An intelligent analysis method based on a drone data base

Abstract

This invention relates to the field of intelligent analysis technology for unmanned aerial vehicles (UAVs), and discloses an intelligent analysis method based on a UAV data base. The method includes acquiring flight data such as the UAV's position, speed, and altitude, and storing it in the UAV data base; then, calculating the distribution density of the UAV in a target area in real time based on this flight data to obtain distribution density analysis results; next, combining environmental monitoring data, using these results to assess potential risks and hazards during flight; subsequently, generating path optimization decisions based on these risks and hazards, and dynamically adjusting the decisions in response to user interaction to generate personalized flight plans; finally, integrating multi-dimensional analysis tools to perform statistical analysis on flight trajectory data within a specific area, generating real-time analysis results. This method achieves unified management of UAV data, improves the accuracy of density calculation and the comprehensiveness of risk assessment, enhances the flexibility of path optimization, and provides strong support for UAV operations.

Description

Technical Field

[0001] This invention relates to the field of intelligent analysis technology for unmanned aerial vehicles (UAVs), specifically to an intelligent analysis method based on a UAV data platform. Background Technology

[0002] With the rapid development of drone technology, its applications have broadly covered multiple fields such as surveying and exploration, environmental monitoring, emergency rescue, and logistics transportation. In actual operation, the effective management and utilization of drone flight data has become a crucial factor affecting operational efficiency and safety. Currently, although most drone systems can collect basic data such as position, speed, and altitude during flight, this data is often scattered across different devices or platforms, lacking a unified data integration and management mechanism. This results in poor data sharing, low utilization, and difficulty in achieving comprehensive control over the drone's flight status.

[0003] In drone swarm operations, real-time monitoring of drone density within a target area is crucial. Existing technologies often rely on preset parameters or static models for density calculation, failing to dynamically adjust based on real-time flight data. This leads to discrepancies between calculated results and actual conditions, impacting the accuracy of subsequent decisions. Furthermore, risk assessments during drone flights tend to focus on single factors, such as airspace restrictions or equipment malfunctions, neglecting comprehensive analysis incorporating environmental monitoring data. This makes it difficult to fully identify potential risks and increases the probability of flight safety accidents.

[0004] In terms of path optimization, traditional methods typically generate a single optimized path based on fixed algorithms, lacking responsiveness to user interactions. When users need to modify the path according to actual operational requirements, existing systems struggle to quickly and dynamically adjust, failing to generate flight plans that meet individual user needs and reducing operational flexibility and adaptability. Furthermore, for the analysis of UAV flight trajectory data within a specific area, existing tools are mostly limited to single-dimensional statistics, lacking the integration of multi-dimensional analysis tools. This makes it difficult to extract more valuable information from the trajectory data and provides a comprehensive reference for subsequent operational optimization or decision-making.

[0005] These problems hinder the full realization of the technological advantages of drones in practical applications, thus restricting the further development of the drone industry towards intelligence and efficiency. Therefore, there is an urgent need for an intelligent analysis method capable of unified management of drone data, real-time density calculation, comprehensive risk assessment, dynamic path optimization, and multi-dimensional trajectory analysis to address the shortcomings of current technology. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent analysis method based on a UAV data base to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides an intelligent analysis method based on a UAV data base, the method comprising:

[0008] Acquire drone flight data, including location information, speed information, and altitude information, and store the drone flight data in the drone data base;

[0009] Based on the UAV flight data, the distribution density of the UAV in the target area is calculated in real time to obtain the distribution density analysis results.

[0010] Using the distribution density analysis results and combined with environmental monitoring data, the potential risks and hazards during the flight of the UAV are assessed.

[0011] Based on the aforementioned potential risks and hazards, a path optimization decision is generated;

[0012] In response to user interaction, the path optimization decision is dynamically adjusted to generate a personalized flight plan;

[0013] It integrates multi-dimensional analysis tools to perform statistical analysis on flight trajectory data within a specific area and generate real-time analysis results.

[0014] Preferably, the process involves acquiring UAV flight data, including location information, speed information, and altitude information, and storing the UAV flight data in a UAV data base, including:

[0015] Real-time acquisition of drone flight data via sensor networks;

[0016] The UAV flight data is preprocessed, including noise filtering and data formatting, to obtain standardized flight data;

[0017] The standardized flight data is uploaded to the UAV data base for centralized storage.

[0018] Preferably, based on the UAV flight data, the distribution density of UAVs in the target area is calculated in real time to obtain the distribution density analysis results, including:

[0019] Using the location information in the standardized flight data, the target area is divided into grid cells;

[0020] Calculate the number of drones in each grid cell and generate an initial distribution density map;

[0021] By combining time series analysis, the initial distribution density map is dynamically updated to obtain the distribution density analysis results.

[0022] Preferably, the distribution density analysis results are used in conjunction with environmental monitoring data to assess potential risks and hazards during the flight of the UAV, including:

[0023] Acquire environmental monitoring data, including meteorological and topographic data;

[0024] Statistical models were used to analyze the correlation between the distribution density analysis results and environmental monitoring data to identify abnormal flight patterns.

[0025] Based on the aforementioned abnormal flight pattern, potential risks and hazards are assessed, including collision risk and environmental impact risk.

[0026] Preferably, based on the potential risks and hazards, a path optimization decision is generated, including:

[0027] Based on the aforementioned potential risks and hazards, high-priority adjustment areas have been identified;

[0028] The drone paths within the high-priority adjustment area are replanned to generate initial optimized paths;

[0029] The initial optimized path is subjected to connectivity verification using a network optimization algorithm to obtain a path optimization decision.

[0030] Preferably, in response to user interaction, the path optimization decision is dynamically adjusted to generate a personalized flight plan, including:

[0031] Parse user-input interaction commands, including manual parameter adjustments;

[0032] Based on the interactive instructions, modify the path parameters in the path optimization decision;

[0033] By using real-time simulation technology, the modified path parameters are verified and processed to generate personalized flight plans.

[0034] Preferably, multi-dimensional analysis tools are integrated to perform statistical analysis on flight trajectory data within a specific area, generating real-time analysis results, including:

[0035] Extract flight trajectory data within a specific area from the drone's data base;

[0036] Calculate the average velocity and density distribution of flight trajectory data to generate basic statistics;

[0037] By combining the personalized flight plan with the basic statistics, contextual integration processing is performed to obtain real-time analysis results.

[0038] Preferably, the basic statistics are subjected to context integration processing to obtain real-time analysis results, including:

[0039] Using data fusion technology, the basic statistics are compared with historical flight data;

[0040] Generate dynamic trend reports to reflect changes in flight behavior;

[0041] Integrate the dynamic trend report into the real-time analysis results.

[0042] Preferably, after generating the dynamic trend report, the method further includes:

[0043] Based on the dynamic trend report, update the flight strategy library in the UAV data base;

[0044] In response to new drone flight data, the process of generating the distribution density analysis results is retried.

[0045] Preferably, in response to new UAV flight data, the process of re-triggering the generation of the distribution density analysis results includes:

[0046] When new drone flight data input is detected, repeat the distribution density calculation step;

[0047] The path optimization decision generation process is optimized using the updated distribution density analysis results.

[0048] Compared with the prior art, the beneficial effects of the present invention are:

[0049] This intelligent analysis method based on a UAV data platform first achieves unified storage and management of UAV flight data by constructing a UAV data platform, changing the previous situation of scattered data storage and low utilization. The unified data platform can integrate scattered flight data, enabling efficient data retrieval and providing a reliable data source for subsequent density calculations, risk assessments, and other stages. It avoids analysis delays or result deviations caused by data dispersion, improving the consistency and efficiency of the entire analysis process.

[0050] In the drone distribution density calculation stage, this method performs dynamic calculations based on real-time acquired flight data. Compared to traditional calculation methods that rely on static models or preset parameters, it can more accurately reflect the actual distribution of drones within the target area. The real-time updated distribution density analysis results allow operators to promptly grasp the distribution status of the drone swarm, avoiding collision risks caused by excessive density or incomplete operational coverage caused by insufficient density, thus ensuring the orderly conduct of swarm operations.

[0051] In terms of risk assessment, this method combines distribution density analysis results with environmental monitoring data, overcoming the limitations of traditional single-factor assessments. The introduction of environmental monitoring data allows for a comprehensive consideration of various factors affecting drone flight safety, making the identification of potential risks more comprehensive and accurate. For example, under severe weather conditions, combining drone distribution density can promptly identify cascading risks that may be triggered by weather factors in high-density areas, enabling proactive countermeasures to reduce the likelihood of accidents and improve the safety of drone flights.

[0052] In terms of path optimization and scheme generation, this method can not only generate initial path optimization decisions based on potential risks and hazards, but also dynamically adjust in response to user interaction, thereby generating personalized flight schemes. This dynamic adjustment capability can fully meet the personalized needs of users in actual operations. When faced with temporary task changes or sudden environmental changes, users can quickly adjust the flight path through interactive operations without relying on complex system remodeling or algorithm reconstruction, greatly improving the flexibility and adaptability of operations and ensuring that UAV operations can better meet the needs of actual scenarios.

[0053] This method integrates multi-dimensional analysis tools to statistically analyze flight trajectory data within a specific area and generate real-time analysis results. The application of these tools breaks through the limitations of traditional single-dimensional analysis, enabling the extraction of data value from multiple perspectives of flight trajectories. The real-time generated analysis results provide operators with more comprehensive operational feedback, such as optimizing operational area division by analyzing trajectory overlap and adjusting flight parameters by analyzing speed change trends, further improving the efficiency and rationality of drone operations and driving the development of drone applications towards greater intelligence. Attached Figure Description

[0054] Figure 1 This is a schematic diagram illustrating the working principle of the intelligent analysis method based on a UAV data base as described in this invention.

[0055] Figure 2 A flowchart for drone flight data acquisition and storage;

[0056] Figure 3 A flowchart for assessing potential risks and hazards in drone flight;

[0057] Figure 4 A flowchart generated for personalized flight plans. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] Please see Figure 1 This invention provides an intelligent analysis method based on a UAV data platform, the method comprising:

[0060] By acquiring drone flight data, including location, speed, and altitude information, and storing this data in the drone data base; based on the drone flight data, the distribution density of drones in the target area is calculated in real time to obtain distribution density analysis results; using the distribution density analysis results, combined with environmental monitoring data, potential risks and hazards during drone flight are assessed; based on potential risks and hazards, path optimization decisions are generated; responding to user interaction, the path optimization decisions are dynamically adjusted to generate personalized flight plans; and multi-dimensional analysis tools are integrated to statistically analyze flight trajectory data within a specific area, generating real-time analysis results.

[0061] See Figure 2 The system's operation begins with a sensor network comprised of various types of sensors, widely deployed in the target airspace and within the UAV swarm itself. This network includes, but is not limited to, GPS / BeiDou modules, inertial measurement units (IMUs), and barometers carried by the UAVs, as well as ground-based auxiliary positioning base stations and meteorological monitoring stations. These sensors continuously capture raw flight data from the UAVs, such as latitude and longitude coordinates, ground velocity, rate of climb, altitude, and heading angle, in a high-frequency synchronous manner. This data is transmitted to the ground control system's data receiving interface in near real-time via encrypted data transmission links and mobile communication networks.

[0062] Before entering the storage stage, the raw data stream undergoes a series of rigorous preprocessing steps. A noise filtering stage is initiated, employing a digital filter based on the Kalman filter algorithm to smooth position and velocity information, effectively suppressing abnormal data fluctuations caused by satellite signal multipath effects, sensor transient errors, or electromagnetic interference. The data formatting module standardizes the filtered data, converting data from different sensor sources and with different protocol formats into a standardized data format defined internally by the system. This format explicitly specifies that latitude and longitude are represented in decimal degrees, velocity is in meters per second, altitude is in meters and based on mean sea level, and timestamps are uniformly Coordinated Universal Time (UTC) accurate to milliseconds. This process transforms the chaotic and heterogeneous raw data stream into standardized flight data with unified dimensions, timing, and structure, improving data processability and consistency.

[0063] After preprocessing, the standardized flight data is uploaded to the UAV data platform, a distributed time-series database with high throughput and low latency, responsible for centralized storage, indexing, and management of massive amounts of flight data. Each standardized flight data record contains a timestamp, a unique UAV identifier, and various standardized fields, and is indexed in a multi-dimensional manner according to time series and spatial location to support subsequent efficient queries and aggregation calculations.

[0064] With the support of the data foundation, the real-time distribution density calculation module begins operation. Based on the latitude and longitude boundaries of the target area, it divides the airspace into uniformly sized square grid cells on a horizontal plane, with each grid cell representing a fixed geographical area. The system uses location information from standardized flight data acquired in real time to calculate and determine the specific grid cell in which each UAV is located at any given moment. The system periodically (e.g., every second) scans all online UAVs and determines their grid affiliation, accumulating the number of UAVs present in each grid cell. Based on the ratio of this number to the grid area, the system generates an initial distribution density map in grid units, which visually displays the density of UAV distribution within the airspace in the form of a heatmap.

[0065] To ensure the density distribution analysis results reflect dynamic changes, the system incorporates time-series analysis. The initial density distribution map is not a static snapshot but rather a dynamically updated data layer. The system overlays multiple initial density distribution maps from continuous time series, calculating the density change trend, average density, and peak density for each grid cell within a sliding time window. By analyzing density evolution over time, the system can identify the aggregation direction, diffusion speed, or persistent hovering area of ​​the UAV swarm. For example, the system might identify a rapid increase in density within a grid cell, indicating that UAVs are converging on that area. Ultimately, the system outputs a dynamic density distribution analysis result that integrates real-time status and short-term historical trends. This result includes not only the current density value but also derived information such as the rate and direction of density change, providing richer and more comprehensive spatial situational awareness for subsequent risk assessment.

[0066] See Figure 3 The system initiates a risk assessment module by first accessing multi-dimensional environmental monitoring data streams from external data sources. Meteorological data originates from a real-time API interface provided by the National Meteorological Information Center, transmitted in JSON format, and includes minute-by-minute updates of wind speed and direction, temperature, relative humidity, precipitation probability, and visibility data within the target airspace. Topographic data utilizes a Digital Elevation Model (DEM) from the Geographic Information System, extracting elevation, slope, and aspect information for each latitude and longitude coordinate point within the target area. All environmental data is assigned the same timestamp and spatial reference system as the UAV flight data, achieving a unified spatiotemporal benchmark.

[0067] The distribution density analysis results serve as the core input, and their data structure includes fields such as grid cell number, center point latitude and longitude, real-time number of drones, density value per unit area, and density change rate over the past five minutes. The system spatially aligns and temporally synchronizes the environmental monitoring data of each grid cell with its corresponding distribution density data. The alignment process uses bilinear interpolation to convert discrete meteorological station data into gridded data; topographic data is directly extracted from the elevation values ​​using the grid center point coordinates.

[0068] The statistical model employs multiple linear regression analysis as its basic framework, using the drone density of a specific grid cell as the dependent variable and the wind speed, precipitation probability, and elevation of that grid cell as independent variables to construct a regression equation. The system performs rolling calculations on data over a continuous ten-minute period to determine the regression coefficients of each environmental factor on the density distribution. For example, in a mountainous area grid, the system might calculate that for every 100-meter increase in elevation, the drone density decreases by 0.8 drones / km²; and for every 5 m / s increase in wind speed, the density decreases by 1.2 drones / km². These quantitative relationships reveal the spatial constraints imposed by environmental factors on drone distribution.

[0069] Abnormal flight patterns are identified through a dual mechanism. The system trains a density prediction model based on historical data, and generates the expected density range for each grid after inputting current environmental parameters. When the measured density value exceeds the upper or lower limit of the prediction range by 15% for three consecutive minutes, a primary anomaly flag is triggered. A mutation detection algorithm scans the environmental data in real time. If extreme changes are detected, such as a sudden increase of 4 levels in wind speed within 30 seconds or a 50% decrease in visibility within 2 minutes, a deep analysis is forcibly initiated regardless of whether the current density is abnormal.

[0070] For marked anomalous grids, the system performs risk coupling analysis, with collision risk assessment focusing on areas of abnormal density: when the measured density of a grid exceeds the airspace safety threshold (e.g., 25 drones per square kilometer) and continues to increase, the system calculates the minimum distance between adjacent drones. If the distance is lower than the aircraft's safe buffer distance (e.g., 15 meters for consumer drones), a collision risk is identified. Simultaneously, the system detects the velocity vector distribution of drones within the anomalous grids. If multiple drones are found to have a heading intersection angle greater than 45 degrees and a relative speed exceeding 10 meters per second, the collision risk level is further upgraded.

[0071] Environmental impact risk assessment focuses on areas with unique terrain. In the grid above ecological protection zones marked by DEM data, the system immediately activates the assessment program when it detects a density value exceeding the zero threshold (i.e., flight activity). The system calls the environmental sensitivity layer for that area. If the drone's flight altitude is below the canopy elevation + 50 meters (e.g., in forest areas) or the straight-line distance to a water source is less than 300 meters, an ecological disturbance risk is identified. For areas with sudden weather changes, when heavy rain causes drone activity within grids with visibility below 500 meters, the system determines the equipment damage risk level based on the matching degree between rainfall intensity and the drone's waterproof rating database.

[0072] All risk assessment results are encoded into structured risk data packets, containing elements such as risk type, risk level, risk location, impact range, and duration. A five-level classification system is used for risk levels, with level one being the lowest risk and level five being an emergency risk. For example, in a case of sudden weather change, the system might output: "Risk type: Equipment damage; Risk level: Level four; Location: Grid G-07; Impact range: Radius 800 meters; Duration: Estimated 25 minutes." These risk data packets are pushed to the path planning system in real time via a message queue and simultaneously displayed as dynamic layers on the geographic information platform, forming a visualized airspace risk situation map.

[0073] See Figure 4The system receives structured risk data packets from the risk assessment module. These packets contain key information such as risk type, level, geographical extent, and duration. The system first performs spatial clustering analysis on this risk data, merging geographically adjacent high-risk areas (typically levels four and five) with the same risk type to form contiguous "high-priority adjustment areas." Each area is assigned a unique area identifier, a set of polygon boundary coordinates, and a comprehensive risk weight value. This weight is determined by the level, scope of impact, and duration of each risk point within the region, and is calculated as follows:

[0074] ;

[0075] in: Represents the first in this region The risk level is rated from 1 to 5. The area affected by this risk point (square kilometers). This represents the estimated duration (in minutes) of the risk point. This represents the maximum duration baseline value set by the system (e.g., 180 minutes). Represents the total number of risk points within the region. Weight The higher the value, the greater the urgency and intensity of the adjustments needed in that region.

[0076] After identifying the high-priority adjustment areas, the path replanning module is activated. This module first obtains the current position, speed, heading, and predetermined original flight path point sequence of all affected UAVs. For each UAV path that needs to traverse or lie within the adjustment area, the system uses an improved A* algorithm for local path replanning, constrained by avoiding the entire risk area polygon. The algorithm expands the risk area polygon outward by a certain distance (e.g., 50 meters) to form a "no-fly buffer zone" and searches for available waypoints outside this buffer zone. The goal of the replanning is to find a smooth new path that minimizes the increase in total distance and lies entirely outside the buffer zone, while connecting the original path's entry and exit points. These locally replanned path segments of all affected UAVs together constitute the initial optimized path set.

[0077] The network optimization algorithm performs connectivity verification on the initial set of optimized paths. This algorithm models the airspace as a three-dimensional graph network, where nodes are potential waypoints and edges are feasible routes between nodes. The verification process mainly checks for two types of conflicts: first, conflicts between drones, i.e., whether the optimized paths of different drones have excessively close intersections in space and time; and second, conflicts of airspace resources, i.e., whether a large number of drones will flood into a newly generated air corridor at the same time, causing new congestion. The verification algorithm uses a spatiotemporal cube search method to simulate the travel process of drones on each path and detect whether there are spatiotemporal overlaps. If a conflict is found, it is further analyzed based on the severity of the conflict and... The weights are then used to fine-tune the relevant paths sequentially, for example, by introducing a height-level hierarchical strategy (allowing conflicting drones to fly at different altitudes) or a time-delay strategy (allowing a drone to hover briefly to stagger its flight time). After multiple rounds of iterative verification and adjustment, a set of conflict-free and executable path optimization decisions is finally generated.

[0078] The decision is pushed to the human-computer interaction interface and presented to the operator in a visual format. In the interface, the original path, risk areas, optimized path, and conflict detection results are clearly marked with different colors and legends. Operators can interact with the system using graphical tools, such as manually dragging a waypoint, adjusting the preset flight speed of a drone, or setting a new temporary hovering point. The system parses these interactive commands, extracts the adjustment parameters (such as new latitude and longitude coordinates and speed values), and directly modifies the corresponding path parameters in the path optimization decision.

[0079] The modified path parameters are immediately sent to a real-time simulation environment for verification. This simulation environment constructs a digital twin airspace consistent with the real world, including data on terrain, buildings, and real-time wind fields. The simulation engine runs at a higher temporal resolution (e.g., 10 frames per second), simulating the entire flight process of the UAV according to the modified parameters. During the simulation, multiple indicators are continuously monitored: whether the UAV can strictly follow the new path; whether energy consumption is within the allowable range; whether the minimum distance between the UAV and other UAVs is always higher than the safety threshold; and whether the flight attitude is stable under simulated wind speed and direction conditions. The simulation results generate a verification report, recording the UAV status in detail for each frame. If the report shows that all indicators meet the requirements, the system generates a personalized flight plan based on the finally verified path parameters; if a problem is detected (such as insufficient spacing), the problem point will be fed back to the interactive interface, prompting the operator to make adjustments again until the simulation verification is completely successful. The final plan includes a complete set of flight commands, including latitude and longitude, altitude, arrival time, speed, and turning radius for each waypoint.

[0080] Multi-dimensional analysis tools extract flight trajectory data within a specific geographic area from the drone data base. The extraction area is determined by operators using geofencing tools to draw a polygonal region on a digital map; the system automatically converts this polygon into latitude and longitude boundary conditions. A query filters out all drone trajectory points that entered this area within the last 30 minutes. Each trajectory data point includes a timestamp, drone ID, longitude, latitude, altitude, velocity vector, and associated personalized flight plan number. The data is then sorted chronologically to form the original trajectory dataset.

[0081] The basic statistics calculation engine spatially grids the original dataset, dividing the target area into 500m × 500m grid cells, each assigned a unique grid code. The calculation process consists of two parallel threads: the velocity statistics thread iterates through the instantaneous velocity values ​​of all trajectory points within each grid, removes outliers (such as hovering points with zero velocity), and calculates the arithmetic mean; the density statistics thread counts the peak number of drones appearing in each grid per minute, then divides it by the grid area to obtain the density distribution value. See Table 1 for the basic statistics of the grid cells at a given time.

[0082] Table 1: Basic statistics of flight trajectories in a specific area.

[0083]

[0084] The context integration processing stage deeply correlates basic statistics with personalized flight plans. The system obtains details of the flight plan being executed within each grid using the plan number index, including preset routes, task types (such as inspection, surveying, and logistics), and priority weights. The integration algorithm establishes a three-dimensional correlation model: in the spatial dimension, it matches grid statistics with path segments in the plan; in the task dimension, it adjusts the density values ​​with task weights based on the rule that logistics tasks take precedence over surveying tasks; and in the temporal dimension, it marks whether any plan change events occurred during the statistics collection period. For example, if a logistics drone is detected temporarily detouring to grid G-1026 for path optimization during the statistical period, the system will append a task type tag to the density value of that grid and associate it with the triggering reason for the detour decision (such as avoiding weather risks).

[0085] Data fusion technology was then initiated, accessing historical databases of statistics for the same region and time period (e.g., 10:00-10:30 AM on the same day last week). A sliding window comparison mechanism was used to align the current average velocity sequence with historical velocity sequences along the time axis, calculating the velocity difference rate at corresponding time points. Density data underwent heatmap overlay analysis to generate a difference matrix between the current density distribution and the historical baseline density distribution. The fusion process paid particular attention to the impact of personalized flight plan changes; when a grid's current density increase exceeded 50% of the historical average, all path optimization decision records applied to that grid during that time period were automatically retrieved.

[0086] The dynamic trend report generation module extracts behavioral patterns based on the fusion results, identifying three typical trends: persistent trends (e.g., density increases in grid G-1024 for three consecutive statistical periods), abrupt trends (e.g., a sudden increase in density in grid G-1026 due to temporary route changes), and periodic trends (e.g., logistics drones passing through grid G-1023 at fixed times each day). The report presents three core dimensions in a structured manner: a spatial hotspot distribution map marking areas of abnormal density growth, a speed evolution curve showing speed changes along major routes, and a mission impact matrix quantifying the contribution of different mission types to regional traffic flow. Each trend is associated with specific flight plan adjustment records; for example, when describing the abrupt trend in grid G-1026, it is simultaneously associated with "09:47 Path Optimization Decision ID-PO-8816".

[0087] The synthesis of real-time analysis results employs a multi-layered overlay architecture. The base layer contains updated grid statistics tables; the intermediate layer embeds visualization elements from dynamic trend reports (such as heatmap layers and trend curves); and the decision layer extracts key conclusions from the reports to form text summaries. All elements are integrated into a unified geographic information platform view, allowing operators to view the integrated analysis results for any time period via a time slider. When a user selects to view the details of grid G-1026, the interface simultaneously presents the current density value, the growth rate compared to the same period in history, a list of logistics task numbers that led to the growth, and related route optimization decision records, forming a closed-loop analysis chain.

[0088] The system receives dynamic trend reports from multi-dimensional analysis tools. These reports include pattern recognition results of flight behavior within specific areas, such as areas of continuous density growth, flight segments with abnormal speed fluctuations, and the impact weight of mission type on traffic flow. The system activates its strategy update engine to parse key parameters from the reports: extracting optimal detour path patterns for areas of continuous trends, extracting suggested cruise speed correction values ​​for flight segments with abnormal speeds, and extracting priority adjustment coefficients for high-traffic mission types. These parameters are converted into a storable format for the flight strategy library according to preset field mapping rules. The update operation employs a version control mechanism, creating new version records in the flight strategy table of the database while retaining historical versions for retrospective analysis. Each record includes fields such as strategy number, effective time, applicable area code, path pattern binary object, speed threshold, and mission priority mapping table. After the update is complete, the system broadcasts a strategy library version update notification to all online UAVs.

[0089] When a new flight data packet is uploaded by the UAV's onboard system or ground control station, the input interface of the data base triggers an event listening mechanism. After the data packet is preprocessed to generate standardized flight data, the system automatically detects the time interval between its timestamp and the most recent distribution density calculation. If the interval exceeds a preset threshold (e.g., 30 seconds), or if the data packet contains flight data from a high-priority area (e.g., an emergency response UAV), the distribution density analysis result regeneration process is immediately initiated. This process fully reuses the gridded density calculation method defined in Example 1: calling the latest map data to re-divide the target area into grid cells; spatially associating the location information in the new data with the grid boundaries; counting the number of UAVs in each grid cell in real time; and combining the density sequence of the previous five minutes to calculate the density change rate and generate a dynamically updated distribution density heatmap. The entire process runs in a distributed computing framework, achieving millisecond-level response through task sharding.

[0090] The newly generated density distribution analysis results are then input into the risk assessment module, which loads the updated flight strategy library and applies the risk determination rules from the new strategies to the current density distribution. For example, in areas using new detour patterns, the system raises the density safety threshold for those areas; in flight segments applying speed correction values, the system simultaneously adjusts the relative speed benchmark for collision risk determination. Environmental monitoring data is also involved in the assessment, but the processing workflows for meteorological and topographic data remain independent. The potential risk hazard data package output by the assessment will include an analysis of the impact of the strategy update, such as "Strategy S-V2.1 reduces the collision risk level of grid G-2105".

[0091] The path optimization decision generation module receives updated risk data packets and distribution density analysis results. The decision algorithm first checks the matching degree between the risk area and the new path pattern in the flight strategy library, prioritizing the initialization using pre-stored optimized path templates. For new risk patterns not covered by the strategy library, the system initiates a real-time path planning algorithm, but references task priority parameters from the strategy library as constraints during the planning process. After connectivity verification, the generated initial optimized path is compared with historical optimization decisions in the strategy library. When the similarity exceeds a threshold, historical verification results are directly invoked to accelerate the decision-making process.

[0092] The user interaction process synchronously links to the updated strategy library. When operators manually adjust the flight path, the interface automatically loads the speed limits and airspace rules from the new strategy as operational boundaries. The real-time simulation verification environment calls the aircraft performance model in the strategy library to update simulation parameters, such as adopting a new battery consumption rate calculation formula. The final personalized flight plan will indicate the version number of the strategy library used and reference the updated path pattern or speed rules in the plan description.

[0093] The entire implementation process forms a closed-loop feedback loop: dynamic trends drive strategy library updates, new strategies alter the risk assessment rules for distribution density calculation, optimized decisions generate new flight trajectories, and this in turn affects the next round of trend analysis. Through continuous strategy iteration and density recalculation, the system's path decision-making capabilities evolve with the accumulation of operational data. All data versions and strategy versions are fully recorded, supporting backtracking analysis of the changes in decision-making logic along a timeline.

[0094] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0095] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent analysis method based on a UAV data platform, characterized in that, include: Acquire drone flight data, including location information, speed information, and altitude information, and store the drone flight data in the drone data base; Based on the UAV flight data, the distribution density of the UAV in the target area is calculated in real time to obtain the distribution density analysis results. Using the distribution density analysis results and combined with environmental monitoring data, the potential risks and hazards during the flight of the UAV are assessed. Based on the aforementioned potential risks, a path optimization decision is generated, including: Based on the potential risks, high-priority adjustment areas are identified; the paths of drones within the high-priority adjustment areas are replanned to generate initial optimized paths; and network optimization algorithms are used to perform connectivity verification on the initial optimized paths to obtain path optimization decisions. The specific process for determining high-priority adjustment areas is as follows: each area is assigned a unique area identifier, a set of polygon boundary coordinates, and a comprehensive risk weight value. This weight is determined by the level, scope of impact, and duration of each risk point within the region, and is calculated as follows: ； in: Represents the first in this region The risk level value of each risk point This represents the area affected by the risk point. This represents the expected duration of the risk point. This represents the maximum duration baseline value set by the system. Represents the total number of risk points in the area, with weights. The higher the value, the greater the urgency and intensity of the adjustments needed in that region. In response to user interaction, the path optimization decision is dynamically adjusted to generate a personalized flight plan, including: The system parses user-input interaction commands, including manual parameter adjustments; modifies path parameters in the path optimization decision based on the interaction commands; and verifies the modified path parameters using real-time simulation technology to generate a personalized flight plan. Integrating multi-dimensional analysis tools, it performs statistical analysis on flight trajectory data within a specific area, generating real-time analysis results, including: Flight trajectory data within a specific area is extracted from the drone data base; the average speed and density distribution of the flight trajectory data are calculated to generate basic statistics; and the basic statistics are integrated into the context of the personalized flight plan to obtain real-time analysis results.

2. The intelligent analysis method based on UAV data base according to claim 1, characterized in that, Acquire drone flight data, including position information, speed information, and altitude information, and store the drone flight data in a drone data base, including: Real-time acquisition of drone flight data via sensor networks; The UAV flight data is preprocessed, including noise filtering and data formatting, to obtain standardized flight data; The standardized flight data is uploaded to the UAV data base for centralized storage.

3. The intelligent analysis method based on UAV data base according to claim 2, characterized in that, Based on the UAV flight data, the distribution density of UAVs in the target area is calculated in real time, and the distribution density analysis results are obtained, including: Using the location information in the standardized flight data, the target area is divided into grid cells; Calculate the number of drones in each grid cell and generate an initial distribution density map; By combining time series analysis, the initial distribution density map is dynamically updated to obtain the distribution density analysis results.

4. The intelligent analysis method based on UAV data base according to claim 1, characterized in that, Using the distribution density analysis results, combined with environmental monitoring data, the potential risks and hazards during the flight of the UAV are assessed, including: Acquire environmental monitoring data, including meteorological and topographic data; Statistical models were used to analyze the correlation between the distribution density analysis results and environmental monitoring data to identify abnormal flight patterns. Based on the aforementioned abnormal flight pattern, potential risks and hazards are assessed, including collision risk and environmental impact risk.

5. The intelligent analysis method based on a UAV data base according to claim 4, characterized in that, Contextual integration processing is performed on the basic statistics to obtain real-time analysis results, including: Using data fusion technology, the basic statistics are compared with historical flight data; Generate dynamic trend reports to reflect changes in flight behavior; Integrate the dynamic trend report into the real-time analysis results.

6. The intelligent analysis method based on a UAV data base according to claim 5, characterized in that, After generating the dynamic trend report, it also includes: Based on the dynamic trend report, update the flight strategy library in the UAV data base; In response to new drone flight data, the process of generating the distribution density analysis results is retried.

7. The intelligent analysis method based on a UAV data platform according to claim 6, characterized in that, In response to new UAV flight data, the process of generating the distribution density analysis results is retried, including: When new drone flight data input is detected, repeat the distribution density calculation step; The path optimization decision generation process is optimized using the updated distribution density analysis results.

Images