Method and system for constructing low-altitude route risk map based on honeycomb grid units

Through the construction method of low-altitude route risk map based on cellular grid units, the problem of insufficient feasibility and scientificity of UAV route planning in the existing technology is solved, the accuracy and real-time nature of risk assessment are improved, and more scientific is achieved and flexible drone route planning.

CN118551182BActive Publication Date: 2024-11-05PANDA ELECTRONICS
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411017857.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-29
Publication Date
2024-11-05
Estimated Expiration
2044-07-29

AI Technical Summary

Technical Problem

The existing drone route planning methods cannot effectively comprehensively consider the impact of meteorological conditions, drone conditions, driver qualifications and laws and regulations on flight safety, resulting in insufficient feasibility and scientificity of planning routes, and the quadrilateral three-dimensional grid The degree of discretization is not high and the area coverage efficiency is low, which limits the accuracy and real-time performance of drone flight risk assessment.

Method used

Using a low-altitude route risk map construction method based on cellular mesh cells, the geographic surface is meshed through a hexagonal cellular mesh, a three-dimensional hierarchical mesh is constructed, and the UAV flight risks are quantitatively analyzed, and dynamic Bayes is used. The network conducts risk assessment and generates dynamic multi-scale low-altitude route three-dimensional cellular risk maps to improve the accuracy and real-time nature of risk assessment.

Benefits of technology

It improves the safety of drone route planning and the accuracy and real-time nature of risk assessment, meets different airspace control requirements, and provides more scientific risk assessment and more flexible terrain adaptability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118551182B_ABST
    Figure CN118551182B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and system for constructing a low-altitude route risk map based on honeycomb grid units. The method includes: based on different types of airspace control requirements, forming a low-altitude airspace three-dimensional honeycomb hierarchical grid through multi-scale division of the airspace horizontal plane and multi-scale division of the airspace vertical plane; constructing a drone flight risk assessment index system for the low-altitude airspace three-dimensional honeycomb hierarchical grid, after optimization, constructing a drone operation risk assessment model based on a dynamic Bayesian network, and generating a multi-scale regional risk value; combining the multi-scale regional risk value with the geographic location information of the region, designing a honeycomb grid coding index, and generating a dynamic multi-scale low-altitude route three-dimensional honeycomb risk map. The present invention can improve the accuracy and real-time performance of grid drone flight risk assessment, quickly derive low-risk grids for drones in target areas, and improve the safety of planned routes.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) route planning, and in particular to a method and system for constructing a low-altitude route risk map based on a honeycomb grid unit. Background Art

[0002] Faced with the reality of accelerating urbanization, increasing ground traffic congestion and limited underground space development, the development of low-altitude resources and the development of the low-altitude economy will bring unlimited possibilities for the future of urban transportation. In recent years, drones have been widely used in military and civilian fields. For example, in the military field, rescue search, surveillance reconnaissance, military strikes and other actual combat tasks; in the civilian field, urban air transport, forest protection, environmental monitoring and other application analysis. With the in-depth application of drones in various fields, the scale of air traffic has increased sharply, and the route planning and operation safety control of drones have become a cutting-edge hot topic.

[0003] Most of the existing UAV route planning methods are based on longitude and latitude, and the route planning of UAVs is realized by solving the three-dimensional curve equation plus the safety margin. The calculation complexity is high and it can no longer meet the increasingly dense low-altitude traffic control needs of aircraft. The invention patent "Path Planning Method, Device, Electronic Equipment and Route Planning Method (Application No. 202310752065.0)" proposes a three-dimensional path planning method for UAVs based on quadrilateral three-dimensional Beidou grid, which can effectively reduce the calculation complexity. However, this method also has shortcomings. First, it does not comprehensively consider the impact of factors such as meteorological conditions, UAV conditions, driver qualifications and flight restrictions in laws and regulations on the safety of UAV flight, resulting in insufficient feasibility and scientificity of planning routes. Secondly, the discretization degree of the quadrilateral three-dimensional grid is not high, and the regional coverage efficiency is low, which limits the accuracy and real-time performance of UAV flight risk assessment. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention proposes a method and system for constructing a low-altitude route risk map based on honeycomb grid units. The geographical surface is gridded using hexagonal honeycomb grids, and a honeycomb three-dimensional layered grid is constructed to quantitatively analyze the UAV flight risk of each grid. Finally, a dynamic multi-scale three-dimensional cellular risk map of low-altitude routes is generated, which improves the accuracy and real-time performance of risk assessment and enhances the safety of route planning.

[0005] Technical solution: In the first aspect, a method for constructing a low-altitude route risk map based on a honeycomb grid unit comprises the following steps:

[0006] Based on the requirements of different types of airspace control, a three-dimensional honeycomb layered grid of low-altitude airspace is formed through multi-scale division of the horizontal and vertical planes of the airspace;

[0007] Based on the drone flight related data set, a drone flight risk assessment index system for low-altitude airspace three-dimensional cellular layered grid is constructed;

[0008] The risk factors of the UAV flight risk assessment index system are optimized through feature screening algorithms;

[0009] According to the corresponding data set of the preferred risk factors, the UAV operation risk assessment model based on dynamic Bayesian network is used to infer the probability of UAV operation accidents and obtain the flight risk value of each cellular grid;

[0010] The sliding average method is used to compensate the error of the flight risk value of each honeycomb grid, and each honeycomb grid is weighted to calculate the regional risk weighted index. The compensated flight risk value of each honeycomb grid and the regional risk weighted index form a multi-scale grid risk value.

[0011] The multi-scale regional risk values ​​are combined with the geographic location information, and a dynamic multi-scale low-altitude route three-dimensional cellular risk map is generated according to the cellular grid coding index.

[0012] Furthermore, the multi-scale division method of the airspace horizontal plane is as follows:

[0013] Based on the controlled airspace range, determine the upper left corner of the horizontal plane of the corresponding area on the map as the origin of the earth's plane coordinates. The coordinates of this point are the origin coordinates. , and at the same time estimate the maximum horizontal and vertical lengths of the airspace range, and establish A flat honeycomb grid of specifications, where: n is the number of plane honeycomb grids on the horizontal axis, m is the number of plane honeycomb grids on the horizontal vertical axis, l is the minimum circumscribed circle diameter of a single plane honeycomb grid;

[0014] In is the origin, In the plane honeycomb grid map coordinate system established according to the specifications, indexes G(v) and g(x,y) are established for each plane honeycomb grid and the vertex of the grid, respectively, where v is the plane honeycomb grid number, G(v) represents the v A plane honeycomb grid, g(x,y) represents the vertex index value of the plane honeycomb grid, and Greater than the maximum horizontal length of the airspace area, Smaller than the maximum horizontal length of the airspace area, Greater than the maximum longitudinal length of the airspace area, Less than the maximum longitudinal length of the airspace area;

[0015] The multi-scale division method of the vertical plane of the airspace is as follows:

[0016] After dividing the horizontal plane honeycomb grid, based on the six vertices of each plane honeycomb grid, extend upward h The distance forms a hexagonal prism shape and establishes the index g(x,y,z) of the hexagonal prism vertex;

[0017] By estimating the maximum altitude of the controlled airspace, The three-dimensional honeycomb grid coordinate system of the specification, where h is the height of a single three-dimensional honeycomb grid, and Greater than the maximum altitude of the controlled airspace, Less than the maximum altitude of the controlled airspace.

[0018] Furthermore, based on the drone flight related data set, a drone flight risk assessment index system for low-altitude airspace three-dimensional cellular layered grid is constructed, including:

[0019] Obtain drone-related data from meteorological departments, drone flight control systems, drone registration systems, drone integrated supervision platforms, and drone obstacle detection and avoidance systems, remove outliers, and fill in missing values ​​to obtain a drone flight risk assessment dataset;

[0020] Based on the UAV flight risk assessment data set, the risk factors affecting the UAV flight safety are determined, and a UAV flight risk assessment index system is formed. The UAV flight risk assessment index system includes eight indicators: meteorological factors, flight restrictions, air risks, UAV conditions, obstacle conditions, ground conditions, driver conditions, and laws and regulations. Each indicator contains one or more risk factors.

[0021] Furthermore, the UAV operation risk assessment model based on dynamic Bayesian network is used to infer the probability of UAV operation accidents, and the flight risk value of each cellular grid is obtained, including:

[0022] Perform Pearson correlation analysis on the optimal risk factors, identify two risk factors whose absolute value of Pearson correlation coefficient is greater than the specified threshold as having correlation between them, analyze the causal relationship between them, connect these causal relationships with directed arrows, and obtain a static Bayesian network;

[0023] The static Bayesian network is extended on the time axis to generate a dynamic Bayesian network with two time slices. The risk factors that change over time are used as dynamic variables. The adjacent time slice nodes of each dynamic variable are connected using directed edges. The joint prediction probability of UAV flight safety risks across time slices is obtained according to the Bayesian formula.

[0024] For each three-dimensional honeycomb grid, the corresponding joint prediction probability of UAV flight safety risk is obtained through the above calculation method according to the actual situation of its risk factors. This probability value is the UAV flight risk value of the honeycomb grid.

[0025] Furthermore, the sliding average method is used to compensate the error of the flight risk value, including:

[0026] Take each three-dimensional honeycomb grid as the central hexagonal prism and perform the following calculations in sequence: take the six hexagonal prisms that share the same side surfaces as the central hexagonal prism grid and the two hexagonal prism grids that share the same bottom and top surfaces as the central hexagonal prism grid as neighbors, calculate the risk average of the neighboring grids and assign it to the central hexagonal prism grid.

[0027] Furthermore, the calculation formula of the regional risk-weighted index is as follows:

[0028] ;

[0029] In the formula, w n is the risk weight of the n-th three-dimensional honeycomb grid, s n is the risk value of the nth three-dimensional honeycomb grid, which is the joint prediction probability of UAV flight risk obtained through the dynamic Bayesian network, N is the number of three-dimensional honeycomb grids in the region, and V is the weighted risk index of the region.

[0030] Furthermore, the cellular grid code index consists of a 25-bit code in the following order: a 6-bit administrative region code, an 8-bit longitude and latitude code, a 2-bit cellular grid height, a 2-bit cellular grid diameter, a 2-bit multi-scale grid risk value, and a 5-bit serial number.

[0031] In the second aspect, a low-altitude route risk map construction system based on honeycomb grid units includes an airspace honeycomb grid construction module, a grid flight risk assessment module, and a route risk map construction module.

[0032] The airspace cellular grid construction module is used to form a low-altitude airspace three-dimensional cellular layered grid through multi-scale division of the airspace horizontal plane and multi-scale division of the airspace vertical plane based on different types of airspace control requirements;

[0033] A grid flight risk assessment module includes a risk assessment index system construction unit and a multi-scale regional risk assessment unit; the risk assessment index system construction unit is used to construct a UAV flight risk assessment index system for a three-dimensional cellular layered grid in low-altitude airspace according to a UAV flight-related data set; the multi-scale regional risk assessment unit is used to optimize the risk factors of the UAV flight risk assessment index system through a feature screening algorithm, and according to the corresponding data set of the optimized risk factors, the UAV operation risk assessment model based on the dynamic Bayesian network is used to infer the probability of UAV operation accidents, and obtain the flight risk value of each cellular grid; the sliding average method is used to perform error compensation on the flight risk value of each cellular grid, and each cellular grid is weighted to calculate the regional risk weighted index, and the flight risk value of each cellular grid after compensation and the regional risk weighted index form a multi-scale grid risk value;

[0034] The route risk map construction module is used to combine the multi-scale regional risk values ​​with the geographic location information, and generate a dynamic multi-scale low-altitude route three-dimensional cellular risk map based on the cellular grid coding index.

[0035] In a third aspect, a computer device comprises: a memory storing one or more programs; and a processor communicatively coupled to the memory and configured to be executed by the processor, wherein the program, when executed by the processor, implements the steps of the method for constructing a low-altitude route risk map based on cellular grid units as described in the first aspect of the present invention.

[0036] In a fourth aspect, a non-temporary processor-readable storage medium stores processor-executable instructions, which, when executed by a processor of a mobile terminal, implement the steps of a method for constructing a low-altitude route risk map based on cellular grid units as described in the first aspect of the present invention.

[0037] Beneficial effects: Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. A grid unit construction method based on regular hexagons was designed. The regular hexagonal grid is a polygon that can seamlessly fill a plane and has the largest number of edges connected to each vertex. It has the advantages of consistent neighborhood, isotropy, compactness, high sampling rate, and more flexible adaptation to terrain changes. Compared with the regular quadrilateral grid-based division algorithm in the existing UAV operation risk assessment algorithm, the risk assessment has higher accuracy and real-time performance.

[0039] 2. Based on the different flight safety control standards and requirements of different types of airspace in the country, a multi-scale hierarchical spatial grid demarcation method was designed to meet the needs of understanding the overall risk situation of the region and insight into the regional micro-risk characteristics.

[0040] 3. Based on the national UAV operation management standards and regulatory system, and combined with the actual operation of UAVs, a multi-level UAV flight risk assessment system has been established, including indicators such as meteorological conditions, UAV conditions, obstacles and pilot qualifications, to provide support for a more scientific and feasible assessment of UAV operation risks.

[0041] 4. The prominent role of dynamic Bayesian networks in causal uncertainty and correlation analysis was utilized to construct a UAV operation risk assessment model. The analysis results provided a strong basis for risk prevention and control management and route planning of UAVs.

[0042] 5. According to the actual air traffic control application needs of different resolutions, a regional weighted risk assessment index calculation method is designed to highlight the risk characteristics of drone operations in different regions. The sliding average model is used to compensate for the error of grid risk anomalies to improve the safety of planned routes. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a general schematic diagram of the method for constructing a low-altitude route risk map based on honeycomb grid units;

[0044] Figure 2 An example of delineation of a three-dimensional honeycomb grid for low-altitude airspace;

[0045] Figure 3 Schematic diagram of the coordinate division of the horizontal hexagonal grid;

[0046] Figure 4 It is a schematic diagram of the vertical plane height layered hexagonal prism index;

[0047] Figure 5 is the contribution and cumulative contribution of each risk assessment factor calculated by random forest;

[0048] Figure 6 This is a heat map of the correlation of factors affecting drone flight safety events;

[0049] Figure 7 Static Bayesian network structure diagram for drone flight risk assessment;

[0050] Figure 8 Dynamic Bayesian network structure diagram for drone flight risk assessment;

[0051] Fig. 9 A schematic diagram of the coding rules for grid risk assessment;

[0052] Fig.10 This is an example of the final generated three-dimensional cellular risk map. DETAILED DESCRIPTION

[0053] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0054] The present invention proposes a method for constructing a low-altitude route risk map based on a honeycomb grid unit, referring to Figure 1 , the method comprises the following steps:

[0055] Step 1: Construct a three-dimensional cellular grid in the low-altitude airspace.

[0056] In step 1, the present invention forms a low-altitude airspace three-dimensional honeycomb layered grid through multi-scale division of the airspace horizontal plane and multi-scale division of the airspace vertical plane based on different types of airspace control requirements.

[0057] According to the implementation mode of the present invention, a multi-layer geographic information library of the target area is formed through various remote sensing and mapping methods such as space-based, air-based and ground-based. While each layer of the geographic information library is generated, the data information is divided into a plane honeycomb grid of appropriate scale according to different geographical restrictions in the "National Airspace Basic Classification Method", and the plane honeycomb grid is divided into layers of different heights according to the airspace restriction height of the "National Airspace Basic Classification Method", forming a three-dimensional honeycomb layered grid in the low-altitude airspace.

[0058] As an example, data such as roads, terrain, elevation, slope, vegetation, rivers, grassland, cultivated land, digital elevation model DEM (Digital Elevation Model), three-dimensional digital ground model DTM (Digital Terrian Model), and meteorological and hydrological data in the target area are obtained to form a multi-layer geographic information database of the controlled low-altitude airspace area. Based on the multi-layer geographic information database, according to the control requirements for different types of airspace in the "National Airspace Basic Classification Method", the horizontal and vertical surfaces of the controlled low-altitude airspace are divided into multiple scales. For airspace with higher operational needs and safety requirements, a finer grid resolution division is adopted to provide more accurate flight safety control; conversely, a coarser grid resolution division is adopted. The corresponding relationship between airspace type and airspace demarcation accuracy is shown in Table 1. Among them, the horizontal diameter of the grid refers to the diameter of the circumscribed circle of the regular hexagon.

[0059] Table 1 Airspace types and airspace demarcation accuracy

[0060]

[0061] For example, a certain controlled low-altitude airspace includes three types: Class C, Class D, and Class G. According to the airspace types and airspace demarcation accuracy in Table 1, the three-dimensional honeycomb grid demarcation diagram is as follows: Figure 2As shown. The grid described in Table 1 refers to a spatial grid, also known as a three-dimensional honeycomb grid, which is a hexagonal prism, that is, the concept of a honeycomb grid unit in the present invention. Unless otherwise specified, the grid in the present invention refers to a three-dimensional honeycomb grid. The three-dimensional honeycomb grid is formed by dividing the horizontal plane into hexagons and setting the vertical plane in layers. For the sake of clarity of description, the horizontal plane hexagon is also called a plane honeycomb grid or a plane grid to distinguish it.

[0062] Taking Class C airspace (according to the National Airspace Basic Classification Method, Class C airspace is located above general aviation airports with control towers, usually a single-ring structure with a radius of 5 kilometers and a runway pavement-airport elevation of 600 meters (inclusive)) as an example, the division method is as follows.

[0063] Horizontal hexagonal grid (plane honeycomb grid) division method:

[0064] (1) After determining the scope of the controlled airspace, determine the earth plane coordinates of a point in the upper left corner of the horizontal plane of the area on the map, use the coordinates of this point as the origin, and establish the origin coordinates . At the same time, estimate the maximum horizontal and vertical lengths of the airspace range and establish The plane honeycomb grid of the specification is a regular hexagonal plane grid, where n is the number of plane honeycomb grids on the horizontal axis, and m is the number of plane honeycomb grids on the horizontal axis. l is the minimum circumscribed circle diameter of a single plane honeycomb grid, in meters. It must be greater than the maximum horizontal length of the airspace area. It must be smaller than the maximum horizontal length of the airspace area. It must be greater than the maximum longitudinal length of the airspace area. It must be smaller than the maximum longitudinal length of the airspace area.

[0065] (2) In is the origin, In the plane grid map coordinate system established according to the specifications, indexes G(v) and g(x,y) are established for each plane honeycomb grid and the vertex of the grid, respectively, where: v is the plane honeycomb grid number, G(v) represents the vth plane honeycomb grid, and g(x,y) represents the vertex index value of the plane honeycomb grid, with the coordinates in the plane grid map coordinate system as the index. The schematic diagram of plane honeycomb grid drawing is as follows Figure 3 As shown in the figure, the vertex index value of each plane honeycomb grid can be obtained by basic mathematical operations, which will not be described here.

[0066] Vertical surface height stratification method:

[0067] After dividing the horizontal plane honeycomb grid, based on the six vertices of each plane honeycomb grid, extend upward hThe distance forms a regular hexagonal prism (hexagonal prism for short), and the index g(x,y,z) of the vertex of the regular hexagonal prism is established. Taking the first regular hexagonal prism as an example, the index diagram is as follows Figure 4 The vertex index value of each hexagonal grid can be obtained by simple mathematical operations, which will not be described here.

[0068] Estimate the maximum altitude of the controlled airspace and establish The three-dimensional honeycomb grid coordinate system of the specification, where: h is the height of a single three-dimensional honeycomb grid, in meters, where It must be greater than the maximum altitude of the controlled airspace. It must be lower than the maximum altitude of the controlled airspace.

[0069] The present invention designs a method for constructing honeycomb grid units based on regular hexagons. Regular hexagonal grids are polygons that can seamlessly fill a plane and each vertex has the most edges connected. They have the advantages of consistent neighborhood, isotropy, compactness, high sampling rate, and more flexible adaptation to terrain changes. Compared with the regular quadrilateral grid-based division algorithm in the existing UAV operation risk assessment algorithm, the accuracy and real-time performance of risk assessment are higher. In addition, according to the relevant standards and requirements of different flight safety control in different types of airspace in the country, a multi-scale hierarchical spatial grid demarcation method is designed to meet the needs of managers for the overall risk situation of the region and the insight into the micro-risk characteristics of the region.

[0070] Step 2: Construct a three-dimensional cellular grid drone flight risk assessment index system.

[0071] According to the implementation mode of the present invention, a multi-level risk assessment system is established in a three-dimensional honeycomb grid, including indicators such as geographical restrictions, airspace restrictions, obstacles, meteorological conditions, aircraft parameters, and drone pilot qualifications. Specifically, the following steps are included:

[0072] Step 2.1: Data collection and preprocessing. Data on weather, flight performance, flight control requirements, drone pilot conditions, and obstacles are collected from meteorological departments, drone flight control systems, drone registration systems, drone integrated supervision platforms, and drone obstacle detection and avoidance systems. Meteorological data include temperature, humidity, rainfall, wind speed, etc. Flight performance data include maximum turning angle, flight speed, drone size, drone weight, etc. Flight control requirements include airspace type, altitude limit, time limit, speed limit, etc. Drone pilot conditions include drone pilot qualification level, flight duration, etc. Obstacle data include obstacle approach rate, obstacle size, obstacle speed, etc. A box plot is used to detect outliers in the data set and fill in the missing values ​​to obtain a drone flight risk assessment data set.

[0073] Step 2.2: Determine the drone flight risk assessment indicator system, including indicators such as meteorological factors, flight restrictions, air risks, drone conditions, obstacle conditions, ground conditions, pilot conditions, and laws and regulations.

[0074] Meteorological factors: Weather conditions are an important factor affecting the safety of drone flights. For example, severe weather conditions such as strong winds and heavy rains may cause the drone to lose control or affect the flight effect.

[0075] Flight restrictions: Different airspaces have restrictions on drone flight altitude, flight time period, and flight speed. If the restrictions are not followed, there may be risks such as exceeding the wireless communication range, intersecting with other vehicles, and aircraft crashes.

[0076] Air risks: Due to their high speed and strong maneuverability, drones are prone to collisions with obstacles, loss of control and other flight accidents. In addition, electromagnetic interference may cause drones to be unable to communicate and navigate.

[0077] Drone situation: In actual flight, the drone's maximum turning angle, flight speed, drone weight and other maneuverability performance, as well as drone battery failure, propeller damage and other faults will also affect the risk of drone collision.

[0078] Obstacles: During the flight of a drone, obstacles such as birds and other vehicles may suddenly appear in a local area. This sudden information is difficult to predict and can easily have a significant impact on the flight of the drone.

[0079] Ground conditions: If the drone loses control and there are a high density of people on the ground, serious injuries may occur.

[0080] Pilot situation: Pilots with different qualifications have different abilities to operate drones and deal with emergencies, which in turn affects the risks of drone operations.

[0081] Laws and regulations: Different regions have different laws and regulations that restrict the flight paths and areas of drones. Risk assessments need to consider these restrictions and assess the compliance of drone flights.

[0082] The UAV flight risk assessment index system established in the present invention is shown in Table 2, which includes eight categories and 22 risk factors.

[0083] Table 2 UAV flight risk assessment indicator system

[0084]

[0085] Based on the national UAV operation management standards and regulatory system and combined with the actual operation of UAVs, this invention has established a multi-level UAV flight risk assessment system including indicators such as meteorological conditions, UAV conditions, obstacles and pilot qualifications, providing support for a more scientific and feasible assessment of UAV operation risks.

[0086] Step 3: Construct a UAV operation risk assessment model based on a dynamic Bayesian network to evaluate the risk value of multi-scale three-dimensional cellular grids. This includes the following steps:

[0087] Step 3.1: Optimization of risk factors.

[0088] The original drone flight risk sample data set has many factors, and too many redundant factors may bring unnecessary noise and affect the model performance. Therefore, the random forest feature selection algorithm is used to screen out the optimal factor set from the original high-dimensional drone accident risk factor data set to improve the calculation efficiency and fitting accuracy of the model.

[0089] Random forest is an integrated supervised learning model based on decision trees. By calculating the contribution of each factor on each tree, the importance of the factor to the occurrence of drone flight accidents can be evaluated. The corresponding algorithm steps are as follows:

[0090] (1) Randomly extract K training data sets with replacement from the original sample set and K corresponding out-of-bag test data sets ;

[0091] (2) Set k = 1, in the training set D k Build a decision tree on B k Make predictions and count the prediction accuracy of out-of-bag data UAV flight accidents, recorded as ;

[0092] (3) For B k Factor X in i ,i=1,2,3,...,n apply random noise disturbance, and count the prediction accuracy of out-of-bag data ;

[0093] (4) For k = 1, 2, ..., K, repeat steps (2) and (3);

[0094] (5) Use the following formula to measure the factor X i Importance :

[0095] ,

[0096] (6) Arrange the factors in descending order of importance, remove redundant factors based on their cumulative contribution rates, and generate the optimal factor set.

[0097] According to the method described in step 3.1, 22 UAV flight safety risk factors were screened to obtain the contribution ranking of each factor and the cumulative contribution as follows: Figure 5 As shown in the figure, the cumulative contribution of the first 14 UAV flight risk factors is as high as 95.07%. The contribution of the last 8 factors is small and can be regarded as redundant factors. Therefore, the first 14 factors are selected as the input data of the Bayesian model: temperature, rainfall, wind speed, electromagnetic interference, obstacle approach rate, UAV failure, mid-air collision, UAV speed, UAV size, driver qualification, compliance, airspace type, population density, and ground loss.

[0098] Step 3.2: Operational risk distribution assessment based on dynamic Bayesian network.

[0099] Bayesian network (BN) is a graphical mathematical model for probabilistic reasoning. It uses a directed acyclic graph to describe the dependencies between variables. Given observed sample data or expert prior knowledge, BN (static Bayesian network) can use Bayesian probability theory to perform probabilistic reasoning on the target variable, and is widely used in uncertainty problems involving multiple variables. Dynamic Bayesian network (DBN) introduces the concept of time series on the basis of BN, and infers dynamic time-varying random processes by characterizing the time series transfer and causal relationship of random variables. The present invention discretizes the risk factor data set obtained by screening as the input of the dynamic Bayesian network model, and uses the maximum likelihood estimation method to learn the conditional probabilities of the initial network and the transition network.

[0100] The specific steps include:

[0101] Step 3.2.1: Construct a static Bayesian network for drone risk assessment.

[0102] The present invention constructs a static BN based on the 14 risk factors with the highest contribution rate obtained by screening. First, a Pearson correlation analysis is performed on each factor to draw a correlation heat map, such as Figure 6As shown. There is a certain correlation between two factors with an absolute value of the correlation coefficient greater than 0.1. For example, the correlation coefficient between population density and ground loss is 0.33, which is a positive correlation. The higher the population density, the greater the ground loss caused when the drone crashes. During the flight of the drone, the driver's qualifications, the flight speed of the drone, and the type of airspace in which it flies must comply with the provisions of laws and regulations related to drone flight; electromagnetic interference, excessive temperature, wind speed, rainfall and other factors may cause drone communication failures, navigation system failures, fuselage structure damage and other failures; drone failures, approaching obstacles, strong winds, and heavy rain may cause drones to have mid-air collisions / crashes and other accidents. In addition, whether a drone has an accident or not is also related to the driver's qualifications. Normally, the more qualified the driver is, the stronger the ability to operate the drone under harsh conditions, and the less likely it is to have an accident; the degree of ground loss is related to the population density of the ground below the flight, the occurrence of mid-air collisions / crashes, and the size of the drone; the comprehensive drone flight compliance risk, mid-air collision / crash risk, and ground loss risk form the drone operation risk. Use directed arrows to connect these causal relationships to obtain a static BN structure as shown below: Figure 7 shown.

[0103] The Bayesian network is intuitively a directed acyclic graph consisting of three elements: nodes and directed arcs, prior probability, and conditional probability. When applied to the risk assessment of drone safety incidents, the nodes in the Bayesian network are represented by random variables. It represents the various risk factors (risk factors) that affect the risk of drone flight accidents. Risk factor X i In the model, the probability of occurrence of the parent node, the risk factor X i The conditional probability can be expressed as , then the joint probability distribution of the overall risk of drone flight is As shown in the following formula, this formula can be used to predict the probability of drone flight safety risk event X.

[0104]

[0105] Step 3.2.2: Construct the UAV risk assessment dynamic Bayesian network DBN.

[0106] DBN is a probability model that integrates the original network structure and time information on the basis of the static Bayesian network and has the function of processing time series data. The DBN model uses directed edges to connect dynamic variables on the time axis, and mines the time series correlation through the transfer probability, so as to integrate the time factor to infer the probability of drone operation risk.

[0107] In the short term, risk factors such as temperature, wind speed, rainfall, drone speed, obstacle approach rate, drone failure, mid-air collision, flight operation risk, flight restriction time, and ground loss have strong temporal correlation and a large range of change over time. Their status at the previous moment and the current moment will continue to affect the drone flight risk and have a time accumulation effect. Therefore, the time factor is introduced into the BN model, and the status of risk factors such as temperature, wind speed, rainfall, drone speed, obstacle approach rate, drone failure, mid-air collision, flight operation risk, compliance, and ground loss at different times are combined for analysis.

[0108] The present invention constructs a DBN model of 2 time slices, and infers the risk probability of UAV operation based on the status of each factor at the previous moment and the current moment, so as to reflect the probability of an accident occurring in the flight of an aircraft at the current moment.

[0109] The static BN is extended on the time axis to generate a DBN with 2 time slices, such as Figure 8 As shown in the figure. In the DBN network, parameters such as temperature, drone speed, obstacle approach rate, and wind speed are dynamic variables. Therefore, directed edges are used to connect adjacent time slice nodes to reflect the time changes of the variables. The other static variables do not need to be connected between time slices. In order to ensure the stability and authenticity of the model results, the dynamic Bayesian network is trained and tested using the 5-fold cross-validation method, that is, the data set is divided into 5 mutually exclusive subsets of equal size by stratified sampling, and 4 of the subsets are used as training sets each time, and the remaining 1 subset is used as a test set. The node parameters of the dynamic Bayesian network are learned using the maximum likelihood estimation method, and then the probability of drone accidents is predicted for the test set samples. The drone operation risks are divided into different levels according to the probability of drone accidents.

[0110] Specifically, the DBN of drone flight safety risk is defined as , the Bayesian network at the initial moment is represented by B1, Represents the combined graph of the Bayesian network at different time slices. Based on the initial distribution and the conditional distribution between adjacent time slices, the dynamic Bayesian network is expanded to the Tth time slice, and a joint prediction probability of drone flight safety risk across multiple time slices can be obtained:

[0111]

[0112] Since different three-dimensional honeycomb grids correspond to different geographical locations, the variable values ​​of the risk factors corresponding to each grid are different. Therefore, for each three-dimensional honeycomb grid, a risk probability value can be obtained according to the above calculation method. This probability value is a risk prediction for the drone when flying through the grid.

[0113] The greater the predicted probability value of a safety incident that may occur during the flight of the drone, the higher the corresponding risk level. The corresponding relationship between the predicted probability and the risk level in the present invention is shown in Table 3 below.

[0114] Table 3 Correspondence between prediction probability and UAV operation risk level

[0115]

[0116] The present invention utilizes the prominent role of dynamic Bayesian networks in the analysis of causal uncertainty and correlation, constructs a UAV operation risk assessment model, determines the UAV operation risk level according to the predicted probability, and the analysis results provide a strong basis for risk prevention and control management and route planning of UAVs.

[0117] Step 3.3: Neighborhood grid risk value error compensation based on sliding average method.

[0118] The grid risk assessment value is affected by the obstacle detection system of the drone. For example, the sensor device in the obstacle detection system may introduce errors in the process of mutual inductance equipment synthesis and voltage transformer secondary circuit voltage drop, causing the obstacle detection result to deviate from the actual situation, which in turn affects the accurate assessment of the risk. These errors are limited by the internal structure and measurement principle of the metering device and are extremely difficult to detect and eliminate.

[0119] The present invention adopts a sliding average method to reduce the abnormal fluctuation level of the risk value in the grid. This method uses six hexagonal prisms that share the same side surfaces as the central hexagonal prism grid, and two hexagonal prism grids that share the same bottom and top surfaces as the central hexagonal prism grid as neighborhoods, calculates the risk average value of the neighborhood hexagonal prism grids, and assigns it to the central hexagonal prism grid.

[0120] Each hexagonal prism is taken as the central hexagonal prism grid for averaging. Let the central hexagonal prism grid V o The eight neighborhood hexagonal grids V1, V2, V3, ..., V8 are processed by sliding average. o The risk value is:

[0121]

[0122] In the above formula, R o is the hexagonal grid V o R1, R2, R3, …, R8 are the risk values ​​corresponding to the hexagonal grids V1, V2, V3, …, V8.

[0123] Step 3.4: Regional weighted risk assessment.

[0124] In actual air traffic control, there is a need to grasp the overall risk pattern and trend of the region, as well as a need for more detailed insights into the regional micro-risk characteristics. For example, for different control areas, such as above a residential area, above a hospital, above an important infrastructure, above a nature reserve, etc., the need to grasp the risk pattern and risk characteristics is different. According to the actual air traffic control application needs of different coarse and fine resolutions, the present invention designs a method for calculating the regional weighted risk index.

[0125]

[0126] In the formula, w n is the risk weight of the n-th three-dimensional honeycomb grid, s n is the risk value of the nth three-dimensional honeycomb grid, N is the number of three-dimensional honeycomb grids in the region, and V is the weighted risk index of the region.

[0127] In the present invention, according to the risk level of the UAV operation, the risk weight assessment value of the three-dimensional honeycomb grid is determined as shown in Table 4. The grid risk level described in Table 4 refers to the UAV operation risk level obtained above.

[0128] Table 4 Grid weighted risk assessment weights

[0129]

[0130] The present invention uses a sliding average model to compensate for the errors of grid risk anomalies and improve the safety of planned routes. According to the actual air traffic control application needs of different coarse and fine resolutions, a regional weighted risk assessment index calculation method is designed to highlight the risk characteristics of drone operations in different regions. In the present invention, the compensated flight risk value and regional risk weighted index of each honeycomb grid form a multi-scale grid risk value, so that the drone operation risk can be comprehensively grasped from a single grid and regional level. For example, a large-scale grid can be a large area composed of many grids; a small-scale grid can be a small area composed of fewer grids, or it can be a single grid.

[0131] Step 4: Combine the multi-scale grid risk values ​​and geographic spatial locations to generate a dynamic multi-scale low-altitude route three-dimensional cellular risk map based on the grid coding mechanism.

[0132] The present invention establishes a grid risk assessment coding mechanism to quantify the safety risk of each spatial grid to the flight of unmanned equipment, and utilizes the efficient computing characteristics of the grid code to quickly obtain the low-risk grids that the unmanned equipment can pass through in the target area, thereby planning a reasonable route.

[0133] Each grid risk assessment is uniquely identified by an identification code, which consists of 25 digits in the following order: administrative area code (6 digits), latitude and longitude code (8 digits), cell height (2 digits), cell diameter (2 digits), risk value (2 digits), and serial number (5 digits). Fig. 9 , where the description of each field is shown in Table 5.

[0134] Table 5 Description of each field of the cellular grid code

[0135]

[0136] Fig.10 An example of the final generated three-dimensional cellular risk map is shown, which is an example of risk maps for three different areas: low risk, relatively low risk, and medium risk.

[0137] Based on the same technical concept as the method embodiment, the present invention also provides a low-altitude route risk map construction system based on honeycomb grid units, including an airspace honeycomb grid construction module, a grid flight risk assessment module, and a route risk map construction module. The connection relationship between each module is as follows: Figure 1 shown.

[0138] The airspace honeycomb grid construction module is used to form a low-altitude airspace three-dimensional honeycomb layered grid through multi-scale division of the airspace horizontal plane and multi-scale division of the airspace vertical plane based on different types of airspace control requirements. The method of multi-scale division of the airspace horizontal plane and vertical plane height stratification refers to the description in the above method embodiment.

[0139] The grid flight risk assessment module is used to build a UAV operation risk assessment model based on a dynamic Bayesian network based on the cellular three-dimensional grid UAV flight risk assessment index system, and evaluate the multi-scale grid risk value.

[0140] It includes risk assessment indicator system construction unit and multi-scale regional risk assessment unit.

[0141] The risk assessment index system construction unit is: determining the drone flight risk assessment index system, including meteorological factors, flight restrictions, air risks, drone conditions, obstacle conditions, ground conditions, driver conditions and legal and regulatory compliance indicators, and each type of indicator contains one or more assessment factors.

[0142] The multi-scale regional risk assessment unit: optimizes the risk factors of the drone flight risk assessment index system through a feature screening algorithm, and uses a drone operation risk assessment model based on a dynamic Bayesian network to infer the probability of drone operation accidents according to the corresponding data set of the optimized risk factors, and obtains the flight risk value of each honeycomb grid; uses a sliding average method to compensate the error of the flight risk value of each honeycomb grid, and assigns weights to each honeycomb grid to calculate the regional risk weighted index. The compensated flight risk value of each honeycomb grid and the regional risk weighted index form a multi-scale grid risk value. The specific calculation method involved can refer to the description of the method embodiment.

[0143] The route risk map construction module is used to combine the multi-scale regional risk values ​​with the regional geographic location information to generate a dynamic multi-scale low-altitude route three-dimensional cellular risk map. The cellular grid coding dynamic index is designed, and the identification code content includes administrative area code, current time, risk value and risk level data. By using the efficient calculation characteristics of the grid unit, the low-risk grid of the drone in the target area can be obtained, and then a reasonable route can be quickly planned.

[0144] The present invention also provides a computer device, comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and when the programs are executed by the processors, the steps of the method for constructing a low-altitude route risk map based on cellular grid units as described above are implemented.

[0145] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method for constructing a low-altitude route risk map based on cellular grid units as described above are implemented.

[0146] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0147] The present invention is described with reference to flowcharts of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process in the flowchart, as well as a combination of processes in the flowchart, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart. Figure 1 A device that specifies functions in a process or multiple processes.

[0148] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A function specified in a process or multiple processes.

[0149] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 The steps of a specified function in a process or multiple processes.

Claims

1. A method for constructing a low-altitude route risk map based on honeycomb grid units. It is characterized in that The following steps are involved: Based on the requirements of different types of airspace control, a three-dimensional honeycomb layered grid of low-altitude airspace is formed through multi-scale division of the horizontal and vertical planes of the airspace; According to the drone flight related data set, a drone flight risk assessment index system for low-altitude airspace three-dimensional cellular layered grid is constructed, including: obtaining drone related data from the meteorological department, drone flight control system, drone registration system, drone integrated supervision platform and drone obstacle detection and avoidance system, removing abnormal values ​​and filling vacant values, and obtaining a drone flight risk assessment data set; according to the drone flight risk assessment data set, determining the risk factors affecting drone flight safety, and forming a drone flight risk assessment index system, the drone flight risk assessment index system includes eight indicators: meteorological factors, flight restrictions, air risks, drone conditions, obstacle conditions, ground conditions, driver conditions and laws and regulations, and each indicator contains one or more risk factors; The risk factors of the UAV flight risk assessment index system are optimized through feature screening algorithms; According to the corresponding data set of the preferred risk factors, the UAV operation risk assessment model based on dynamic Bayesian network is used to infer the probability of UAV operation accidents and obtain the flight risk value of each cellular grid; The sliding average method is used to compensate the error of the flight risk value of each honeycomb grid, and each honeycomb grid is weighted to calculate the regional risk weighted index. The compensated flight risk value of each honeycomb grid and the regional risk weighted index form a multi-scale grid risk value. The multi-scale grid risk values ​​are combined with the geographic location information, and a dynamic multi-scale low-altitude route three-dimensional cellular risk map is generated according to the cellular grid coding index.

2. The method according to claim 1, It is characterized in that The multi-scale division method of the airspace horizontal plane is as follows: Based on the controlled airspace range, determine the upper left corner of the horizontal plane of the corresponding area on the map as the origin of the earth's plane coordinates. The coordinates of this point are the origin coordinates. , and at the same time estimate the maximum horizontal and vertical lengths of the airspace range, and establish A flat honeycomb grid of specifications, where: n is the number of plane honeycomb grids on the horizontal axis, m is the number of plane honeycomb grids on the horizontal vertical axis, l is the minimum circumscribed circle diameter of a single plane honeycomb grid; In is the origin, In the plane honeycomb grid map coordinate system established according to the specifications, indexes G(v) and g(x,y) are established for each plane honeycomb grid and the vertex of the grid, respectively, where v is the plane honeycomb grid number, G(v) represents the v A plane honeycomb grid, g(x,y) represents the vertex index value of the plane honeycomb grid, and Greater than the maximum horizontal length of the airspace area, Less than the maximum horizontal length of the airspace area, Greater than the maximum longitudinal length of the airspace area, Less than the maximum longitudinal length of the airspace area; The multi-scale division method of the vertical plane of the airspace is as follows: After dividing the horizontal plane honeycomb grid, based on the six vertices of each plane honeycomb grid, extend upward h The distance forms a hexagonal prism shape and establishes the index g(x,y,z) of the hexagonal prism vertex; By estimating the maximum altitude of the controlled airspace, The three-dimensional honeycomb grid coordinate system of the specification, where h is the height of a single three-dimensional honeycomb grid, and Greater than the maximum altitude of the controlled airspace, Less than the maximum altitude of the controlled airspace.

3. The method according to claim 1, It is characterized in that The UAV operation risk assessment model based on dynamic Bayesian network is used to infer the probability of UAV operation accidents, and the flight risk value of each cellular grid is obtained, including: Perform Pearson correlation analysis on the optimal risk factors, identify two risk factors whose absolute value of Pearson correlation coefficient is greater than the specified threshold as having correlation between them, analyze the causal relationship between them, connect these causal relationships with directed arrows, and obtain a static Bayesian network; The static Bayesian network is extended on the time axis to generate a dynamic Bayesian network with two time slices. The risk factors that change over time are used as dynamic variables. The adjacent time slice nodes of each dynamic variable are connected using directed edges. The joint prediction probability of UAV flight safety risks across time slices is obtained according to the Bayesian formula. For each three-dimensional honeycomb grid, the corresponding joint prediction probability of UAV flight safety risk is obtained through the above calculation method according to the actual situation of its risk factors. This probability value is the UAV flight risk value of the honeycomb grid.

4. The method according to claim 2, It is characterized in that The sliding average method is used to compensate the error of the flight risk value, including: Take each three-dimensional honeycomb grid as the central hexagonal prism and perform the following calculations in sequence: take the six hexagonal prisms that share the same side surfaces as the central hexagonal prism grid and the two hexagonal prism grids that share the same bottom and top surfaces as the central hexagonal prism grid as neighbors, calculate the risk average of the neighboring grids and assign it to the central hexagonal prism grid.

5. The method according to claim 1, It is characterized in that The regional risk-weighted index is calculated as follows: ; In the formula, w n is the risk weight of the n-th three-dimensional honeycomb grid, s n is the risk value of the nth three-dimensional honeycomb grid, which is the joint prediction probability of UAV flight risk obtained through the dynamic Bayesian network, N is the number of three-dimensional honeycomb grids in the region, and V is the weighted risk index of the region.

6. The method according to claim 1, It is characterized in that The cellular grid code index consists of 25-bit codes in the following order: 6-bit administrative area code, 8-bit longitude and latitude code, 2-bit cellular grid height, 2-bit cellular grid diameter, 2-bit multi-scale grid risk value, and 5-bit serial number.

7. A low-altitude route risk map construction system based on honeycomb grid units, It is characterized in that include: The airspace cellular grid construction module is used to form a low-altitude airspace three-dimensional cellular layered grid through multi-scale division of the airspace horizontal plane and multi-scale division of the airspace vertical plane based on different types of airspace control requirements; The grid flight risk assessment module includes a risk assessment index system construction unit and a multi-scale regional risk assessment unit; the risk assessment index system construction unit is used to construct a UAV flight risk assessment index system for a three-dimensional cellular layered grid in low-altitude airspace according to a UAV flight-related data set; the multi-scale regional risk assessment unit is used to optimize the risk factors of the UAV flight risk assessment index system through a feature screening algorithm, and according to the corresponding data set of the optimized risk factors, the UAV operation risk assessment model based on the dynamic Bayesian network is used to infer the probability of UAV operation accidents, and obtain the flight risk value of each cellular grid; The sliding average method is used to compensate the error of the flight risk value of each honeycomb grid, and each honeycomb grid is weighted to calculate the regional risk weighted index. The compensated flight risk value of each honeycomb grid and the regional risk weighted index form a multi-scale grid risk value. The route risk map construction module is used to combine the multi-scale grid risk value with the geographic location information and generate a dynamic multi-scale low-altitude route three-dimensional cellular risk map based on the cellular grid code index; The risk assessment index system construction unit constructs a UAV flight risk assessment index system for low-altitude airspace three-dimensional cellular hierarchical grids based on the UAV flight related data set, including: Obtain drone-related data from meteorological departments, drone flight control systems, drone registration systems, drone integrated supervision platforms, and drone obstacle detection and avoidance systems, remove outliers, and fill in missing values ​​to obtain a drone flight risk assessment dataset; Based on the UAV flight risk assessment data set, the risk factors affecting the UAV flight safety are determined, and a UAV flight risk assessment index system is formed. The UAV flight risk assessment index system includes eight indicators: meteorological factors, flight restrictions, air risks, UAV conditions, obstacle conditions, ground conditions, driver conditions, and laws and regulations. Each indicator contains one or more risk factors.

8. A computer device, It is characterized in that include: A memory storing one or more programs; and a processor, communicatively coupled to the memory and configured to be executed by the processor, and when the program is executed by the processor, the steps of the method for constructing a low-altitude route risk map based on cellular grid units as described in any one of claims 1 to 6 are implemented.

9. A non-transitory processor-readable storage medium having processor-executable instructions stored thereon, It is characterized in that When the processor executable instructions are executed by the processor of the mobile terminal, the steps of the method for constructing a low-altitude route risk map based on cellular grid units as described in any one of claims 1-6 are implemented.