Urban unmanned aerial vehicle operation comprehensive risk assessment method and system

Abstract

The invention discloses an urban unmanned aerial vehicle operation comprehensive risk assessment method and system, and belongs to the technical field of unmanned aerial vehicle low-altitude safety management. The urban unmanned aerial vehicle operation comprehensive risk assessment method comprises the following steps: firstly, based on urban GIS data and a fishing net segmentation technology, constructing a three-dimensional dynamic scene model fusing building topology, available low-altitude airspace and population flow characteristics; establishing a direct death risk model, an indirect death risk model and a property loss model in combination with unmanned aerial vehicle operation parameters and time-varying population distribution data, and performing weighted synthesis based on normalized risk values to obtain a comprehensive risk model; and based on a space-time risk thermodynamic diagram generation technology, outputting an urban low-altitude risk situation map with different time periods, different regions and different heights. According to the method, the time-varying characteristic and the multi-dimensional influence of the operation risk of the unmanned aerial vehicle can be quantified for an urban low-altitude unmanned aerial vehicle operation risk scene of refined modeling, and decision support is provided for urban unmanned aerial vehicle path planning, airspace management and emergency response.

Description

Technical Field

[0001] The present invention relates to the technical field of low-altitude safety management of unmanned aerial vehicles (UAVs), and in particular to a comprehensive risk assessment method and system for urban UAV operations. Background Art

[0002] With the widespread use of drones in urban environments, the third-party risks they bring when performing missions, especially casualties and property losses, have become key challenges that need to be addressed urgently. The complexity and variability of urban low-altitude airspace require a more refined and dynamic assessment of the risks during drone flights. However, most existing risk assessment methods rely on static models, which usually ignore the impact of time-varying factors in urban environments, especially the significant impact of dynamic changes in population density within a day on the risk of death. Therefore, when facing complex urban environments, traditional risk assessment methods based on static data are difficult to accurately reflect the spatiotemporal fluctuation characteristics of risks during drone flights, and it is difficult to provide timely and effective warnings and decision support for the safe flight of drones.

[0003] In addition, existing risk assessment models usually focus on a single risk factor, such as the probability of a drone’s system failure or collision, but pay insufficient attention to multi-dimensional, multi-factor comprehensive assessments. Factors such as building density and population mobility in different areas of a city will have varying degrees of impact on the operational safety of drones, and traditional assessment methods fail to effectively integrate these factors. Summary of the invention

[0004] The purpose of the present invention is to provide a comprehensive risk assessment method and system for urban drone operations, aiming to provide a more accurate means of drone operation risk assessment. By combining factors such as urban dynamic population distribution and building density, a comprehensive risk assessment model that adapts to spatiotemporal changes is constructed, so that third-party risks during drone flight can be more accurately assessed and more effective decision support can be provided.

[0005] In order to solve the above technical problems, the first aspect of the present application provides a comprehensive risk assessment method for urban drone operations, comprising the following steps:

[0006] Step 1: Use urban GIS data and fishing net segmentation to build a 3D dynamic scene model that integrates building topology, available low-altitude airspace, and population flow characteristics;

[0007] Step 2: Based on the drone operation parameters and time-varying population distribution and building data, establish the direct death risk, indirect death risk and property loss risk models, and use the normalized risk value to perform a weighted synthesis of the comprehensive risk assessment model;

[0008] Step 3: Based on the generation of spatiotemporal risk heat map, output the urban low-altitude risk situation map by time period, region and altitude.

[0009] Accordingly, the second aspect of the present application provides a comprehensive risk assessment system for urban drone operations, comprising:

[0010] The 3D dynamic scene model construction module is used to use urban GIS data and fishing net segmentation to build a 3D dynamic scene model that integrates building topology, available low-altitude airspace and population flow characteristics;

[0011] A comprehensive risk assessment model building module is used to establish direct mortality risk, indirect mortality risk and property loss risk models based on drone operation parameters and time-varying population distribution and building data, and to synthesize a comprehensive risk assessment model by weighting based on normalized risk values;

[0012] The urban low-altitude risk situation map output module is used to generate urban low-altitude risk situation maps based on time-space risk heat maps, and output urban low-altitude risk situation maps by time period, region, and altitude.

[0013] Compared with the prior art, the present invention has the following beneficial effects:

[0014] 1. This paper proposes a UAV operation risk assessment method based on time-varying population density and a multi-dimensional risk assessment framework. Compared with the traditional static model, this paper can consider the time-varying characteristics of population distribution in the urban environment, significantly improving the accuracy and timeliness of risk assessment. By comprehensively considering multiple factors such as building density, available airspace, and population mobility, it can more finely reflect the risk changes in different areas, different heights, and different time periods, effectively making up for the lack of adaptability of existing technologies to complex urban environments.

[0015] 2. The present invention realizes accurate risk visualization on the ArcGIS platform, and improves the operability and intuitiveness of risk assessment results. Through the setting of symbol system and highly layered query function, combined with spatiotemporal dynamic data, the present invention provides accurate risk situation maps by region, altitude and time period, which can provide efficient and practical decision support for drone path planning, airspace management and emergency response, and significantly enhance the safety and management efficiency of urban airspace. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is an overall implementation flow chart of the comprehensive risk assessment method for urban drone operations of the present invention.

[0017] Figure 2 This is a diagram of the urban functional area division of Baijia Lake area in Jiangning District, Nanjing City according to an embodiment of the present invention.

[0018] Figure 3This is a scene diagram of urban available low-altitude airspace modeling according to an embodiment of the present invention.

[0019] Figure 4 This is a heat map of population distribution at 12:00 on August 20, 2024 according to an embodiment of the present invention.

[0020] Figure 5 This is a risk map at a height of 40 meters at 12:00 on August 20, 2024, which only considers the direct risk of death in an embodiment of the present invention.

[0021] Figure 6 This is a risk map at a height of 40 meters at 12:00 on August 20, 2024, which only considers the indirect death risk in an embodiment of the present invention.

[0022] Figure 7 This is a risk map at a height of 40 meters at 12:00 on August 20, 2024, which only considers the risk of property loss in an embodiment of the present invention.

[0023] Figure 8 This is a risk map at a height of 40 meters at 24:00 on August 20, 2024, which takes into account comprehensive risks in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The present invention is described below based on examples, but the present invention is not limited to these examples. In the following detailed description of the present invention, some specific details are described in detail to facilitate the understanding of those skilled in the art.

[0025] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described below with reference to the accompanying drawings.

[0026] See also Figure 1 In one embodiment, the present invention proposes a comprehensive risk assessment method for urban drone operations, comprising the following steps:

[0027] Step 1: Use urban GIS data and fishing net segmentation to build a 3D dynamic scene model that integrates building topology, available low-altitude airspace, and population flow characteristics;

[0028] Step 2: Based on the drone operation parameters and time-varying population distribution and building data, establish the direct death risk, indirect death risk and property loss risk models, and use the normalized risk value to perform a weighted synthesis of the comprehensive risk assessment model;

[0029] Step 3: Based on the spatiotemporal risk heat map generation technology, output the urban low-altitude risk situation map by time period, region, and altitude.

[0030] Specifically, step 1 includes:

[0031] Step 11: Import the 2D vector data of urban buildings in ArcGIS Pro, and convert the original data into an appropriate projection coordinate system through projection and coordinate transformation operations. Next, set the 3D extrusion parameters, use the building height information to convert the 2D building data into a 3D structural model, and generate an accurate 3D building entity.

[0032] Step 12: Based on the specific geographic scope of the study area, the fishnet tool provided by ArcGIS Pro is used to divide the area into regular grid units, and a marked point, i.e., a fishnet point, is generated at the center point of each grid. Subsequently, the coordinates of the input point feature set are read, and new feature sets are inserted at different height layers by setting the height range and step length to complete the conversion of the fishnet point from a two-dimensional plane to a three-dimensional space. The five functional areas mainly included in the modeled urban scene are specifically divided as follows: Figure 2 As shown;

[0033] Step 13: Use ArcPy programming tools to perform spatial queries, and combine the layer management function to analyze the overlap between each layer of fishing net points and buildings at each altitude layer, and mark the covered fishing net points as obstacle points. Create a temporary layer in each altitude layer, locate the points that overlap with the buffer surface through the spatial query function, and calculate and record the number of overlapping points. De-duplicate the data by creating a point ID set to avoid repeated processing of the same fishing net points. Save the unavailable fishing net points that have been screened out in a separate layer to form the final unavailable point data set. Traverse the three-dimensional fishing net points, mark the fishing net points that are not in the unavailable point set as available points, and insert them into a new feature class. Finally, ensure that the retained fishing net points are completely within the available range to form the final available airspace data set. In the embodiment, the upper limit of the modeling height of the airspace is set to 120 meters, and the interval between fishing net points is set to 10 meters. In order to ensure the safety of the building, a 20-meter buffer zone is set to define the safe distance between the drone and the building. The available low-altitude airspace modeling scene diagram is as shown in the following figure. Figure 3 shown.

[0034] Step 14: Using Baidu Huiyan population density data, import the population density data of different time periods into the attribute table of the fishing net points. First, perform kernel density analysis on the population Excel file showing XY data to generate the corresponding population heat map. Then, use ArcPy and Spatial Analyst extension tools to batch extract population density values ​​from multiple raster files and add them to the fields of the point feature class. In this way, the fishing net points can obtain the population density attributes of different time periods, thereby realizing the construction of a three-dimensional scene containing building topology, available low-altitude airspace, and population flow characteristics. Figure 4 The heat map of population distribution in the modeling scenario area at 12:00 on August 20, 2024 is shown.

[0035] Furthermore, in one embodiment of step 2, the selected drone model is DJI Mavic 3-Pro.

[0036] Step 2 specifically includes the following steps:

[0037] Step 21: Construct a direct death risk model based on the aircraft model data and the scenario information in S1 :

[0038]

[0039] in The probability of a drone crashing is determined by the reliability of the drone system’s hardware and software.

[0040] , represents the probability of a drone hitting a person after falling, where Indicates the skin penetration coefficient. Skin penetration occurs at high kinetic energy density. When the drone falls, the carbon fiber frame causes different degrees of damage to human skin. represents the population density on the ground at time t; It is the contact area between the drone and the person when the drone falls and hits the person.

[0041] , represents the probability of a drone crashing into a person and causing death, where S C represents the occlusion factor, which is a real number in the range of (1, 1]; , represents the impact kinetic energy; and Represents two constants used to define the energy threshold, It is the threshold of impact kinetic energy required to cause human death when the shielding factor approaches 0.

[0042] , represents the number of people hit by the drone after it falls.

[0043]

[0044] represents the area expected to be hit by the drone, r UAV is the radius of the drone, r human For the radius of a person, the impact area can be estimated using the radius of the drone and the radius of the person; is the population density on the ground at time t.

[0045] Step 22: When a drone crashes into a building, there is a risk of indirect casualties. The indirect death risk model D is constructed using the regional building density division method. i :

[0046]

[0047] in, is the probability of the drone crashing; , represents the number of buildings affected by the impact after the drone falls, Indicates the impact area after the drone crashes. Represents the building density within the functional area m in the scene; Represents the average number of casualties resulting from collisions between drones and buildings.

[0048] Step 23: The potential collision between drones and buildings will cause property loss risk. Using the density and height distribution characteristics of buildings in the area, a property loss risk model is constructed:

[0049] is the probability density function of the lognormal distribution, which is used to calculate a specific height The property loss risk density under is the risk cost of property loss during drone operation, where It is a functional area Building density within the area; Indicates functional area The geometric mean of the building heights; Indicates functional area logarithmic standard deviation of building heights;

[0050]

[0051] is the height factor, which indicates the attenuation of the impact of building density on risk with height. For functional areas The decay rate parameter within.

[0052] Step 24: Integrate the three risk models to construct the final comprehensive risk assessment model:

[0053]

[0054] in, , , is the risk weight; , , They represent the three risk values ​​of direct death risk value, indirect death risk value and property loss risk value after normalization:

[0055] .

[0056] Furthermore, in one embodiment, in step 3, based on the technology of generating a spatiotemporal risk heat map, a city low-altitude risk situation map by time period, region, and altitude can be output. Specifically, the following steps are included:

[0057] Step 31: Combine the risk model in step 2 with specific drone parameters and scene information, and use ArcPy tools to add the independent risk values ​​and comprehensive risk values ​​of the fishing net points corresponding to each time period to the attribute table. That is, each fishing net point represents a square-sized airspace, and the three sub-risks and comprehensive risks of this airspace are calculated using the formula in step 2. These quantified risk values ​​are imported into the fishing net points corresponding to their respective square airspaces, so that the fishing net points have the attributes of representing risks.

[0058] Step 32: In ArcGIS Pro, visualize different risks by setting symbol systems. Based on the four risk fields (comprehensive risk, Dd, Di, Cp_r) added in step 31, select the fields to be displayed for symbolization. By setting gradient colors or other symbol styles, highlight the spatial distribution of different risk values ​​to help visually identify high-risk areas and risk changes in different time periods. Figure 5-Figure 7 The drone operation risk map at an altitude of 40 meters at 12:00 on August 20, 2024, considering only the direct death risk, indirect death risk, and property loss risk, is shown respectively. From the figure, it can be clearly and intuitively observed that the different risk distributions at specific times and altitudes in the case study area, and the specific values ​​are clear at a glance, realizing the quantitative characterization and visual expression of risks.

[0059] Step 33: Define the query tool to display the corresponding risk distribution according to different altitudes. Set query conditions for each time period and altitude to filter out the corresponding fishing net point data. This can generate a city low-altitude risk situation map with spatiotemporal characteristics, divided by time period, region, and altitude, showing the risk changes of different altitudes in different time periods, and finally forming a complete risk situation map. Figure 8 The comprehensive risk map of drone operations at an altitude of 40 meters at 12:00 on August 20, 2024 is displayed, with the weights of the three risks set to 0.5, 0.25, and 0.25 respectively. The figure shows the distribution of risk levels in different areas at this time and altitude, as well as the spatial location of high-risk areas. This multi-dimensional analysis method based on time, altitude, and space makes the risk visualization of low-altitude airspace more accurate and can provide more reliable support for the safe operation of drones.

[0060] The present invention has many specific application paths. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements can be made without departing from the principle of the present invention. These improvements should also be regarded as the protection scope of the present invention.

Claims

1. A comprehensive risk assessment method for urban drone operations, characterized in that: The following steps are involved: Step 1: Use urban GIS data and fishing net segmentation to build a 3D dynamic scene model that integrates building topology, available low-altitude airspace, and population flow characteristics; Step 2: Based on the drone operation parameters and time-varying population distribution and building data, establish the direct death risk, indirect death risk and property loss risk models, and use the normalized risk value to perform a weighted synthesis of the comprehensive risk assessment model; Step 3: Based on the generation of spatiotemporal risk heat map, output the urban low-altitude risk situation map by time period, region and altitude.

2. The comprehensive risk assessment method for urban drone operation according to claim 1 is characterized in that: Step 1 includes: Step 11: Import the 2D vector data of urban buildings in ArcGIS Pro, and convert the original data into an appropriate projection coordinate system through projection and coordinate transformation operations; set 3D extrusion parameters, and use the building height information to convert the 2D building data into a 3D structural model to generate an accurate 3D building entity; Step 12: According to the specific geographic scope of the study area, use the fishnet tool provided by ArcGIS Pro to set appropriate pixel width and height, divide the area into regular grid cells, and generate annotation points at the center of each grid, namely fishnet points; read the coordinates of the input point feature set, insert new feature sets at different height layers by setting the height range and step size, and complete the conversion of fishnet points from two-dimensional plane to three-dimensional space; Step 13: Use ArcPy programming tools to perform spatial queries, and combine the layer management function to analyze the overlap between each layer of fishing net points and buildings at each height layer, and mark the covered fishing net points as obstacle points; create a temporary layer in each height layer, locate the points overlapping with the buffer surface through the spatial query function, and calculate and record the number of overlapping points; deduplicate the data by creating a point ID set; save the filtered unusable fishing net points in a separate layer to form the final unusable point dataset; traverse the three-dimensional fishing net points, mark the fishing net points that are not in the unusable point set as available points, and insert them into the new feature class; finally, ensure that the retained fishing net points are completely within the available range to form the final available airspace dataset; Step 14: Using Baidu Smart Eye population density data, import the population density data of different time periods into the attribute table of the fishing net points to complete the construction of a three-dimensional scene including building topology, available low-altitude airspace and population flow characteristics.

3. The comprehensive risk assessment method for urban drone operation according to claim 1 is characterized in that: The method for constructing the direct mortality risk model in step 2 includes: Construct a direct death risk model based on the aircraft model data and the scenario information in S1 : ； in is the probability of the drone crashing; represents the probability of a drone hitting a person after falling, where represents the skin penetration coefficient, represents the population density on the ground at time t; The contact area between the drone and the person when the drone falls and hits the person; represents the probability of a drone crashing into a person and causing death, where S C represents the occlusion factor, which is a real number in the range of (1, 1]; , represents the impact kinetic energy; and Represents two constants used to define the energy threshold, It is the threshold of impact kinetic energy required to cause human death when the shielding factor approaches 0; , represents the number of people hit by the drone after it falls; ； represents the area expected to be hit by the drone, r UAV is the radius of the drone, which depends on the model of the drone, r human The radius of a person; is the population density on the ground at time t.

4. The comprehensive risk assessment method for urban drone operation according to claim 3 is characterized in that: The method for constructing the indirect mortality risk model in step 2 includes: Constructing the indirect mortality risk model D using the regional building density division method i : ； in, is the probability of the drone crashing; , represents the number of buildings affected by the impact after the drone falls, Indicates the impact area after the drone crashes. Represents the building density within the functional area m in the scene; Represents the average number of casualties resulting from collisions between drones and buildings.

5. The comprehensive risk assessment method for urban UAV operation according to claim 4 is characterized in that: The method for constructing the property loss risk model in step 2 includes: Using the density and height distribution characteristics of buildings in the region, a property loss risk model is constructed: ； is the probability density function of the lognormal distribution, which is used to calculate a specific height The property loss risk density under is the risk cost of property loss during drone operation, where It is a functional area Building density within the area; Indicates functional area The geometric mean of the building heights; Indicates functional area logarithmic standard deviation of building heights; is the height factor, which indicates the attenuation of the impact of building density on risk with height. For functional areas The decay rate parameter within.

6. A comprehensive risk assessment method for urban UAV operation according to claim 5, characterized in that: The method of weighted synthesis comprehensive risk assessment model in step 2 includes: The three risk models are integrated to construct the final comprehensive risk assessment model: ； in, , , is the risk weight; , , They represent the normalized direct death risk value, indirect death risk value, and property loss risk value respectively: 。 7. The comprehensive risk assessment method for urban drone operation according to claim 1 is characterized in that: Step 3 includes: Step 31: Combining the risk model in step 2 with specific drone parameters and scene information, using ArcPy tools, add the independent risk values ​​and comprehensive risk values ​​of the fishing net points corresponding to each time period into the attribute table; Step 32: In ArcGIS Pro, visualize the different risks by setting symbol systems; Step 33: By defining the query tool, the corresponding risk distribution is displayed according to different altitude layers; query conditions are set for each time period and altitude layer to filter out the corresponding fishing net point data; and a city low-altitude risk situation map with spatiotemporal characteristics is generated by time period, region, and altitude.

8. A comprehensive risk assessment system for urban drone operations, characterized in that: include: The 3D dynamic scene model construction module is used to use urban GIS data and fishing net segmentation to build a 3D dynamic scene model that integrates building topology, available low-altitude airspace, and population flow characteristics; A comprehensive risk assessment model building module is used to establish direct mortality risk, indirect mortality risk and property loss risk models based on drone operation parameters and time-varying population distribution and building data, and to synthesize a comprehensive risk assessment model by weighting based on normalized risk values; The urban low-altitude risk situation map output module is used to generate urban low-altitude risk situation maps based on time-space risk heat maps, and output urban low-altitude risk situation maps by time period, region, and altitude.

Images