Search

CN-122020030-A - Method and equipment for dynamically scheduling low-altitude airspace based on multi-source heterogeneous data

CN122020030ACN 122020030 ACN122020030 ACN 122020030ACN-122020030-A

Abstract

The invention discloses a method and equipment for dynamically scheduling a low-altitude airspace based on multi-source heterogeneous data, and relates to the fields of low-altitude airspace management technology, artificial intelligence and multi-source data fusion. The method comprises the steps of constructing a four-dimensional space-time grid containing space and time based on Beidou grid codes, low-altitude airspace geographic boundaries and UTC time references, aligning original data based on the four-dimensional space-time grid, inputting the aligned original data into a fusion model, extracting to obtain multi-dimensional space-time features, and generating an optimal flight path and a scheduling instruction based on the space features, the time features and environmental features. The invention systematically solves the core problems of low resource utilization rate, large data fusion error, inaccurate conflict detection, insufficient scene adaptation capability, low target recognition precision, unbalanced transmission safety and real-time performance and the like in the traditional low-altitude airspace management.

Inventors

  • BAI YUFENG
  • WAN FEN
  • XIONG FEI

Assignees

  • 湖北邮电规划设计有限公司

Dates

Publication Date
20260512
Application Date
20260119

Claims (10)

  1. 1. The method for dynamically scheduling the low-altitude airspace based on the multi-source heterogeneous data is characterized by comprising the following steps of: generating a first time-space grid sequence based on the Beidou grid code, the low-altitude airspace geographic boundary and the UTC time reference by dividing a multi-level space grid, constructing a time window and formulating a grid coding rule, wherein the first time-space grid sequence comprises a plurality of first time-space grids formed by ordering first grid identifiers; based on the first space-time grid sequence, the real-time position data of the aircraft and the identity data, obtaining a second space-time grid sequence with self-adaptive flow density through flow density monitoring, dynamic threshold calculation, grid segmentation and merging and boundary smoothing, wherein the second space-time grid sequence comprises a plurality of second space-time grids formed by sorting the second grid identifications; Carrying out space-time alignment on the multi-source original data based on the second space-time grid to obtain space-time aligned data, inputting the space-time aligned data into a fusion model, and obtaining multi-dimensional space-time feature vectors, wherein the multi-dimensional space-time features comprise space features, time features and environment features; and generating an optimal flight path and a scheduling instruction based on the spatial characteristics, the temporal characteristics and the environmental characteristics.
  2. 2. The method of claim 1, wherein generating a first time-space grid sequence based on the Beidou grid code, the low-space domain geographic boundary and the UTC time reference by dividing a multi-level space grid, constructing a time window and formulating a grid coding rule, the first time-space grid sequence comprising a plurality of first time-space grids ordered by first grid identifiers comprises: Dividing a space into a plurality of levels of space grids with space identifiers based on Beidou grid codes and low-altitude airspace geographic boundaries, wherein the space identifiers comprise space codes, longitude and latitude boundaries, altitude boundaries, grid center coordinates, single grid areas and single grid volumes; Based on UTC time reference data, establishing a time window sequence with a space domain state and time dynamic association relation, wherein the time window sequence comprises a plurality of time windows with window identifications in time continuity; And combining the space identifier, the time window and the UTC timestamp to form a first grid identifier, and obtaining a first time-space grid.
  3. 3. The method of claim 1, wherein the obtaining a second space-time grid sequence with adaptive traffic density based on the first space-time grid sequence, the real-time location data of the aircraft, and the identification data by traffic density monitoring, dynamic threshold calculation, grid segmentation and merging, and boundary smoothing, the second space-time grid sequence comprising a plurality of second space-time grids ordered by second grid identification comprises: Periodically acquiring real-time position data of the aircraft and identity data of the aircraft, screening legal aircraft by the identity data, counting the number of the legal aircraft in each first time-space grid, and determining the real-time flow density of the legal aircraft according to the number of the legal aircraft and the corresponding first time-space grid; Establishing dynamic association between the flow density and the threshold according to the flow density of the previous acquisition period and the initialization threshold to obtain a dynamic threshold, wherein the dynamic threshold is a real-time threshold for determining legal aircrafts in a unit area; establishing a grid granularity logic association between the real-time traffic density and the dynamic threshold value and the first time-space grid to obtain a self-adaptive time-space grid; And smoothing the boundary of the self-adaptive space-time grid to form a transition region with a virtual grid identifier, an adjacent grid identifier, equivalent granularity, flow density after smoothing and boundary coordinates, and filling the self-adaptive space-time grid with the transition region to obtain a second space-time grid.
  4. 4. The method of claim 1, wherein counting the number of legitimate aircraft in each first time-space grid comprises: determining a time window to which each legal aircraft belongs according to the positioning time stamp of the legal aircraft; Retrieving and associating a first grid identification based on the longitude and latitude of the legal aircraft and the altitude of the legal aircraft in the belonging time window; And counting the number of legal aircrafts associated with the same first grid identification, and obtaining the number of legal aircrafts in each first time-space grid.
  5. 5. A method according to claim 3, characterized in that the initialization threshold is based on historical flow data of 1 year in a certain city, and the maximum value of the flow density in the period of 80% of the flat peak period obtained by the flow density distribution in the flat peak period is counted; Dynamic threshold The calculation formula is as follows: Lambda is the weighting factor, For the dynamic threshold of the previous acquisition period, The number of legal aircraft in a certain first time-space grid for the previous acquisition period.
  6. 6. The method of claim 5, wherein establishing a logical association of the real-time traffic density and the dynamic threshold with the grid granularity of the first spatio-temporal grid results in an adaptive spatio-temporal grid, comprising: If the number of legal aircrafts in a certain first space-time grid is detected to be larger than a dynamic threshold value and the minimum sub-grid level of the space-time grid is not larger than 5, splitting the minimum sub-grid level of the space-time grid into 4 sub-grids; and if the number of legal aircrafts in a certain first space-time grid and the minimum sub-grid is detected to be smaller than a half dynamic threshold value and the minimum sub-grid level of the space-time grid is larger than zero, merging the minimum sub-grids of the space-time grid into the same father grid.
  7. 7. The method of claim 1, wherein the spatiotemporal alignment of the multi-source raw data based on the second spatiotemporal grid to obtain spatiotemporal alignment data, inputting the spatiotemporal alignment data into the fusion model, and wherein the multi-dimensional spatiotemporal features comprise spatial features, temporal features, and environmental features, comprising: ext> collectingext> lowext> -ext> orbitext> satelliteext> remoteext> sensingext> dataext>,ext> Dopplerext> radarext> detectionext> dataext>,ext> 5ext> Gext> -ext> Aext> baseext> stationext> generalext> senseext> dataext>,ext> weatherext> stationext> monitoringext> dataext> andext> aircraftext> airborneext> sensorext> dataext>,ext> andext> carryingext> outext> standardizationext>,ext> anomalyext> removalext> andext> deletionext> compensationext> toext> obtainext> multiext> -ext> sourceext> originalext> dataext>;ext> Based on the second space-time grid, performing space-time alignment on the multi-source original data to obtain space-time aligned data; Constructing a fusion model; And inputting the space-time alignment data into a fusion model, and obtaining multidimensional space-time feature vectors.
  8. 8. The method of claim 1, wherein the fusion model comprises an input layer, STGNN layers, LSTM layers, fusion layers, and an output layer; The input layer is used for inputting space-time alignment data; The STGNN layers are constructed based on the graph adjacency matrix and used for extracting spatial and local time sequence characteristics; The LSTM layer is used for extracting the long-time sequence characteristics; The fusion layer is used for adaptively distributing the weights of the space characteristics and the time sequence characteristics based on the attention mechanism weighting fusion calculation layer; the output layer is used for outputting the airspace state characteristic vector.
  9. 9. The method of claim 1, wherein generating optimal flight path and scheduling instructions based on the spatial features, the temporal features, and the environmental features comprises: establishing conflict levels and early warning operation rules associated with the conflict levels according to the spatial characteristics and the time characteristics; Determining a dynamic safety interval of the aircraft according to the environmental characteristics and the spatial characteristics; constructing a multi-objective optimization function based on the time features and the space features, taking the dynamic safety interval as a constraint condition of the multi-objective optimization function, and determining an optimal flight path according to the optimization function value; And generating a scheduling instruction according to the optimal flight path, wherein the scheduling instruction comprises an instruction for controlling the aircraft to execute corresponding early warning operation based on the conflict level and the early warning operation rule.
  10. 10. Electronic equipment for dynamically scheduling low-altitude airspace based on multi-source heterogeneous data, which is characterized by comprising at least one processor, at least one memory and a bus, wherein the at least one memory and the bus are connected with the processor, the processor and the memory are used for completing mutual communication through the bus, and the processor is used for calling program instructions in the memory to execute the method for dynamically scheduling the low-altitude airspace based on the multi-source heterogeneous data according to any one of claims 1 to 9.

Description

Method and equipment for dynamically scheduling low-altitude airspace based on multi-source heterogeneous data Technical Field The invention relates to the field of low-altitude airspace management technology, artificial intelligence and multi-source data fusion, in particular to a method and equipment for dynamically scheduling a low-altitude airspace based on multi-source heterogeneous data. The method is suitable for low-altitude economy typical application scenes such as urban low-altitude logistics, emergency medical rescue, border monitoring, agricultural plant protection and the like, and achieves airspace utilization rate improvement, flight safety guarantee and multi-scene flexible adaptation through fine dynamic division and intelligent scheduling of low-altitude airspace resources. Background With the rapid development of low-altitude economy, the low-altitude flight demands of scenes such as urban low-altitude logistics, emergency medical rescue, border monitoring, agricultural plant protection and the like are increasing day by day, the number of low-altitude aircrafts (aircrafts, eVTOL and the like) is greatly increased, and extremely high requirements are provided for the fine management and intelligent scheduling of low-altitude airspace resources. The traditional static grid division mode (such as fixed 1km multiplied by 1km grid) cannot adapt to the flow difference between the peak period and the flat peak period of the logistics, so that the space domain congestion in a high flow area and the resource waste in a low flow area are caused, the space domain utilization rate is generally lower than 40%, and a dynamic adjustment mechanism based on real-time flow is lacked. Ext> theext> lowext> -ext> altitudeext> supervisionext> relatesext> toext> multiext> -ext> sourceext> heterogeneousext> dataext> suchext> asext> satelliteext> remoteext> sensingext>,ext> radarext> detectionext>,ext> 5ext> Gext> -ext> Aext> baseext> stationsext>,ext> meteorologicalext> monitoringext> andext> theext> likeext>,ext> spaceext> -ext> timeext> accurateext> alignmentext> isext> difficultext> toext> achieveext> inext> theext> priorext> artext>,ext> theext> dataext> fusionext> errorext> rateext> isext> upext> toext> 15ext>%ext>,ext> andext> thereforeext>,ext> theext> perceptionext> ofext> anext> airspaceext> stateext> isext> inaccurateext>,ext> andext> theext> reliabilityext> ofext> schedulingext> decisionsext> isext> affectedext>.ext> In addition, when the aircraft and the man-machine are in mixed operation, the existing two-dimensional collision detection model does not consider factors such as the quality and the altitude difference of the aircraft, the collision detection accuracy is less than 90%, and the safety requirement of the mixed operation cannot be met. In addition, the requirements of logistics distribution and medical rescue on space domain priority, path planning and safety interval are obviously different, special scheduling logic is required for extended scenes such as border monitoring and agricultural plant protection, the existing system lacks an intelligent algorithm aiming at the special scenes, and flexible adaptation of one system to multiple scenes cannot be realized. Therefore, a low-altitude airspace management technology capable of realizing airspace dynamic division, multi-source data accurate fusion, multi-target safe scheduling, multi-scene flexible adaptation and data safe transmission is needed to systematically solve the technical pain. Disclosure of Invention The invention provides a method for dynamically scheduling a low-altitude airspace based on multi-source heterogeneous data, which systematically solves the problems of low resource utilization rate, large data fusion error, inaccurate conflict detection, insufficient scene adaptation capability, low target recognition precision, unbalanced data transmission safety and real-time performance and the like in the traditional low-altitude airspace management technology by constructing a four-dimensional space-time grid system, designing a flow density self-adaptive dynamic grid division algorithm, constructing an air-space-ground integrated multi-source data fusion framework (comprising STGNN +LSTM fusion model), developing a multi-target optimized intelligent scheduling model, formulating a scene dynamic scheduling strategy (comprising a CNN target classification model), and constructing a light-weight efficient encryption transmission mechanism. In view of the above, the present invention provides a method for dynamically scheduling a low-altitude space based on multi-source heterogeneous data, which overcomes or at least partially solves the above problems, and the technical scheme is as follows: the method for dynamically scheduling the low-altitude airspace based on the multi-source heterogeneous data is characterized by comprising the following steps of: generating a first time-space grid sequence based on the Beidou grid cod