Method and device for predicting positioning deviation of terminal logistics UAV in urban low-altitude scenes
By constructing a route deviation database and a spatio-temporal geographic weighted regression model, predicting the route deviation of urban low-altitude drones is solved, and the accuracy and efficiency of route planning are improved.
Patent Information
- Application Number
- CN202410605733.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-15
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-05-15
AI Technical Summary
When positioning the terminal logistics drone in urban low-altitude scenarios, the positioning accuracy is not high due to the differences in shading and meteorological conditions of high-rise buildings in urban areas. It is difficult for the existing technology to effectively predict route deviations, which affects flight safety and efficiency.
By obtaining the planning and actual route data of logistics drones, dividing route segments and building historical deviation databases, extracting building feature data, and using spatiotemporal and spatial relationship model of positioning deviations is established, predicting route deviations, and optimizing route planning.
It improves the positioning accuracy and flight accuracy of drones, reduces operating costs, and meets the needs of the future urban low-altitude logistics drone market.
Smart Images

Figure CN118482725B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the research field of logistics UAV flight, and in particular to a method and device for predicting positioning deviation of terminal logistics UAV in urban low-altitude scenes. Background Art
[0002] With the development of drone technology and the advancement of the work of building a strategic emerging industry of low-altitude economy, the drone logistics industry presents huge market potential. Among them, the emergence of terminal logistics drone application scenarios provides an effective solution to the problem of the "last mile" of goods delivery in remote areas and urban scenarios with high timeliness requirements. On the one hand, most terminal logistics drones are light and small, and their operating altitude is below 120m, that is, the operating area (space) of terminal logistics drones is mostly below the city skyline. At the same time, the positioning of light and small drones often relies only on satellite positioning methods such as Beidou and GPS to achieve positioning, but due to the obstruction of high-rise buildings in the city, there are spatial differences (heterogeneity) in the accuracy of satellite positioning; on the other hand, due to the different heights and distributions (three-dimensional spatial forms) of urban buildings, there are also spatial differences in meteorological conditions such as wind direction, wind speed, and temperature in different areas of the city.
[0003] At present, in order to ensure the safety of the planned routes of UAVs, the safety of the actual operation of the planned routes of UAVs is mainly ensured by improving the positioning accuracy (such as multi-source data fusion positioning, adding RTK equipment, cooperative target perception integration, etc.) and combining the field verification flight before operation with real-time monitoring during operation. Among them, the former relies on the sensor equipment carried by the UAV itself and requires the effective layout of low-altitude meteorological stations, communication navigation and monitoring equipment, mainly for medium and large UAVs. On the one hand, this method increases the cost of infrastructure investment. At the same time, the installation of sensors on UAVs will undoubtedly reduce the endurance of UAVs, and the economy is poor; the latter can be achieved through field verification flights without increasing infrastructure investment and reducing the endurance of UAVs. It is mainly for micro-light UAVs, but considering the density of low-altitude UAV routes in cities in the later period and the exponential growth of the number of routes, on the one hand, the field verification flight before operation cannot meet the needs of the future urban low-altitude logistics UAV market, and on the other hand, real-time monitoring during operation will undoubtedly increase operating costs and the work pressure of UAV controllers.
[0004] In summary, in order to ensure the operational safety of the planned routes of terminal logistics drones in urban low-altitude scenarios, the key point is to improve the positioning accuracy of terminal logistics drones or to effectively predict the flight positioning deviations at each location of the planned route. The former is not practical for micro-light terminal logistics drones by adding airborne sensors and ground infrastructure (sensors cannot be installed or sensors can be installed but seriously affect the endurance), while the latter is limited by the fact that there are many factors that affect the flight positioning deviations. There is currently no effective way to predict the flight positioning deviations at each location of the planned route, such as the spatial differences (heterogeneity) in positioning accuracy in different areas of urban low altitude. Based on this, when using terminal logistics drones in urban low-altitude scenarios, how to effectively predict the positioning deviations of spatial differences in urban areas and improve the accuracy of flight positioning of logistics drones is a technical problem that needs to be solved urgently. Summary of the invention
[0005] The purpose of the present invention is to solve the technical problems pointed out by the background technology, and to provide a method and device for predicting the positioning deviation of terminal logistics drones in urban low-altitude scenes, establish a segmented correspondence between the planned routes of logistics drones and the actual operating routes, and realize a historical deviation database that relies on historical data to obtain segmented matching; then extract the building feature data, and use the spatiotemporal geographic weighted regression method to construct a spatiotemporal relationship model of the positioning deviation between the building feature data and the route feature segments, thereby realizing the prediction of the logistics drone route positioning deviation.
[0006] The purpose of the present invention is achieved through the following technical solutions:
[0007] A method for predicting positioning deviation of a terminal logistics UAV in a low-altitude urban scene, the method comprising:
[0008] S1. Obtain the planned route data of the logistics drone and the corresponding actual operation route data to form a route deviation data group, and perform corresponding segment matching on the route deviation data group according to the planned route and the actual operation route of the logistics drone. The segment matching is the planned route segment and the actual route segment, and the deviation data corresponding to the segment is obtained and a historical deviation database is constructed;
[0009] S2. Acquire urban building data and construct an urban building model based on it, extract building feature data from the urban building model, and extract route feature segments from the historical deviation database, where the route feature segment is a planned route segment or an actual route segment or a center route segment between the planned route segment and the actual route segment; use a spatiotemporal geographic weighted regression method to construct a spatiotemporal relationship model of positioning deviations between building feature data and route feature segments;
[0010] S3. Plan a new route for the logistics UAV and the corresponding new route data. Use the positioning deviation spatiotemporal relationship model to match the new route data according to the route feature segments. Use the historical deviation database to match and extract the deviation data corresponding to the route segments. Obtain the deviation data corresponding to each route segment of the new route data as the positioning deviation prediction result.
[0011] In order to better implement the present invention, in step S1, the segment spacing size is determined by the following method: calculate the deviation value between the planned route and the actual operating route of the logistics drone in the route deviation data group at the corresponding position and take the maximum deviation value as the segment spacing size.
[0012] Preferably, in step S2, the method for acquiring urban building data is as follows: using a drone equipped with a multi-view camera to take aerial photos of the target urban area and obtain aerial photography data, and then obtaining urban building data through data fusion.
[0013] Preferably, in step S1, the deviation data in the historical deviation database is expressed in vectors, and the historical deviation database is a three-dimensional vector set {V1, V2, ..., V n}, where V n Represents the three-dimensional vector of the nth group of matching segments, V n =[x n -x' n ,y n -y' n , z n -z' n ],(x n ,y n , z n )、(x' n , y' n , z' n ) represent the position coordinates of the planned route segment and the actual route segment in the nth group of matching segments respectively.
[0014] Preferably, the building characteristic data include characteristic parameters of CV, B, Sh, Sk, K, FP, MP, FPA, MPA, and Tg.
[0015] Preferably, the route feature segment is a planned route segment, the number of the building feature data is P, and the spatiotemporal relationship expression between the building feature data and the route feature segment in the positioning deviation spatiotemporal relationship model is as follows:
[0016]
[0017] Among them, V n Represents the three-dimensional vector of the nth group of matching segments, V n =[x n -x'n ,y n -y' n , z n -z' n ],(x n ,y n , z n )、(x' n , y' n , z' n ) represent the position coordinates of the planned route segment and the actual route segment in the nth group of matching segments, respectively, (x n ,y n , z n ) is the route characteristic segment; β0(x n ,y n , z n ) represents the intercept of the route characteristic segment; U nk represents the data corresponding to the k-th building feature and the n-th route feature segment, 1≤k≤P; β k (x n ,y n , z n ) represents the fitting coefficient of the kth building feature and the nth route feature segment; ε n Represents random error.
[0018] Preferably, the present invention also includes the following method:
[0019] S4. Plan the starting point position coordinates and the end point position coordinates of the proposed route, traverse the historical deviation database and extract the route segment combination with the lowest deviation data, the route segment combination is a continuous route segment combination from the starting point position coordinates to the end point position coordinates, and use the route segment combination as the proposed route.
[0020] A positioning deviation prediction device for terminal logistics UAV in urban low-altitude scenes, comprising:
[0021] at least one processor;
[0022] at least one memory for storing at least one program;
[0023] When at least one program is executed by at least one processor, the at least one processor implements the method for predicting positioning deviation of terminal logistics drones in urban low-altitude scenarios of the present invention.
[0024] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0025] (1) The present invention groups the route deviation data into segments for matching, establishes the segmentation correspondence between the planned route and the actual operating route of the logistics UAV, and realizes the construction of a historical deviation database for segmented matching based on historical data; then extracts the building feature data, and uses the spatiotemporal geographic weighted regression method to construct a spatiotemporal relationship model of the positioning deviation between the building feature data and the route feature segments, thereby realizing the prediction of the logistics UAV route positioning deviation.
[0026] (2) The present invention constructs a positioning deviation spatiotemporal relationship model, which can extract segmented route segments of newly planned route data, and uses the positioning deviation spatiotemporal relationship model to match the newly planned route data according to the route feature segments. The deviation data corresponding to the route segments are extracted using the historical deviation database to obtain the deviation data corresponding to each route segment of the newly planned route data as the positioning deviation prediction result, which has the advantages of high prediction efficiency and high prediction accuracy.
[0027] (3) The present invention can first plan the starting point position coordinates and the end point position coordinates of the planned route, then traverse the historical deviation database and extract the route segment combination with the lowest deviation data, and then combine and plan the optimal planned route with the highest positioning accuracy, thereby improving the positioning accuracy of the logistics drone flight and meeting the needs of the rapid development of new low-altitude economic application scenarios of logistics drones in the future. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION
[0029] The present invention is further described in detail below in conjunction with embodiments:
[0030] Example
[0031] like Figure 1 As shown, a method for predicting positioning deviation of a terminal logistics UAV in a low-altitude urban scene includes:
[0032] S1. Obtain the planned route data of the logistics UAV and the corresponding actual operation route data to form a route deviation data group, and perform corresponding segment matching on the route deviation data group according to the planned route and the actual operation route of the logistics UAV. The segment matching is the planned route segment and the actual route segment. The planned route segment and the actual route segment form a matching segment group, obtain the deviation data corresponding to the segment and construct a historical deviation database. The historical deviation database stores the deviation data according to the segment groups (including the planned route segments and the actual route segments).
[0033] In some embodiments, the segment spacing size of this embodiment is obtained by the following method: calculate the deviation value between the planned route and the actual operating route of the logistics drone in the route deviation data group at the corresponding position and take the maximum deviation value as the segment spacing size.
[0034] In some embodiments, the deviation data in the historical deviation database is expressed in vectors, and the historical deviation database is a three-dimensional vector set {V1, V2, ..., V n}, where V n Represents the three-dimensional vector of the nth group of matching segments, V n =[x n -x' n ,y n -y' n , z n -z' n ],(x n ,y n , z n )、(x' n , y' n , z' n ) represent the position coordinates of the planned route segment and the actual route segment in the nth group of matching segments. If the maximum deviation value between the planned route and the actual route of the logistics drone in the route deviation data group is taken, the segment spacing size d = max{|x n -x' n |,|y n -y' n |,|z n -z' n |}.
[0035] S2. Obtain urban building data and use it to build an urban building model, extract building feature data from the urban building model, and extract route feature segments from the historical deviation database. The route feature segment is the planned route segment or the actual route segment or the center route segment between the planned route segment and the actual route segment. Use the spatiotemporal geographic weighted regression (GTWR) method to build a spatiotemporal relationship model of positioning deviation between building feature data and route feature segments.
[0036] In some embodiments, the method for acquiring urban building data is as follows: using a drone equipped with a multi-view camera to take aerial photos of the target urban area and obtain aerial photography data, and then obtaining urban building data through data fusion. The specific method is exemplified as follows: using a drone equipped with a multi-view camera, taking aerial photos of urban buildings included in the historical routes of the terminal logistics drone and using Context Capture software to perform oblique photography modeling to obtain an urban oblique photography model, the urban oblique photography model includes urban building location data and elevation data (DEM data), the location data and elevation data of the urban oblique photography model are imported into software such as SuperMap iServer, SVSModeler or OSketch for individual operations, and a file (generally in .CSV format data) of information such as the longitude and latitude of surrounding buildings at n equally spaced locations, the total height of the floors, the number of floors, and the DEM height value is obtained, thereby obtaining an urban building model.
[0037] Extract building characteristic data from the urban building model. In some embodiments, the building characteristic data of the present invention include characteristic parameters of CV, B, Sh, Sk, K, FP, MP, FPA, MPA, and Tg (i.e., characteristic parameters of coefficient of variation CV, burst information B, information entropy Sh, skewness coefficient Sk, kurtosis coefficient K, first peak FP, maximum peak MP, first peak area FPA, maximum peak area MPA, and transition point Tg). Extract building feature data from the urban building model (for example, extract feature data from a CSV format data file). The building feature data of the present invention mainly includes CV, B, Sh, Sk, K, FP, MP, FPA, MPA, and Tg, a total of ten building features, to achieve building feature description. See Qingfeng Guan, Yao Yao, Teng Ma, Ye Hong, Yongpan Bie & Jianjun Lyu, "Under the Dome: A 3D Urban Texture Model and Its Relationship with Urban Land Surface Temperature" published in Annals of the American Association of Geographers, DOI: 10.1080 / 24694452.2021.1972790. In this embodiment, UrbanTexture v1.0 software can be used to extract building feature data. CV (representing coefficient of variation) is the full name of Coeffcient of variation, which reflects the absolute value of the degree of data dispersion. B (representing burst information) is the full name of Burst statistics, which can be used to measure the burstiness of spatial distribution. Sh (information entropy) is the full name of Shannon entropy (Sh), which can be used to measure the uncertainty or uncertainty of the system. Sk (skewness coefficient) is the full name of Skewness (Sk), which can describe the symmetry of the distribution deviation. K (kurtosis coefficient) is the full name of Kurtosis (k), which describes the degree of suddenness. FP (first peak) is the full name of The first peak (FP), which is the peak feature at the first peak. MP (maximum peak) is the full name of The maximum peak (MP), which is the peak feature at the maximum peak. FPA (first peak area) is the full name of The first peak area (FPA), which is the area from the origin to the first peak position. MPA (maximum peak area) is the full name of The maximum peak area (MPA), which is the integral from the origin to the maximum peak position.The full name of Tg (transition point) is the ratio from the first peak to the first valley. The building characteristic data corresponding to the route characteristic segment n adopts {CV. n , B n , Sh n , Sk n , K n , FP n , M.P. n , FPA n , MPA n , Tg n}Express.
[0038] The spatial-temporal relationship model of the positioning deviation of the present invention can be constructed by using ArcGIS 10.X software (combined with GTWR Beta 1.0 script), and the spatial-temporal relationship model of the positioning deviation stores data of the spatial-temporal geographic weighted regression method. n ,y n , z n ) is given as follows: The number of building feature data is P (in this embodiment, CV n , B n , Sh n , Sk n , K n , FP n , M.P. n , FPA n , MPA n , Tg n A total of 10), the spatiotemporal relationship expression between the building feature data and the route feature segment in the positioning deviation spatiotemporal relationship model is as follows:
[0039]
[0040] Among them, V n Represents the three-dimensional vector of the nth group of matching segments, V n =[x n -x' n ,y n -y' n , z n -z' n ],(x n ,y n , z n )、(x' n , y' n , z' n ) represent the position coordinates of the planned route segment and the actual route segment in the nth group of matching segments, respectively, (x n ,y n , zn ) is the route characteristic segment. n ,y n , z n ) represents the intercept of the route characteristic segment. nk Indicates the data corresponding to the k-th building feature and the n-th route feature segment, 1≤k≤P. β k (x n ,y n , z n ) represents the fitting coefficient of the kth building feature and the nth route feature segment. ε n Represents random error.
[0041] For route characteristic segment n, β k (x n ,y n , z n ) can be averaged To represent, that is, all historical data of route feature segment n corresponds to β k (x n ,y n , z n )’s average value.
[0042]
[0043] Generally, the route feature segment will have a time series, (x n ,y n , z n , t n ) represents the latitude, longitude, elevation and time of route feature segment n; β k (x n ,y n , z n , t n ) represents the kth building feature and the nth route feature segment at time t n The fitting coefficient of ; taking into account the time series, the spatiotemporal relationship expression is as follows:
[0044]
[0045] S3. Plan a new route for the logistics UAV and the corresponding new route data. Use the positioning deviation spatiotemporal relationship model to match the new route data according to the route feature segments. Use the historical deviation database to match and extract the deviation data corresponding to the route segments. Obtain the deviation data corresponding to each route segment of the new route data as the positioning deviation prediction result.
[0046] In some embodiments, the present invention further comprises the following method:
[0047] S4. Plan the starting point position coordinates and the end point position coordinates of the proposed route, traverse the historical deviation database and extract the route segment combination with the lowest deviation data, the route segment combination is a continuous route segment combination from the starting point position coordinates to the end point position coordinates, and use the route segment combination as the proposed route.
[0048] A device for predicting positioning deviation of a UAV for terminal logistics in a low-altitude urban scene comprises at least one processor and at least one memory. The memory is used to store at least one program; when the at least one program is executed by the at least one processor, the at least one processor implements the method for predicting positioning deviation of a UAV for terminal logistics in a low-altitude urban scene of this embodiment.
[0049] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for predicting positioning deviation of terminal logistics drones in urban low-altitude scenarios, characterized by: The methods include: S1. Obtain the planned route data of the logistics drone and the corresponding actual operation route data to form a route deviation data group, and perform corresponding segment matching on the route deviation data group according to the planned route and the actual operation route of the logistics drone. The segment matching is the planned route segment and the actual route segment, and the deviation data corresponding to the segment is obtained and a historical deviation database is constructed; S2. Obtain urban building data and use it to construct an urban building model, extract building feature data from the urban building model, and extract route feature segments from the historical deviation database. The route feature segment is a planned route segment or an actual route segment or a central route segment between the planned route segment and the actual route segment; use the spatiotemporal geographic weighted regression method to construct a positioning deviation spatiotemporal relationship model between the building feature data and the route feature segment; the route feature segment selects the planned route segment, the number of building feature data is P, and the spatiotemporal relationship expression between the building feature data and the route feature segment in the positioning deviation spatiotemporal relationship model is as follows: Among them, V n Represents the three-dimensional vector of the nth group of matching segments, V n =[x n -x' n ,y n -y' n , z n -z' n ],(x n ,y n , z n )、(x' n , y' n , z' n ) represent the position coordinates of the planned route segment and the actual route segment in the nth group of matching segments, respectively, (x n ,y n , z n ) is the route characteristic segment; β0(x n ,y n , z n ) represents the intercept of the route characteristic segment; U nk represents the data corresponding to the k-th building feature and the n-th route feature segment, 1≤k≤P; β k (x n ,y n , z n ) represents the fitting coefficient of the kth building feature and the nth route feature segment; ε n represents random error; S3. Plan a new route for the logistics UAV and the corresponding new route data. Use the positioning deviation spatiotemporal relationship model to match the new route data according to the route feature segments. Use the historical deviation database to match and extract the deviation data corresponding to the route segments. Obtain the deviation data corresponding to each route segment of the new route data as the positioning deviation prediction result.
2. The method for predicting positioning deviation of terminal logistics UAV in urban low-altitude scene according to claim 1 is characterized by: In step S1, the segment spacing size is determined by the following method: calculate the deviation value between the planned route of the logistics drone and the actual operating route in the route deviation data group at the corresponding position and take the maximum deviation value as the segment spacing size.
3. The method for predicting positioning deviation of terminal logistics drones in urban low-altitude scenarios according to claim 1 is characterized by: In step S2, the method for acquiring urban building data is as follows: using a drone equipped with a multi-view camera to take aerial photos of the target urban area and obtain aerial photography data, and then obtaining urban building data through data fusion.
4. The method for predicting positioning deviation of a UAV for terminal logistics in an urban low-altitude scene according to claim 1 is characterized in that: In step S1, the deviation data in the historical deviation database is expressed in vectors. The historical deviation database is a three-dimensional vector set {V1, V2, ..., V n }, where V n Represents the three-dimensional vector of the nth group of matching segments, V n =[x n -x' n ,y n -y' n , z n -z' n ],(x n ,y n , z n )、(x' n , y' n , z' n ) represent the position coordinates of the planned route segment and the actual route segment in the nth group of matching segments respectively.
5. The method for predicting positioning deviation of terminal logistics UAV in urban low-altitude scene according to claim 1 is characterized by: The building characteristic data include characteristic parameters of CV, B, Sh, Sk, K, FP, MP, FPA, MPA, and Tg, wherein CV reflects the absolute value of the degree of data dispersion, B is used to measure the suddenness of spatial distribution, Sh is used to measure the uncertainty of the system, Sk is used to describe the symmetry of the distribution deviation, K is used to describe the degree of suddenness, FP represents the peak feature at the first peak, MP represents the peak feature at the maximum peak, FPA represents the area from the origin to the first peak position, MPA represents the integral from the origin to the maximum peak position, and Tg represents the ratio from the first peak to the first valley.
6. The method for predicting positioning deviation of a UAV for terminal logistics in an urban low-altitude scene according to claim 1 is characterized by: Also includes the following methods: S4. Plan the starting point position coordinates and the end point position coordinates of the proposed route, traverse the historical deviation database and extract the route segment combination with the lowest deviation data, the route segment combination is a continuous route segment combination from the starting point position coordinates to the end point position coordinates, and use the route segment combination as the proposed route.
7. A positioning deviation prediction device for terminal logistics drones in urban low-altitude scenarios, characterized by: include: at least one processor; at least one memory for storing at least one program; When at least one program is executed by at least one processor, the at least one processor implements the method according to any one of claims 1 to 6.
