=αf(X)+βd(X,X base ) Among them, α and β are weight coefficients used to balance the impact of classification results and anomaly detection results.

9. The unmanned aerial vehicle low-altitude flight record management system according to claim 7, characterized in that: The multi-dimensional evaluation indicators of the record evaluation module when evaluating the low-altitude flight performance of the UAV specifically include safety indicators, efficiency indicators, compliance indicators and task completion quality indicators; The safety indicators include flight altitude stability and obstacle avoidance effect, the efficiency indicators include mission completion time and energy consumption rate, the compliance indicators include airspace restriction compliance and flight parameter compliance, and the mission completion quality indicators include data collection accuracy and target recognition accuracy; The time dimension in the multi-dimensional extraction of UAV low-altitude flight data includes flight duration, mission execution speed and key event time points; the spatial dimension includes flight trajectory, coverage area and altitude change; and the associated dimension includes equipment status, environmental factors and operation command response.

10. The unmanned aerial vehicle low-altitude flight record management system according to claim 1, characterized in that: The recording and playback module extracts key data of the UAV during low-altitude flight according to the low-altitude flight report generated by the recording and evaluation module, and uses 3D modeling to reproduce the low-altitude flight process of the UAV in a real environment, and displays the dynamic trajectory, flight posture and interaction of the UAV with the surrounding environment; When reproducing the low-altitude flight trajectory of a drone, quickly locate the flight trajectory of the drone at the low-altitude flight stage through the timeline, and control the display of the drone's low-altitude flight trajectory through the playback speed and viewing angle, zoom in on local details or switch to different angles to observe the entire low-altitude flight process of the drone; The recording and playback module also automatically indicates potential collision risk points between the drone and surrounding obstacles during the display of the drone's low-altitude flight trajectory.

Images