UAV collision risk warning method and system based on dynamic spatial grid

Abstract

The present invention discloses a method and system for early warning of UAV collision risks based on a dynamic spatial grid, relating to the field of aircraft control technology. The method comprises: obtaining historical operating data of a target area based on preset spatiotemporal collection constraints, extracting a set of spatial and UAV reference indicators; constructing a dynamic spatial grid comprising an analysis grid and a shared grid; using the shared grid as a data collection constraint, traversing the analysis grid to initially screen analysis objects, obtaining a set of associated objects comprising a subject and a guest aircraft; performing a collision assessment based on preset discrimination conditions, and outputting a collision risk early warning result. This method thereby achieves the technical effect of improving early warning accuracy and efficiency, and facilitating efficient and parallel early warning.

Description

Technical Field

[0001] The present invention relates to the field of aircraft control technology, and in particular to a method and system for early warning of unmanned aerial vehicle (UAV) collision risks based on a dynamic spatial grid. Background Art

[0002] With the increasing number of drones and the increasing complexity of their operating areas, efficient and accurate early warning of drone collision risks is becoming increasingly important. Existing drone collision risk warning methods are mostly based on fixed grid modeling, which cannot dynamically adjust grid parameters based on actual operating conditions. This can lead to problems with meshes that are either too coarse or too fine in some areas.

[0003] Specifically, traditional fixed-grid modeling methods primarily rely on fixed grid division rules, failing to fully consider the differences between drone operating areas and between drones. This makes it impossible to achieve finer-grained grid division in high-collision risk areas, hindering accurate analysis and early warning of drone collision risks. Furthermore, the fixed nature of the grid division leads to a significant amount of redundant computation during the overall calculation process, impacting both analysis accuracy and the real-time nature of early warnings. Summary of the Invention

[0004] The present invention provides a UAV collision risk warning method and system based on dynamic spatial grids to solve the technical problem in the existing technology that the fixed grid division accuracy is insufficient, which affects the risk warning effectiveness and accuracy, thereby improving the warning accuracy and efficiency and facilitating the technical effect of efficient parallel warning.

[0005] In a first aspect, the present invention provides a UAV collision risk warning method based on a dynamic spatial grid, wherein the UAV collision risk warning method based on a dynamic spatial grid comprises:

[0006] Based on the preset spatiotemporal acquisition constraints, historical operation data of the target operation area is obtained, and a multidimensional reference indicator set is correspondingly extracted, wherein the multidimensional reference indicator set includes a spatial reference indicator set and a UAV reference indicator set.

[0007] Based on the multi-dimensional reference indicator set, a dynamic space grid is constructed, wherein the dynamic space grid includes an analysis space grid and a sharing space grid.

[0008] Taking the shared space grid as the UAV data collection constraint, the analysis space grid is traversed to perform preliminary screening of analysis objects to obtain a set of associated analysis objects, wherein the set of associated analysis objects includes an analysis subject machine and an analysis object machine.

[0009] The collision risk warning result is outputted by performing collision evaluation and warning on the associated analysis object set according to the preset collision judgment conditions.

[0010] In a feasible implementation, based on preset spatiotemporal collection constraints, historical operating data of the target operating area is obtained, and a multi-dimensional reference indicator set is correspondingly extracted, including:

[0011] The spatial constraint is set according to the area boundary of the target operation area, and the backtracking time length of the target operation area is used as the time constraint to form the spatiotemporal acquisition constraint.

[0012] Using the spatiotemporal acquisition constraint as an index, the regional log library of the target operation area is traversed to extract the corresponding UAV operation logs and collision event logs, and output them as the historical operation data.

[0013] The historical operation data is parsed to obtain a heat reference index, a spatial weight reference index, and a collision rate reference index in a target operation area, and output as the spatial reference index set.

[0014] Traverse the historical operation data to extract a list of drone models, and interact with external data sources based on the drone model list to collect corresponding value reference indicators and task weight reference indicators, and output them as the drone reference indicator set.

[0015] In a feasible implementation, a dynamic space grid is constructed based on the multi-dimensional reference indicator set, wherein the dynamic space grid includes an analysis space grid and a shared space grid, including:

[0016] The target operation area is spatially divided based on the spatial reference index set to obtain the analysis space grid.

[0017] Based on the analysis space grid, the shared space grid is constructed by combining the drone reference index set and the space reference index set.

[0018] A grid mapping relationship is established between the analysis space grid and the shared space grid, and the grid mapping relationship, the analysis space grid, and the shared space grid are correspondingly output as the dynamic space grid.

[0019] In a feasible implementation, performing a spatial division on the target operation area based on the spatial reference index set to obtain the analysis space grid includes:

[0020] A region division factor is calculated for a target operating region based on the heat reference index, the spatial weight reference index, and the collision rate reference index.

[0021] According to the area division factor and the preset standard grid parameters, the grid size distribution of the target operation area is configured, and M grid units with variable grid boundaries are correspondingly formed.

[0022] The M grid units are traversed to be numbered and their boundaries are marked, and the output is the analysis space grid.

[0023] In a feasible implementation, based on the analysis space grid and in combination with the drone reference index set and the spatial reference index set, the shared space grid is constructed, including:

[0024] The M grid centers of the shared space grid are defined with the grid center of each grid cell in the analysis space grid as the center point.

[0025] The UAV reference index and the space reference index are integrated to obtain a shared space radius set, and an initial shared space grid is generated by combining M grid centers, wherein the shared space radius set includes M space expansion radii.

[0026] Analyze and calculate the grid monomer spatial overlap rate of the initial shared space grid, and determine whether the grid monomer spatial overlap rate meets the overlap rate control limit.

[0027] If satisfied, the consistency measure of the grid monomer spatial overlap rate distribution and the heat reference index is calculated, the initial shared space grid is consistency checked, and the space expansion radius is adjusted and expanded according to the consistency check result to iteratively update the initial shared space grid.

[0028] The initial shared space grid that passes the consistency check is output, and a corresponding mapping relationship is established with the analysis space grid.

[0029] In a feasible implementation, the shared space grid is used as a constraint for drone data collection, and the analysis space grid is traversed to perform preliminary screening of analysis objects to obtain a set of related analysis objects, including:

[0030] The boundary of each grid cell in the shared space grid is used as the data collection range, the object machine to be analyzed is determined, and the object machine state data set is correspondingly extracted.

[0031] The UAVs contained in each grid cell in the analysis space grid in real time are taken as analysis subjects, and corresponding subject state data sets are collected.

[0032] The real-time speed algebraic sum is calculated by combining the object machine state data set and the subject machine state data set, and the predicted intersection distance is calculated by combining the warning time window.

[0033] Calculate the distance difference between the predicted intersection distance and the spatial straight-line distance between the analysis object machine and the analysis subject machine, and correspondingly eliminate the analysis object machines whose distance difference does not meet the set safety distance, and output the retained analysis subject machine and the analysis object machine as the associated analysis object set.

[0034] In a feasible implementation, collision evaluation and warning are performed on the associated analysis object set according to preset collision judgment conditions, and a collision risk warning result is output, including:

[0035] The motion state parameters and path planning parameters are updated and collected according to the association analysis object set, and a dynamic intersection determination model is constructed accordingly.

[0036] According to the dynamic intersection determination model, the predicted relative shortest distance of each analysis subject machine-analysis object machine pair is calculated.

[0037] The intrinsic clearance distance of the association analysis object set is obtained, and a collision risk coefficient is calculated according to the intrinsic clearance distance, the predicted relative closest distance and a preset reference reaction time.

[0038] The collision risk warning level is matched according to the collision risk coefficient, and the collision risk warning level and the corresponding analysis subject machine-analysis object machine pair are output as the collision risk warning result.

[0039] In a second aspect, the present invention further provides a UAV collision risk warning system based on a dynamic spatial grid, wherein the UAV collision risk warning system based on a dynamic spatial grid comprises:

[0040] The multidimensional indicator extraction module is used to obtain historical operation data of the target operation area based on preset spatiotemporal acquisition constraints, and correspondingly extract a multidimensional reference indicator set, wherein the multidimensional reference indicator set includes a spatial reference indicator set and a drone reference indicator set.

[0041] The dynamic space grid construction module is used to construct a dynamic space grid based on the multi-dimensional reference indicator set, wherein the dynamic space grid includes an analysis space grid and a shared space grid.

[0042] The analysis object determination module is used to use the shared space grid as the drone data collection constraint, traverse the analysis space grid to perform preliminary screening of analysis objects, and obtain a set of associated analysis objects, wherein the set of associated analysis objects includes an analysis subject machine and an analysis object machine.

[0043] The risk warning and output module is used to perform collision evaluation and warning on the associated analysis object set according to preset collision judgment conditions, and output collision risk warning results.

[0044] The beneficial effects of the present invention are: according to the set spatiotemporal acquisition constraints, the historical operation data of the target operation area is obtained, and a multi-dimensional reference indicator set is extracted based on this, and the indicator set includes spatial dimension reference indicators and UAV dimension reference indicators; based on the multi-dimensional reference indicator set, an adaptive dynamic spatial grid structure is constructed, and the spatial grid is divided into an analysis grid for information analysis and a shared grid for data coordination; using the shared spatial grid as the data acquisition boundary constraint of the UAV, the target object traversal and screening are performed in the analysis grid, and the analysis object set associated with the current task is identified, and the object set includes the analysis subject UAV and the associated object UAV; for the analysis object set, collision risk analysis and warning are performed according to the preset collision judgment rules, and the corresponding collision risk prompt result is output. The UAV collision risk warning method and system based on dynamic spatial grid disclosed by the present invention solves the technical problem of insufficient fixed grid division accuracy and affecting the risk warning effectiveness and accuracy, and achieves the technical effect of improving warning accuracy and warning efficiency, and facilitating efficient parallel warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 The figure is a flow chart of the UAV collision risk warning method based on dynamic spatial grid of the present invention.

[0046] Figure 2 This is a structural diagram of the UAV collision risk warning system based on dynamic spatial grids of the present invention.

[0047] In the accompanying drawings, the components represented by the reference numerals are described as follows:

[0048] Multi-dimensional indicator extraction module 11, dynamic space grid construction module 12, analysis object determination module 13, risk warning and output module 14. DETAILED DESCRIPTION

[0049] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention, rather than all, are shown in the drawings.

[0050] Example 1, as Figure 1 The figure is a flow chart of a UAV collision risk warning method based on a dynamic spatial grid according to the present invention, wherein the UAV collision risk warning method based on a dynamic spatial grid comprises:

[0051] S100: Based on preset spatiotemporal acquisition constraints, historical operation data of a target operation area is obtained, and a multidimensional reference indicator set is correspondingly extracted, wherein the multidimensional reference indicator set includes a spatial reference indicator set and a UAV reference indicator set.

[0052] Specifically, the target operating area refers to the geographic space that the drone is currently in or about to enter during a current or pre-determined mission. Spatiotemporal collection constraints are a set of conditions used to limit the collection and screening of drone flight data to specific time and spatial ranges, ensuring the relevance and representativeness of the data.

[0053] Specifically, the multidimensional reference indicator set is a multidimensional feature set extracted from historical operation data that meets the spatiotemporal acquisition constraints and is used as the basis for grid division. The multidimensional reference indicator set includes a spatial reference indicator set and a UAV reference indicator set. The spatial reference indicator set is used to describe the spatial structure characteristics within the operation area, such as grid density, obstacle distribution and airspace congestion, and the UAV reference indicator set is used to describe the behavioral characteristics of historical UAVs, such as the UAV model, performance parameters, mission type and other information.

[0054] Through the above process, the historical operation status of UAVs in the target operation area can be fully grasped, providing a solid data foundation and reference basis for subsequent dynamic spatial grid construction and collision risk warning.

[0055] In some embodiments, based on preset spatiotemporal collection constraints, historical operating data of the target operating area is obtained, and a multi-dimensional reference indicator set is correspondingly extracted, including:

[0056] According to the regional boundary of the target operation area, spatial constraints are set, and the backtracking time of the target operation area is used as the time constraint to form the spatiotemporal acquisition constraints; with the spatiotemporal acquisition constraints as the index, the regional log library of the target operation area is traversed to extract the corresponding drone operation logs and collision event logs, and output them as the historical operation data; the historical operation data is parsed to obtain the heat reference index, spatial weight reference index and collision rate reference index in the target operation area, and output them as the spatial reference index set; the historical operation data is traversed to extract the drone model list, and according to the drone model list, the corresponding value reference index and task weight reference index are collected through interaction with external data sources to output them as the drone reference index set.

[0057] Specifically, spatial constraints are set according to the boundaries of the target operating area to ensure that the collected data is relevant to the current task area; temporal constraints are determined based on the preset backtracking time length to limit the time window of historical data to ensure the timeliness of the data.

[0058] Specifically, historical operational data refers to raw operational information extracted from regional log repositories while meeting the aforementioned spatiotemporal collection constraints, including drone operation logs and collision event logs. The multidimensional reference indicator set, generated by analyzing historical operational data, is a structured set of indicators used to support subsequent risk assessment and model building. Specifically, it includes a spatial reference indicator set and a drone reference indicator set.

[0059] Among them, the spatial reference indicator set is used to characterize the spatial operation characteristics of the target area, such as heat distribution, heavy flight areas and historical collision density; the drone reference indicator set is used to characterize the performance and mission behavior characteristics of typical drones in the area, such as value level, mission frequency and flight risk level.

[0060] Specifically, we first set spatial constraints based on the target operating area's boundaries, while also using the target operating area's backtracking time as a temporal constraint. This creates spatiotemporal collection constraints, ensuring the relevance of the acquired data to the target operating area. Next, we traverse the target operating area's regional log library using the spatiotemporal collection constraints as an index, extracting drone operation logs and collision event logs. These logs record detailed information about the drone's historical operations, including key data such as flight trajectory, speed, altitude, and the time and location of collision events, forming a complete historical operation dataset.

[0061] Furthermore, historical operational data is analyzed to calculate reference indicators for popularity (such as the average number of drone appearances per unit time), spatial density (such as the density distribution of drones in different airspace locations), and collision rate (such as the collision rate per thousand flights). These indicators reflect the statistical characteristics of drone activity density, spatial distribution, and collision risk within the target operational area, ultimately forming a spatial reference indicator set. Simultaneously, a list of drone models is compiled from historical operational data. Combined with manufacturer-provided technical parameters and value assessment data from industry databases, value reference indicators (such as purchase cost and maintenance expenses) and mission density (such as mission priority and payload capacity) are collected for each drone model to form a drone reference indicator set.

[0062] Through the above steps, the historical operation data of the target operation area can be comprehensively and systematically collected and organized, and then a set of multi-dimensional reference indicators that are critical to the dynamic space grid division can be extracted, providing detailed data support and technical basis for the subsequent dynamic space grid construction and collision risk assessment.

[0063] S200: Constructing a dynamic space grid based on the multi-dimensional reference indicator set, wherein the dynamic space grid includes an analysis space grid and a shared space grid.

[0064] Specifically, the dynamic spatial grid is a grid structure system with spatiotemporal adaptability and functional layering characteristics constructed based on a multidimensional reference indicator set, which includes two substructures: the analytical spatial grid and the shared spatial grid.

[0065] Among them, the analysis space grid is used to divide multiple drones that need collision risk warning into relatively balanced subsets according to the characteristics of their local space, thereby realizing the decomposition and averaging of the collision risk warning task, facilitating subsequent processing and parallel disposal, and helping to improve modeling accuracy and resource utilization efficiency; the shared space grid is used to constrain the scope of real-time information sharing among multiple drones to ensure efficient and low-redundancy risk perception and collaborative warning.

[0066] Through the above process, a dynamic spatial grid can be constructed based on the characteristics and risk levels of drone activities in different areas, which can not only reflect the spatial distribution but also adapt to the dynamic changes of drones, providing a flexible and accurate spatial analysis framework for real-time monitoring of drones and collision risk warning.

[0067] In some embodiments, a dynamic spatial grid is constructed based on the multidimensional reference indicator set, wherein the dynamic spatial grid includes an analysis spatial grid and a shared spatial grid, including:

[0068] The target operation area is spatially divided based on the spatial reference index set to obtain the analysis space grid; based on the analysis space grid, the shared space grid is constructed in combination with the UAV reference index set and the spatial reference index set; a grid mapping relationship between the analysis space grid and the shared space grid is established, and the grid mapping relationship, the analysis space grid and the shared space grid are correspondingly output as the dynamic space grid.

[0069] Specifically, based on spatial reference indicators extracted from historical operational data (such as regional heat distribution and historical conflict frequency), the target operational area is spatially divided using methods such as density clustering, adaptive grid partitioning, or information entropy optimization. This yields an analytical spatial grid with differentiated spatial characteristics to support subsequent risk modeling and structural analysis. For example, refining the grid granularity in the urban core airspace reduces the number of drones requiring analysis within a single grid, thereby improving forecast response efficiency and facilitating parallel forecasting.

[0070] Specifically, based on the analysis of the spatial grid, a shared spatial grid is generated by combining the drone reference indicator set with the spatial reference indicator set. This shared spatial grid aims to provide a more flexible data collection framework for drones. For example, for drones performing high-priority missions, the shared spatial grid can appropriately expand its grid size (i.e., spatial expansion radius) to provide a wider sensing and monitoring range. Preferably, multiple grid cells in this shared spatial grid can overlap, meaning that any drone can be simultaneously sensed and monitored by multiple other drones.

[0071] Furthermore, a grid mapping relationship is established between the analysis space grid and the shared space grid, ensuring that each grid cell in the analysis space grid accurately corresponds to the corresponding area in the shared space grid, facilitating data connection. Ultimately, the grid mapping relationship, analysis space grid, and shared space grid can be integrated and output to form a complete dynamic space grid system.

[0072] The dynamic spatial grid constructed through the above process can adapt to the distribution changes and mission requirements of drones in real time, providing a flexible and accurate spatial analysis tool for efficient monitoring and collision risk warning of drones, which helps to improve the intelligence level and safety of airspace management.

[0073] In some implementations, performing a spatial division of the target operating area based on the spatial reference index set to obtain the analysis space grid includes:

[0074] Based on the heat reference index, the spatial weight reference index and the collision rate reference index, a regional division factor is calculated for the target operating area; based on the regional division factor and the preset standard grid parameters, the grid size distribution of the target operating area is configured, and M grid monomers with variable grid boundaries are formed accordingly; the M grid monomers are traversed, numbered and boundary-marked, and output as the analysis space grid.

[0075] Specifically, the analysis space grid is a grid collection used to limit the spatial range of a single collision risk analysis. This grid provides a basic spatial framework for subsequent collision risk analysis by dividing the target operating area into multiple grid cells with variable boundaries. The regional division factor is a numerical value calculated based on the heat reference index, spatial weight reference index and collision rate reference index, and is used to quantify the risk distribution within the target operating area. According to the regional division factor and the preset standard grid parameters, the size of each grid cell can be quantitatively determined, thereby achieving differentiated treatment of risk areas with different characteristics and refining the high-risk areas, thereby improving the overall analysis efficiency while ensuring the accuracy of risk identification.

[0076] Specifically, first, the regional division factor of the target operation area is calculated based on the heat reference index, spatial weight reference index, and collision rate reference index. For example, a weighted superposition model or a normalized scoring mechanism is used to comprehensively score each regional unit to form a regional division factor to reflect the risk sensitivity and mission importance of each area. Then, based on the regional division factor and the preset standard grid parameters, the grid size distribution of the target operation area is configured, and M grid cells with variable grid boundaries are formed. Among them, smaller grids are configured in high-factor areas to improve local analysis accuracy and reduce the amount of single grid analysis. In low-factor areas, larger grids can be configured to improve computational efficiency. Finally, the M grid cells are traversed, a unique identifier is assigned to each grid cell, and its spatial boundary coordinates are recorded, and the analysis space grid with a complete structure is output.

[0077] Through the above process, it is possible to achieve refined modeling of high-risk areas and simplified processing of low-risk areas while ensuring the integrity of airspace coverage, thereby improving the response efficiency and computing resource utilization of collision risk analysis. Especially in high-density flight environments, by refining the grid in high-heat and high-collision rate areas, the number of objects required to be processed in each analysis can be significantly reduced, and the system's parallel computing capabilities and real-time dynamic response capabilities can be enhanced, thereby providing high-quality spatial basic support for subsequent shared space grid construction and risk warning mechanisms.

[0078] In some implementations, constructing the shared spatial grid based on the analysis spatial grid and combining the drone reference indicator set with the spatial reference indicator set includes:

[0079] Taking the grid center of each grid monomer in the analysis space grid as the center point, define M grid centers of the shared space grid; fuse the UAV reference index and the space reference index to obtain a shared space radius set, and generate an initial shared space grid in combination with the M grid centers, wherein the shared space radius set includes M space expansion radii; analyze and calculate the grid monomer spatial overlap rate of the initial shared space grid, and determine whether the grid monomer spatial overlap rate meets the overlap rate control limit; if so, calculate the consistency measure of the grid monomer spatial overlap rate distribution and the heat reference index, perform a consistency check on the initial shared space grid, and adjust and expand the space expansion radius according to the consistency check result to iteratively update the initial shared space grid; output the initial shared space grid that passes the consistency check, and establish a corresponding mapping relationship with the analysis space grid.

[0080] Specifically, the shared spatial grid is a dynamic grid structure constructed through spatial expansion based on spatial grid analysis, combined with real-time operational status and airspace risk characteristics. It supports data sharing, collaborative perception, and early warning linkage among multiple UAV systems. In other words, the shared spatial grid can form an extended spatial region with overlapping areas by setting the grid center and spatial expansion radius, thereby achieving cross-grid information fusion and collaborative decision-making. The generation of this shared spatial grid relies on the fusion calculation of the UAV reference index set and the spatial reference index set.

[0081] Specifically, first, the grid center of each grid unit in the analysis space grid is used as the center point to define M grid centers of the shared space grid, ensuring that the shared space grid maintains a one-to-one correspondence with the analysis space grid in spatial structure; then, the drone reference index and the spatial reference index are fused, and a fusion model such as weighted fusion is used to calculate the spatial expansion radius of each grid unit to form a shared space radius set, and the initial shared space grid is generated with the grid center and the corresponding radius as parameters; among them, each spatial expansion radius determines the coverage range of the corresponding shared space grid, and the spatial size of the initial grid is usually several multiples of the corresponding analysis space grid size.

[0082] Specifically, the spatial overlap rate between adjacent grid cells in the initial shared space grid is then analyzed and calculated, and it is determined whether the preset overlap rate control limit (such as a preset interval) is met. If so, it means that the current initial shared space grid division is relatively moderate (preliminary qualified), and the consistency measure between the spatial overlap rate distribution of the grid cells and the spatial distribution of the heat reference index can be further calculated. For example, indicators such as the Pearson correlation coefficient, KL divergence, or spatial distribution similarity can be used for consistency verification. The purpose of this consistency verification is to ensure that the areas with high overlap rates of the initial shared space grid correspond to the areas with high heat, thereby ensuring the rationality and adaptability of the obtained shared space grid.

[0083] Specifically, if the consistency check fails, the spatial expansion radius is adjusted based on the consistency check result, and the shared spatial grid structure is iteratively updated and iterative consistency checks are performed until the consistency requirements are met. At this point, the initial shared spatial grid that passes the consistency check can be output and a mapping relationship can be established between it and the analysis spatial grid, enabling the linkage of the two types of grid structures.

[0084] The above process constructs a shared space grid system corresponding to the analysis space grid structure, which has spatial overlap and dynamic adaptability. It effectively expands the perception boundary of a single analysis grid and enhances the collaborative early warning capability of multiple UAV systems in high-risk areas, especially in areas with intensive tasks and frequent path intersections. By increasing the expansion radius and overlap rate of the shared space grid, the information fusion and risk identification accuracy of the local area can be significantly improved. At the same time, the consistency verification mechanism is used to ensure that the structural distribution of the shared space grid matches the airspace heat characteristics, thereby optimizing the overall resource scheduling efficiency and risk response capability of the system while ensuring the continuity of spatial coverage.

[0085] S300: Using the shared space grid as a constraint for drone data collection, traverse the analysis space grid to perform preliminary screening of analysis objects, and obtain a set of associated analysis objects, wherein the set of associated analysis objects includes an analysis subject machine and an analysis object machine.

[0086] Specifically, initial analysis object screening involves the preliminary selection of drones within the analysis grid, within the constraints of the shared spatial grid. This process is used to initially determine which drones are at high risk of collision. The associated analysis object set is the set of drones identified after initial screening for further collision risk analysis, including both subject and object drones. Subject drones are drones within the current analysis grid, while object drones are other drones within the shared spatial grid that may interact with the subject drone.

[0087] By initially screening the analysis objects based on the shared spatial grid as a constraint, the scope of the collision risk analysis can be quickly narrowed down, focusing the analysis objects on those drone pairs that may actually pose a collision risk. This avoids unnecessary data collection and calculation, reduces the waste of system resources, and thus improves the efficiency and accuracy of collision risk warnings.

[0088] In some embodiments, the shared space grid is used as a constraint for drone data collection, and the analysis space grid is traversed to perform preliminary screening of analysis objects to obtain a set of associated analysis objects, including:

[0089] The boundary of each grid cell in the shared space grid is used as the data collection range to determine the object machine for analysis and extract the object machine status data set accordingly; the drones contained in each grid cell in the analysis space grid in real time are used as the analysis subject machines, and the subject machine status data set is collected accordingly; the real-time speed algebraic sum is calculated by combining the object machine status data set and the subject machine status data set, and the predicted intersection distance is calculated in combination with the warning time window; the distance difference between the predicted intersection distance and the spatial straight-line distance between the analysis object machine and the analysis subject machine is calculated, and the analysis object machines whose distance difference does not meet the set safety distance are correspondingly eliminated, and the retained analysis subject machines and analysis object machines are output as the associated analysis object set.

[0090] Specifically, for each cell in the analysis space grid, the corresponding cell in the shared space grid is called, and the data collection range is set to the boundaries of the cells in the shared space grid, limiting the drone objects involved in the analysis to be within this spatial range. Next, the drone currently contained in each cell in the analysis space grid is identified as the analysis subject, and its current state data set is collected, including but not limited to position, velocity vector, heading angle, mission type, etc. Simultaneously, other drones within the corresponding shared space grid boundary are identified as analysis objects, and their state data sets are simultaneously extracted.

[0091] Specifically, then, traverse each pair of analysis subject machine and analysis object machine combination, calculate the real-time speed algebraic sum of the combination based on the corresponding state data set, and calculate the predicted intersection distance in combination with the warning time window; at the same time, calculate the spatial straight-line distance between the analysis subject machine and the analysis object machine based on the corresponding state data set; then calculate the distance difference between the predicted intersection distance and the above-mentioned spatial straight-line distance, and use the distance difference as a judgment basis to compare with the set safety distance. If the distance difference is the smaller one, it is considered that the analysis object machine may have a spatial intersection with the subject machine within the warning window, and the object machine is retained; otherwise, it is eliminated.

[0092] Furthermore, for each grid cell of the analysis space grid, the retained analysis subject machine and its corresponding analysis object machine set are output as the associated analysis object set under the grid for subsequent collision risk modeling and warning calculation.

[0093] This method allows for rapid, preliminary screening of potential collision targets based on the drones' current motion status and path characteristics within the confines of a shared spatial grid, improving computational efficiency and accuracy. For example, in a complex drone operating environment, there may be a large number of drones, but not all of them pose a collision risk. This process allows for rapid screening of drone pairs of concern, thereby improving the real-time and reliability of early warnings.

[0094] S400: performing collision evaluation and early warning on the associated analysis object set according to preset collision judgment conditions, and outputting a collision risk early warning result.

[0095] Specifically, collision judgment criteria refer to the preset standards used to determine whether there is a collision risk between drones, typically including factors such as safe distance, relative speed, and flight path. Collision assessment is the process of assessing the risk of drone pairs in the associated analysis set based on the collision judgment criteria.

[0096] Specifically, the collision risk warning result is the final conclusion on the UAV collision risk based on the collision evaluation, which usually includes the collision risk level, warning information and related UAV pair information.

[0097] For example, the system first evaluates the collision risk of each pair of drones in the associated analysis set based on the pre-defined collision criteria. The predicted relative closest distance between each pair of drones is calculated, and a collision risk coefficient is calculated based on the aforementioned collision criteria. Finally, the collision risk coefficient is matched to a corresponding collision risk warning level, and the collision risk warning result is output.

[0098] Through precise collision evaluation, the above process can ultimately determine which drone pairs are at risk of collision and issue early warning information in a timely manner so that appropriate risk avoidance measures can be taken.

[0099] In some embodiments, performing collision evaluation and warning on the associated analysis object set according to preset collision judgment conditions and outputting collision risk warning results includes:

[0100] According to the associated analysis object set, the motion state parameters and path planning parameters are updated and collected, and a dynamic intersection judgment model is constructed accordingly; according to the dynamic intersection judgment model, the predicted relative closest distance of each analysis subject machine-analysis object machine pair is calculated; the intrinsic clearance distance of the associated analysis object set is obtained, and the collision risk coefficient is calculated based on the intrinsic clearance distance, the predicted relative closest distance and the preset reference reaction time; the collision risk warning level is matched according to the collision risk coefficient, and the collision risk warning level and the corresponding analysis subject machine-analysis object machine pair are output as the collision risk warning result.

[0101] Specifically, motion state parameters refer to the UAV's real-time flight status data, including speed, acceleration, position coordinates, and flight direction. Path planning parameters refer to the UAV's flight path information, including starting point, end point, altitude, and flight path. The dynamic intersection determination model is a mathematical model based on the UAV's motion state and path planning information, used to predict the relative motion and intersection between UAVs.

[0102] Specifically, the predicted relative closest distance refers to the minimum distance that can be reached between the subject and the object being analyzed within a preset time window. The intrinsic clearance distance refers to the minimum safe distance that must be maintained between drones to avoid collision, determined by the size and flight characteristics of the drones. The collision risk coefficient is a quantitative indicator used to assess the degree of collision risk between drones. It is calculated by combining the predicted relative closest distance, intrinsic clearance distance, and reference reaction time. The collision risk warning level is a risk level based on the collision risk coefficient, which is used to concisely and efficiently indicate the severity of the collision risk.

[0103] Specifically, for each subject and object pair, the system first updates and collects the corresponding motion state parameters (including position, velocity, acceleration, and heading angle) and path planning parameters (including flight path, mission target, and obstacle avoidance strategy) in real time. Based on these parameters, a dynamic convergence determination model is constructed to predict spatial convergence trends. Then, based on this constructed dynamic convergence determination model, the system calculates the relative minimum distance between each subject and object pair within the prediction time window—that is, the distance between the closest points in the two aircraft's trajectories at a given moment in time.

[0104] At the same time, for each subject-object pair, the corresponding intrinsic clearance distance is obtained. This distance is dynamically set based on factors such as the flight type, platform size, control accuracy, and mission priority of the two aircraft, representing the minimum safe separation between the two aircraft without external interference. For example, the larger the size and weight of the UAV platform and the lower its flight stability, the greater the corresponding intrinsic clearance distance. This intrinsic clearance distance can be obtained through interaction with the manufacturer.

[0105] Furthermore, by combining the predicted relative closest distance, the two intrinsic clearance distances of the subject machine and the object machine pair, and the preset reference reaction time, the collision risk coefficient of each object pair can be calculated. For example, the collision risk coefficient can be calculated using the following calculation model:

[0106] ;

[0107] in, is the collision risk factor, is the intrinsic clearance distance (such as the sum of two intrinsic clearance distances), Predict the relative closest distance, is the relative speed modulus of the two machines, For the preset reference reaction time, A small value to prevent division by zero. A larger risk factor indicates a higher potential collision risk.

[0108] Furthermore, based on the calculated collision risk coefficient, the system traverses a preset collision risk warning level mapping table to match the warning level. Each analysis subject machine-analysis object machine pair is then classified into a corresponding risk level (e.g., high risk, medium risk, low risk, or safe). The system then outputs a collision risk warning result containing the following: the analysis subject machine identifier, the corresponding analysis object machine identifier, the collision risk warning level, an optional warning time, a warning location, or an avoidance suggestion. The collision risk warning level mapping table contains the correspondence between collision risk coefficients and multiple risk levels, such as the range of collision risk coefficient values ​​corresponding to different risk levels.

[0109] Through the above-mentioned collision assessment and risk warning method, dynamic modeling and hierarchical warning of potential collision relationships in the drone group operation environment are realized. Among them, the risk coefficient and level mapping mechanism enables the warning results to have a clear response level, which facilitates the subsequent command and dispatch and conflict avoidance strategy execution.

[0110] In summary, the UAV collision risk warning method based on dynamic spatial grid provided by the present invention has the following technical effects:

[0111] Through the set spatiotemporal collection constraints, the historical operation data of the target operation area is obtained, and a multi-dimensional reference indicator set is extracted based on this, which includes spatial dimension reference indicators and UAV dimension reference indicators; based on the multi-dimensional reference indicator set, an adaptive dynamic spatial grid structure is constructed, and the spatial grid is divided into an analysis grid for information analysis and a shared grid for data coordination; using the shared spatial grid as the data collection boundary constraint of the UAV, the target objects are traversed and screened in the analysis grid, and the analysis object set associated with the current task is identified, and the object set includes the analysis subject UAV and the associated object UAV; for the analysis object set, collision risk analysis and warning are performed according to the preset collision judgment rules, and the corresponding collision risk prompt results are output, thereby improving the warning accuracy and warning efficiency, and facilitating the technical effect of efficient parallel warning.

[0112] Example 2, as Figure 2 This is a schematic diagram of the structure of the UAV collision risk warning system based on dynamic spatial grids of the present invention. For example, Figure 1 The flowchart of the UAV collision risk warning method based on dynamic spatial grid in the present invention can be shown as follows: Figure 2 The structure shown is implemented.

[0113] Based on the same concept as the UAV collision risk warning method based on dynamic spatial grid in the above embodiment, the present invention also provides a UAV collision risk warning system based on dynamic spatial grid, which includes:

[0114] The multidimensional indicator extraction module 11 is used to obtain historical operation data of the target operation area based on preset spatiotemporal acquisition constraints, and correspondingly extract a multidimensional reference indicator set, wherein the multidimensional reference indicator set includes a spatial reference indicator set and a drone reference indicator set.

[0115] The dynamic space grid construction module 12 is configured to construct a dynamic space grid based on the multi-dimensional reference indicator set, wherein the dynamic space grid includes an analysis space grid and a shared space grid.

[0116] The analysis object determination module 13 is used to use the shared space grid as the drone data collection constraint, traverse the analysis space grid to perform preliminary screening of analysis objects, and obtain a set of associated analysis objects, wherein the set of associated analysis objects includes an analysis subject machine and an analysis object machine.

[0117] The risk warning and output module 14 is used to perform collision evaluation and warning on the associated analysis object set according to preset collision judgment conditions, and output a collision risk warning result.

[0118] In some embodiments, the multi-dimensional indicator extraction module 11 includes:

[0119] The spatiotemporal acquisition constraint setting unit is configured to set spatial constraints according to the area boundary of the target operation area, and to use the backtracking time of the target operation area as a time constraint to form the spatiotemporal acquisition constraints.

[0120] The historical operation data extraction unit is used to traverse the regional log library of the target operation area with the said spatiotemporal acquisition constraint as the index, extract the corresponding drone operation log and collision event log, and output them as the said historical operation data.

[0121] The spatial reference indicator set generation unit is used to parse the historical operation data, obtain the heat reference index, spatial weight reference index and collision rate reference index in the target operation area, and output them as the spatial reference indicator set.

[0122] The drone reference indicator set generation unit is used to traverse the historical operation data to extract a list of drone models, and interact with external data sources to collect corresponding value reference indicators and task weight reference indicators based on the drone model list, and output them as the drone reference indicator set.

[0123] In some embodiments, the dynamic spatial grid construction module 12 includes:

[0124] The analysis space grid acquisition unit is used to perform a spatial division on the target operation area based on the spatial reference index set to obtain the analysis space grid.

[0125] A shared space grid construction unit is used to construct the shared space grid based on the analysis space grid and in combination with the drone reference indicator set and the space reference indicator set.

[0126] The grid mapping relationship establishing and outputting unit is used to establish a grid mapping relationship between the analysis space grid and the shared space grid, and correspondingly output the grid mapping relationship, the analysis space grid and the shared space grid as the dynamic space grid.

[0127] In some implementations, the analysis space grid acquisition unit in the dynamic space grid construction module 12 includes:

[0128] The area division factor calculation unit is used to calculate the area division factor of the target operation area according to the heat reference index, the spatial weight reference index and the collision rate reference index.

[0129] The grid size distribution configuration unit is used to configure the grid size distribution of the target operation area according to the area division factor and the preset standard grid parameters, and correspondingly form M grid units with variable grid boundaries.

[0130] The analysis space grid output unit is used to traverse the M grid units to perform numbering and boundary marking, and output the analysis space grid.

[0131] In some implementations, the shared spatial grid construction unit in the dynamic spatial grid construction module 12 includes:

[0132] The shared space grid center definition unit is used to define M grid centers of the shared space grid with the grid center of each grid cell in the analysis space grid as the center point.

[0133] An initial shared space grid generation unit is used to fuse the UAV reference index with the space reference index, obtain a shared space radius set, and generate an initial shared space grid in combination with M grid centers, wherein the shared space radius set includes M space expansion radii.

[0134] The grid monomer space overlap rate analysis and judgment unit is used to analyze and calculate the grid monomer space overlap rate of the initial shared space grid, and judge whether the grid monomer space overlap rate meets the overlap rate control limit.

[0135] The consistency check and shared space grid iterative update unit is used to calculate the consistency measurement of the grid monomer spatial overlap rate distribution and the heat reference index if it is satisfied, perform consistency check on the initial shared space grid, and adjust and expand the space expansion radius according to the consistency check result to iteratively update the initial shared space grid.

[0136] The shared space grid output and mapping relationship establishment unit is used to output the initial shared space grid that has passed the consistency check and establish a corresponding mapping relationship with the analysis space grid.

[0137] In some embodiments, the analysis object determination module 13 includes:

[0138] The object machine state data set extraction unit is used to determine the object machine for analysis and extract the object machine state data set accordingly, using the boundary of each grid cell in the shared space grid as the data collection range.

[0139] The subject machine state data set collection unit is used to take the drone contained in each grid cell in the analysis space grid in real time as the analysis subject machine and collect the subject machine state data set accordingly.

[0140] The predicted intersection distance calculation unit is used to calculate the real-time speed algebraic sum in combination with the object machine state data set and the subject machine state data set, and calculate the predicted intersection distance in combination with the warning time window.

[0141] The associated analysis object set output unit is used to calculate the distance difference between the predicted intersection distance and the spatial straight-line distance between the analysis object machine and the analysis subject machine, and to eliminate the analysis object machines whose distance difference does not meet the set safety distance, and output the retained analysis subject machine and the analysis object machine as the associated analysis object set.

[0142] In some embodiments, the risk warning and output module 14 includes:

[0143] The dynamic intersection determination model construction unit is used to update the collected motion state parameters and path planning parameters according to the association analysis object set, and to construct a dynamic intersection determination model accordingly.

[0144] The predicted relative shortest distance calculation unit is used to calculate the predicted relative shortest distance of each analysis subject machine-analysis object machine pair according to the dynamic intersection judgment model.

[0145] The collision risk coefficient calculation unit is used to obtain the intrinsic clearance distance of the association analysis object set and calculate the collision risk coefficient according to the intrinsic clearance distance, the predicted relative closest distance and a preset reference reaction time.

[0146] The collision risk warning result output unit is used to match the collision risk warning level according to the collision risk coefficient, and output the collision risk warning level and the corresponding analysis subject machine-analysis object machine pair as the collision risk warning result.

[0147] It should be understood that the embodiments mentioned in this specification focus on their differences from other embodiments. The specific embodiments in the aforementioned embodiment one are also applicable to the drone collision risk warning system based on dynamic spatial grid described in embodiment two. For the sake of brevity of the specification, no further elaboration will be given here.

[0148] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the embodiments mentioned above. It should be understood that those skilled in the art can still modify the technical solutions described in the above embodiments or replace some of the technical features therein with equivalents; and such modifications or replacements do not deviate from the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention and are all included in the scope of protection of the present invention.

Claims

1. A UAV collision risk warning method based on dynamic spatial grids, characterized by: include: Based on the preset spatiotemporal collection constraints, historical operation data of the target operation area is obtained, and a multi-dimensional reference indicator set is correspondingly extracted, wherein the multi-dimensional reference indicator set includes a spatial reference indicator set and a UAV reference indicator set; Based on the multi-dimensional reference indicator set, constructing a dynamic space grid, wherein the dynamic space grid includes an analysis space grid and a shared space grid; Using the shared space grid as a constraint for drone data collection, traversing the analysis space grid to perform preliminary screening of analysis objects, and obtaining a set of associated analysis objects, wherein the set of associated analysis objects includes an analysis subject machine and an analysis object machine; Performing collision evaluation and early warning on the associated analysis object set according to preset collision judgment conditions, and outputting collision risk early warning results; Among them, building a dynamic spatial grid includes: Performing a spatial division on the target operation area based on the spatial reference index set to obtain the analysis space grid; Based on the analysis space grid, the shared space grid is constructed by combining the drone reference indicator set and the spatial reference indicator set; Establishing a grid mapping relationship between the analysis space grid and the shared space grid, and correspondingly outputting the grid mapping relationship, the analysis space grid, and the shared space grid as the dynamic space grid; Wherein, obtaining the analysis space grid includes: Calculate the regional division factor of the target operation area based on the heat reference index, space weight reference index and collision rate reference index; According to the region division factor and the preset standard grid parameters, a grid size distribution of the target operation area is configured, and M grid units with variable grid boundaries are correspondingly formed; Traversing M grid units, numbering and marking boundaries, and outputting the analysis space grid; Wherein, constructing the shared space grid includes: Defining M grid centers of the shared space grid with the grid center of each grid cell in the analysis space grid as the center point; Fusing the UAV reference index with the spatial reference index to obtain a shared space radius set, and generating an initial shared space grid by combining M grid centers, wherein the shared space radius set includes M space expansion radii; Analyzing and calculating the grid monomer spatial overlap rate of the initial shared space grid, and determining whether the grid monomer spatial overlap rate meets the overlap rate control limit; If satisfied, then calculate the consistency measure of the grid monomer spatial overlap rate distribution and the heat reference index, perform consistency check on the initial shared space grid, and adjust and expand the space expansion radius according to the consistency check result to iteratively update the initial shared space grid; The initial shared space grid that passes the consistency check is output, and a corresponding mapping relationship is established with the analysis space grid.

2. The UAV collision risk warning method based on dynamic spatial grid according to claim 1, characterized in that: Based on the preset spatiotemporal collection constraints, historical operating data of the target operating area is obtained, and a multi-dimensional reference indicator set is extracted accordingly, including: Setting spatial constraints according to the regional boundaries of the target operating area, and using the backtracking time of the target operating area as a time constraint to form the spatiotemporal acquisition constraints; Using the spatiotemporal acquisition constraints as indexes, traverse the regional log library of the target operation area, extract corresponding drone operation logs and collision event logs, and output them as the historical operation data; Parsing the historical operation data, obtaining a heat reference index, a spatial weight reference index, and a collision rate reference index within a target operation area, and outputting the result as the spatial reference index set; Traverse the historical operation data to extract a list of drone models, and interact with external data sources based on the drone model list to collect corresponding value reference indicators and task weight reference indicators, and output them as the drone reference indicator set.

3. The UAV collision risk warning method based on dynamic spatial grid according to claim 2, characterized in that: Using the shared space grid as the drone data collection constraint, traverse the analysis space grid to perform preliminary screening of analysis objects and obtain a set of related analysis objects, including: Taking the boundary of each grid cell in the shared space grid as the data collection range, determining the object machine to be analyzed and extracting the object machine state data set accordingly; The UAVs contained in each grid cell in the analysis space grid in real time are used as analysis subjects, and corresponding subject state data sets are collected; Calculating a real-time velocity algebraic sum by combining the object machine state data set and the subject machine state data set, and calculating a predicted intersection distance by combining the warning time window; Calculate the distance difference between the predicted intersection distance and the spatial straight-line distance between the analysis object machine and the analysis subject machine, and correspondingly eliminate the analysis object machines whose distance difference does not meet the set safety distance, and output the retained analysis subject machine and the analysis object machine as the associated analysis object set.

4. The UAV collision risk warning method based on dynamic spatial grid according to claim 3, characterized in that: Perform collision evaluation and early warning on the associated analysis object set according to preset collision judgment conditions, and output collision risk early warning results, including: Update and collect motion state parameters and path planning parameters according to the association analysis object set, and construct a dynamic intersection determination model accordingly; Calculating the predicted relative shortest distance between each analysis subject machine and analysis object machine pair according to the dynamic intersection determination model; Obtaining the intrinsic clearance distance of the association analysis object set, and calculating a collision risk coefficient based on the intrinsic clearance distance, the predicted relative closest distance, and a preset reference reaction time; The collision risk warning level is matched according to the collision risk coefficient, and the collision risk warning level and the corresponding analysis subject machine-analysis object machine pair are output as the collision risk warning result.

5. The UAV collision risk warning system based on dynamic spatial grid is characterized by: The method for implementing the UAV collision risk warning method based on dynamic spatial grid according to any one of claims 1 to 4 comprises: A multidimensional indicator extraction module is used to obtain historical operation data of the target operation area based on preset spatiotemporal collection constraints, and correspondingly extract a multidimensional reference indicator set, wherein the multidimensional reference indicator set includes a spatial reference indicator set and a UAV reference indicator set; A dynamic space grid construction module, configured to construct a dynamic space grid based on the multi-dimensional reference indicator set, wherein the dynamic space grid includes an analysis space grid and a shared space grid; An analysis object determination module is configured to use the shared spatial grid as a constraint for drone data collection, traverse the analysis spatial grid to perform a preliminary screening of analysis objects, and obtain a set of associated analysis objects, wherein the set of associated analysis objects includes an analysis subject machine and an analysis object machine; The risk warning and output module is used to perform collision evaluation and warning on the associated analysis object set according to preset collision judgment conditions, and output collision risk warning results.

Images