Global enroute conflict prediction method based on spatial conflict matrix

Abstract

This invention discloses a global aeronautical chart route conflict prediction method based on an airspace conflict matrix, specifically relating to the field of air traffic management technology. It addresses the problem of systematic deviations in conflict probability prediction values ​​caused by existing route conflict prediction methods failing to distinguish the heterogeneity of flight spatial distribution within the same airspace grid during transitional areas of route rules. The method involves acquiring flight plans and track data, dividing the airspace into cell grids, extracting historical track points of flights before they enter a cell grid, and calculating the fractal dimension to determine the flight's route operation status. Based on whether the difference between the fractal dimension and a preset dimension threshold falls within a preset transition interval, the occupancy probability is calculated using a normal distribution, uniform distribution, or mixed probability distribution. A spatiotemporal conflict matrix is ​​constructed based on the occupancy probability, and cell grids with conflict probabilities exceeding a preset probability threshold and their corresponding time periods are identified as route conflict prediction results.

Description

Technical Field

[0001] This invention relates to the field of air traffic management technology, and more specifically, to a method for predicting global aeronautical chart route conflicts based on an airspace conflict matrix. Background Technology

[0002] Within the airspace covered by global aeronautical charts, air traffic management systems typically predict route conflicts by gridding the airspace and constructing a conflict matrix. Existing methods acquire flight plans and track data from global aeronautical charts, project aircraft four-dimensional tracks onto a uniformly divided airspace grid, and calculate the probability that each grid is occupied by multiple aircraft at different times, forming an airspace conflict matrix. Based on this matrix, the system determines which grids have potential conflicts, thus providing a basis for adjusting flight routes or altitude levels. This approach assumes that all flights within the same airspace grid follow the same spatial distribution pattern, such as a normal distribution along a fixed route centerline or a uniform distribution within the grid, and uses this distribution as a unified benchmark for conflict probability calculation.

[0003] However, in the actual operation of global aeronautical charts, different regions may adopt different route operation modes at different times. During the transition period of route rule switching within the same airspace grid, flights following different operation rules may coexist. For example, some flights fly along fixed routes, while others fly along free routes. The spatial distribution characteristics of these two types of flights within the same grid are fundamentally different. Existing conflict matrix construction methods fail to distinguish the heterogeneity of the spatial distribution of flights within the same grid and still use a uniform probability distribution assumption for conflict calculation. This leads to a systematic deviation in the predicted conflict probability values ​​of the conflict matrix in the route rule transition area, reducing the reliability of global aeronautical chart route conflict prediction under conditions of multiple route modes coexisting. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art, this invention provides a global aeronautical chart route conflict prediction method based on an airspace conflict matrix to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: The global aeronautical chart route conflict prediction method based on the airspace conflict matrix includes the following steps: S1: Obtain flight plans and track data of each flight in the global aeronautical chart airspace within a preset time window, and divide the airspace into cell grids; S2: For any cell grid that any flight passes through, extract the historical flight path points before the flight enters the cell grid and calculate the fractal dimension of the historical flight path points. When the fractal dimension is lower than the preset dimension threshold, it is determined that the flight is in a fixed route operation state within the cell grid; otherwise, it is determined that the flight is in a free route operation state. S3: Calculate the difference between the fractal dimension and the preset dimension threshold. When the difference does not fall into the preset transition interval, the occupancy probability of flights operating on fixed routes is calculated using a normal distribution along the centerline of the fixed route, and the occupancy probability of flights operating on free routes is calculated using a uniform distribution within the cell grid. When the difference falls into the preset transition interval, the occupancy probability is calculated by combining the normal distribution and the uniform distribution into a mixed probability distribution based on the difference using a mixing coefficient. S4: Construct a spatiotemporal conflict matrix based on the occupancy probability of each flight in each cell grid at each time period; S5: The cell grids in the spatiotemporal conflict matrix whose conflict probability exceeds the preset probability threshold and their corresponding time periods are determined as the route conflict prediction results.

[0006] Furthermore, the flight plans and tracks of each flight within the global aeronautical chart airspace are acquired within a preset time window, and the airspace is divided into a grid of cells, including: The flight plans of all flights entering the global chart airspace within a preset time window are extracted from the flight plan database of the air traffic management system. The track data corresponding to the flight plans of all flights are extracted from the track database of the air traffic management system. The start point of the preset time window is the current time, and the end point of the preset time window is the current time plus a preset duration. The size of the cell grid along the direction perpendicular to the fixed route centerline is determined by the average spacing of the fixed route centerline in the global chart airspace, and the size of the cell grid along the fixed route centerline is determined by the average distance between adjacent waypoints on the fixed route centerline.

[0007] Furthermore, the preset duration of the preset time window is determined based on the average flight time of all flights flying over the unit grid covered by the global aeronautical chart airspace, and the preset duration is an integer multiple of the average flight time.

[0008] Furthermore, for any cell grid passed by any flight, the historical flight path points before the flight entered the cell grid are extracted and the fractal dimension of the historical flight path points is calculated. When the fractal dimension is lower than a preset dimension threshold, the flight is determined to be in a fixed-path operation state within the cell grid; otherwise, it is determined to be in a free-path operation state, including: All flight points within a preset flight duration before a flight enters a cell grid are extracted from the flight data as historical flight points. The plane containing the historical flight points is divided into a coverage grid with proportionally increasing side lengths using box counting. The number of historical flight points occupying each coverage grid under each side length is counted. The absolute value of the slope of the straight line fitted by the logarithm of the coverage grid side length and the logarithm of the corresponding number of occupied points is used as the fractal dimension. The fractal dimension is compared with a preset dimension threshold. When the fractal dimension is lower than the preset dimension threshold, the flight is determined to be in a fixed-path operation state within the cell grid. When the fractal dimension is not lower than the preset dimension threshold, the flight is determined to be in a free-path operation state within the cell grid.

[0009] Furthermore, a logarithmic coordinate system is constructed with the logarithm of the covered grid side length as the x-axis and the logarithm of the corresponding occupied quantity as the y-axis. The least squares method is used to perform linear fitting on the scattered points to obtain the fitted line, and the absolute value of the slope of the fitted line is taken as the fractal dimension.

[0010] Furthermore, the difference between the fractal dimension and a preset dimension threshold is calculated. When the difference does not fall within a preset transition interval, the occupancy probability of flights operating on fixed routes is calculated using a normal distribution along the centerline of the fixed route, while the occupancy probability of flights operating on free routes is calculated using a uniform distribution within the cell grid. When the difference falls within a preset transition interval, a mixing coefficient based on the difference is used to combine the normal distribution and the uniform distribution into a mixed probability distribution to calculate the occupancy probability, including: The absolute value of the difference between the fractal dimension and the preset dimension threshold is used as the state difference. The state difference is compared with half the width of the preset transition interval. When the state difference is greater than half the width of the preset transition interval, the difference does not fall into the preset transition interval. When the state difference is less than or equal to half the width of the preset transition interval, the difference falls into the preset transition interval. When the difference falls into the preset transition interval, the ratio of the state difference to half the width of the preset transition interval is used as the mixing coefficient. The probability distribution obtained by weighting the normal distribution along the fixed flight path centerline with the mixing coefficient as the weight and the uniform distribution within the cell grid with the complement of the mixing coefficient as the weight is used as the mixed probability distribution. The occupancy probability of the flight within the cell grid is calculated using the mixed probability distribution.

[0011] Furthermore, the mean of the lateral deviations of the historical flight path points before entering the cell grid relative to the centerline of the fixed route are used as the mean of the normal distribution, and the standard deviation of the lateral deviations is used as the standard deviation of the normal distribution to construct a normal distribution along the centerline of the fixed route. When the standard deviation is lower than the preset lower limit of the standard deviation, the preset lower limit of the standard deviation is used to replace the standard deviation for construction.

[0012] Furthermore, based on the occupancy probability of each flight in each cell grid at each time period, a spatiotemporal conflict matrix is ​​constructed, including: For each time period within a preset time window, the occupancy probabilities of all flights within the same time period of the same unit grid are multiplied in pairs. The sum of all pairwise multiplication results is taken as the conflict probability of the unit grid in that time period. A spatiotemporal conflict matrix is ​​constructed with each unit grid as the row, each time period as the column, and each conflict probability as the element value.

[0013] Furthermore, the preset time window is divided into consecutive and non-overlapping time periods according to the preset number of time periods, with each time period having an equal duration. The duration of each time period is the total duration of the preset time window divided by the preset number of time periods.

[0014] Furthermore, the cell grids in the spatiotemporal conflict matrix whose conflict probability exceeds a preset probability threshold and their corresponding time periods are identified as the route conflict prediction results, including: The conflict probability of each cell in the spatiotemporal conflict matrix is ​​traversed for each time period. The conflict probability is compared with a preset probability threshold. Cells with conflict probabilities exceeding the preset probability threshold and their corresponding time periods are extracted. The extracted cell and corresponding time periods are used as the route conflict prediction results.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. By extracting the fractal dimension of flight track points, the flight's route operation status within a cell grid is identified. The fractal dimension quantitatively reflects the degree to which a flight is constrained by a fixed route, based on the clustering and dispersion characteristics of the spatial distribution of track points. Under fixed route operation status, track points exhibit a linear clustering distribution along the fixed route centerline, with the fractal dimension approaching a smaller value. Under free route operation status, track points exhibit a planar dispersion distribution within the cell grid, with the fractal dimension approaching a larger value. This allows for objective differentiation of operation status based solely on the spatial geometric characteristics of the actual flight track, without relying on flight plan declaration information. Based on this, the occupancy probability is calculated using a normal distribution along the fixed route centerline for flights determined to be under fixed route operation status, and using a uniform distribution within the cell grid for flights determined to be under free route operation status. These two probability distributions match the actual spatial distribution patterns of the two types of operation statuses, eliminating the systematic bias introduced by existing methods in the transition area of ​​route rules due to the assumption of a uniform probability distribution.

[0016] 2. By introducing a preset transition interval and a mixed probability distribution in the route operation status determination process, the problem of ambiguous intervals in flight operation status during the route rule transition period is further solved. When the difference between the fractal dimension and the preset dimension threshold falls within the preset transition interval, a mixing coefficient is constructed based on the difference. The normal distribution and the uniform distribution are weighted and combined into a mixed probability distribution to calculate the occupancy probability. The mixing coefficient changes continuously with the difference, and the shape of the mixed probability distribution smoothly transitions between the normal distribution and the uniform distribution. This makes the predicted conflict probability value in the transition region of the conflict matrix change from a binary jump to a continuous gradual change, which conforms to the real physical process of the flight spatial distribution gradually changing from clustering to dispersion during the route rule switch. The route conflict prediction results output by the constructed spatiotemporal conflict matrix can reflect the actual conflict situation under the condition of coexistence of multiple route modes in the global aeronautical chart airspace, providing a more reliable basis for air traffic management to predict route conflicts. Attached Figure Description

[0017] Figure 1 This is a flowchart of the global aeronautical chart route conflict prediction method based on the airspace conflict matrix of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Example: Figure 1 The present invention provides a global aeronautical chart route conflict prediction method based on an airspace conflict matrix, comprising the following steps: S1: Obtain flight plans and track data of each flight in the global aeronautical chart airspace within a preset time window, and divide the airspace into cell grids; S2: For any cell grid that any flight passes through, extract the historical flight path points before the flight enters the cell grid and calculate the fractal dimension of the historical flight path points. When the fractal dimension is lower than the preset dimension threshold, it is determined that the flight is in a fixed route operation state within the cell grid; otherwise, it is determined that the flight is in a free route operation state. S3: Calculate the difference between the fractal dimension and the preset dimension threshold. When the difference does not fall into the preset transition interval, the occupancy probability of flights operating on fixed routes is calculated using a normal distribution along the centerline of the fixed route, and the occupancy probability of flights operating on free routes is calculated using a uniform distribution within the cell grid. When the difference falls into the preset transition interval, the occupancy probability is calculated by combining the normal distribution and the uniform distribution into a mixed probability distribution based on the difference using a mixing coefficient. S4: Construct a spatiotemporal conflict matrix based on the occupancy probability of each flight in each cell grid at each time period; S5: The cell grids in the spatiotemporal conflict matrix whose conflict probability exceeds the preset probability threshold and their corresponding time periods are determined as the route conflict prediction results.

[0020] The flight plans and track data of each flight in the global aeronautical chart airspace within a preset time window are obtained, and the airspace is divided into cell grids. The specific implementation method is as follows.

[0021] The air traffic management system (Air Traffic Management System) has a flight plan database and a track database. The flight plan database stores the flight plans of all flights entering the global chart airspace, while the track database stores the track data corresponding to each flight's flight plan. A preset time window is a time interval starting from the current time and ending at the current time plus a preset duration. For example, if the current time is 12:00:00 UTC and the preset duration is 60 minutes, then the preset time window is the time interval from 12:00:00 UTC to 13:00:00 UTC. At the start time of the preset time window, a query request is sent to the flight plan database to retrieve the flight plans of all flights entering the global chart airspace within the preset time window. Each flight plan includes a flight identifier, the planned entry time into the global chart airspace, the planned departure time from the global chart airspace, the planned waypoint sequence, and the planned flight altitude. Based on the flight identifiers in the retrieved flight plans of all flights, a query request is sent to the track database to retrieve the track data corresponding to the flight identifiers in the flight plans of all flights. Each track data entry contains a flight identifier and a sequence of track points arranged in chronological order. Each track point contains its latitude and longitude, flight altitude, and timestamp.

[0022] The preset duration is determined as follows: The actual flight time of all flights from entering to leaving the global aeronautical chart airspace within the preset statistical days is statistically analyzed, and the average of all actual flight times is calculated as the average flight traverse duration. The preset duration is taken as an integer multiple of the average flight traverse duration, such as one time. The preset statistical days are determined based on the periodicity of airspace operations; for example, the preset statistical days are taken as seven consecutive days of operation.

[0023] The airspace is divided into cell grids as follows: The geographic coordinates of all fixed route centerlines in the global aeronautical chart airspace are extracted. The spacing between any two adjacent fixed route centerlines in the direction perpendicular to the fixed route centerline is calculated. The average of all spacings is taken as the average spacing of the fixed route centerlines. This average spacing is used as the dimension of the cell grid along the direction perpendicular to the fixed route centerline. The straight-line distances between adjacent waypoints on each fixed route centerline are extracted. The average of all straight-line distances between adjacent waypoints is taken as the average distance between adjacent waypoints on the fixed route centerline. This average distance is used as the dimension of the cell grid along the direction of the fixed route centerline. Each cell grid is a rectangular area defined by the dimension along the direction perpendicular to the fixed route centerline and the dimension along the fixed route centerline. The global aeronautical chart airspace is divided into an array of cell grids arranged along and perpendicular to the fixed route centerline. Each cell grid is uniquely identified by a row index and a column index. The row index corresponds to the position along the direction perpendicular to the centerline of the fixed route, and the column index corresponds to the position along the centerline of the fixed route.

[0024] Extract the historical flight path points before the flight enters the cell grid and calculate the fractal dimension of the historical flight path points. When the fractal dimension is lower than the preset dimension threshold, it is determined that the flight is in a fixed route operation state within the cell grid; otherwise, it is determined that the flight is in a free route operation state. The specific implementation method is as follows.

[0025] After acquiring flight plans and track data for all flights within the global aeronautical chart airspace within a preset time window and dividing the data into cell grids, processing is performed on any cell grid that any flight passes through. All track points within a preset flight duration before the flight enters the cell grid are extracted from the track data as historical track points. The preset flight duration is determined based on the typical flight time of a flight within a cell grid. Specifically, it is determined by: statistically analyzing the actual flight duration of all flights passing through a cell grid in the historical track data, calculating the average of all actual flight durations as the average flight duration of the cell grid, and taking twice the average flight duration of the cell grid as the preset flight duration. Historical track points are all track points extracted from the track data within a closed interval from the time the flight enters the cell grid minus the preset flight duration to the time the flight enters the cell grid.

[0026] The fractal dimension of historical track points is calculated using box counting. Box counting covers the plane containing the historical track points with a series of covering grids whose side lengths increase geometrically. The initial side length of the covering grid is determined by the smaller of the dimension of the unit grid along the fixed flight path centerline and the dimension perpendicular to the fixed flight path centerline, for example, 1% of the smaller value. The side length of the covering grid increases in a geometric progression with a common ratio of 2, with the upper limit of the increase being the smaller of the dimension of the unit grid along the fixed flight path centerline and the dimension perpendicular to the fixed flight path centerline. For each covering grid side length, the plane containing the historical track points is divided into an array of square covering grids with side lengths equal to the covering grid side length. The number of square covering grids occupied by at least one historical track point is counted, yielding the occupied number corresponding to the covering grid side length.

[0027] The fractal dimension D is calculated as: D = |k|; where k represents the slope of the fitted line. The fitted line is obtained by linearly fitting the scatter points constructed in a logarithmic coordinate system with the natural logarithm of the covered grid side length as the x-axis and the natural logarithm of the occupied number of the covered grid side length as the y-axis, and then using the least squares method to fit the scatter points. The fractal dimension D is a dimensionless quantity.

[0028] The preset dimension threshold is determined based on statistical analysis of the fractal dimension characteristics of the spatial distribution of track points under fixed-route and free-route operating conditions. Specifically, it is determined as follows: historical track points of flights operating under fixed-route conditions are obtained, the fractal dimension of these historical track points is calculated, and the first distribution interval of the fractal dimension is statistically analyzed; historical track points of flights operating under free-route conditions are obtained, the fractal dimension of these historical track points is calculated, and the second distribution interval of the fractal dimension is statistically analyzed; the average of the upper limit of the first distribution interval and the lower limit of the second distribution interval is taken as the preset dimension threshold. Under fixed-route operating conditions, track points are clustered along the centerline of the fixed route, with a fractal dimension approaching 1; under free-route operating conditions, track points are diffusely distributed within the cell grid, with a fractal dimension approaching 2. The fractal dimension is compared with a preset dimension threshold. When the fractal dimension is lower than the preset dimension threshold, the flight is determined to be in a fixed route operation state within the cell grid. When the fractal dimension is not lower than the preset dimension threshold, the flight is determined to be in a free route operation state within the cell grid.

[0029] This step uses fractal dimension as the criterion for distinguishing route operation status. Compared to directly determining the operation status based on the planned waypoint sequence in the flight plan, this method does not rely on whether the route type field declared in the flight plan is consistent with the actual flight behavior. Instead, it quantitatively identifies the status based on the spatial distribution characteristics of the flight track points. Under fixed route operation status, track points are linearly clustered along the fixed route centerline, while under free route operation status, track points are diffusely distributed in a planar manner within the cell grid. The fractal dimensions corresponding to the two operation statuses are fundamentally different, therefore, fractal dimension can effectively distinguish between the two route operation statuses.

[0030] The difference between the fractal dimension and the preset dimension threshold is calculated. When the difference does not fall into the preset transition interval, the occupancy probability of flights operating on fixed routes is calculated using a normal distribution along the centerline of the fixed route, and the occupancy probability of flights operating on free routes is calculated using a uniform distribution within the cell grid. When the difference falls into the preset transition interval, the occupancy probability is calculated by combining the normal distribution and the uniform distribution into a mixed probability distribution using a mixing coefficient based on the difference. The specific implementation method is as follows.

[0031] After determining whether a flight is operating on a fixed or free route within the cell grid and calculating its fractal dimension, the absolute value of the difference between the fractal dimension and a preset dimension threshold is calculated as the state difference. The state difference is used to measure the degree to which the actual flight behavior of a flight within the cell grid deviates from the ideal boundary between fixed and free route operating states.

[0032] The preset transition interval is a numerical range centered on a preset dimension threshold, extending outwards at half the width of the preset transition interval. The half-width of the preset transition interval is determined based on the degree of overlap in the distribution of the fractal dimensions of flight track points under fixed-route and free-route operating conditions. Specifically, it is determined as follows: historical flight track points of flights operating under fixed-route conditions are obtained, the fractal dimension of these historical flight track points under fixed-route conditions is calculated, and the first standard deviation of the fractal dimension is calculated; historical flight track points of flights operating under free-route conditions are obtained, the fractal dimension of these historical flight track points under free-route conditions is calculated, and the second standard deviation of the fractal dimension is calculated; half of the average of the first and second standard deviations is taken as the half-width of the preset transition interval.

[0033] The state difference is compared with half the width of the preset transition interval. When the state difference is greater than half the width of the preset transition interval, it indicates that the gap between the fractal dimension and the preset dimension threshold exceeds the range of the operational state transition region, the flight's operational state within the cell grid is clear, and the difference does not fall within the preset transition interval. In this case, the following procedures are followed: If the flight is operating on a fixed route within the cell grid, the probability of the flight occupying the cell grid is calculated using a normal distribution along the centerline of the fixed route; if the flight is operating on a free route within the cell grid, the probability of the flight occupying the cell grid is calculated using a uniform distribution within the cell grid.

[0034] The construction method for the normal distribution along the centerline of a fixed flight path is as follows: Historical flight path points before a flight enters the cell grid are extracted from the flight path data. The lateral deviation of these historical flight path points relative to the centerline of the fixed flight path is calculated, where the lateral deviation is the vertical distance from the historical flight path point to the centerline. The mean of all lateral deviations is calculated as the mean of the normal distribution, and the standard deviation of all lateral deviations is calculated as the standard deviation of the normal distribution. When the standard deviation is lower than a preset lower limit, it is replaced by the preset lower limit. The preset lower limit is determined based on the half-width of the fixed flight path within the cell grid. Specifically, the designed half-width of the fixed flight path within the cell grid is extracted, and one-third of this designed half-width is taken as the preset lower limit. Flights flying on fixed routes are constrained by the flight path width, and the standard deviation of the lateral deviation will not infinitely approach zero. When the statistically obtained standard deviation is too small, it indicates that the number of historical flight path point samples is insufficient or there are anomalies. Replacing it with the preset lower limit ensures the rationality of the probability distribution. The normal distribution covers approximately 95% of the probability distribution with the mean as the center and twice the standard deviation as the radius. This interval is matched with the actual width of a fixed airway.

[0035] The probability density function f(x) of a normal distribution along a fixed flight path centerline is specifically expressed as follows: Where x represents the lateral coordinate perpendicular to the fixed route centerline, μ represents the mean of the lateral deviations of historical track points relative to the fixed route centerline, σ represents the standard deviation of the lateral deviations of historical track points relative to the fixed route centerline, and π is pi. The longitudinal range of the flight within the cell grid along the fixed route centerline is obtained by multiplying the flight time within the cell grid along the fixed route centerline by the uniform speed at which the flight traverses the cell grid along the fixed route centerline. The integral value of this normally distributed value within the region enclosed by the longitudinal range and the cell grid dimensions perpendicular to the fixed route centerline is used as the probability of the flight occupying the cell grid.

[0036] Uniform distribution within a cell grid means that a flight has an equal probability of appearing at any position within the cell grid. The probability density function g(x,y) for uniform distribution within a cell grid is calculated as: g(x,y)=1 / (A); where x represents the lateral coordinate perpendicular to the fixed route centerline, y represents the longitudinal coordinate along the fixed route centerline, and A represents the area of ​​the cell grid. The integral value of the flight uniformly distributed within the cell grid is taken as the probability of the flight occupying the cell grid.

[0037] When the state difference is less than or equal to half the width of the preset transition interval, it indicates that the difference between the fractal dimension and the preset dimension threshold is within the range of the operational state transition region, and the difference falls into the preset transition interval. In this case, the ratio of the state difference to half the width of the preset transition interval is used as the mixing coefficient. The mixing coefficient r is calculated as: r = ΔD / (W / 2); where ΔD represents the state difference, and W represents half the width of the preset transition interval. The mixing coefficient ranges from 0 to 1. The mixing coefficient is used as the weight to weight the normal distribution along the fixed flight path centerline, and the complement of the mixing coefficient is used as the weight to weight the uniform distribution within the cell grid. The complement of the mixing coefficient is 1 minus the mixing coefficient. The probability distribution obtained by weighted summation is used as the mixed probability distribution. The probability density function h(x,y) of the mixed probability distribution is calculated as: h(x,y) = r × f(x) + (1-r) × g(x,y). The integral value of the mixed probability distribution within the cell grid is used as the occupancy probability of the flight within the cell grid.

[0038] This step employs a pre-defined transition interval and a mixed probability distribution to handle transitional scenarios in route operation status determination. Compared to a binary division of operation status with no clear distinction and the direct application of a single probability distribution, this approach constructs a mixed probability distribution from the perspective of continuous transition. This makes the calculation of flight occupancy probability more closely resemble the actual physical process of the gradual shift in the spatial distribution of flights from clustered to dispersed during the route rule transition period. Fixed route operation status and free route operation status are not completely separated during the transition period; rather, there exists a fuzzy interval. Flights whose fractal dimension falls within this interval exhibit spatial distribution characteristics of both operation statuses. The mixed probability distribution quantitatively characterizes the proportion of each characteristic through a mixing coefficient.

[0039] Based on the occupancy probability of each flight in each cell grid at each time period, a spatiotemporal conflict matrix is ​​constructed, and the specific implementation method is as follows.

[0040] After calculating the occupancy probability of each flight in each cell grid, the preset time window is divided into continuous and non-overlapping time periods. The number of preset time periods is determined based on the time resolution requirements of air traffic management for conflict prediction; the higher the time resolution requirement, the larger the number of preset time periods. For example, when the total duration of the preset time window is 60 minutes and the time resolution requirement is 1 minute, the number of preset time periods is 60, and the duration of each time period is the total duration of the preset time window (60 minutes) divided by the number of preset time periods (60), i.e., the duration of each time period is 1 minute. The time periods are arranged chronologically, and each time period is uniquely identified by a time period number, ranging from 1 to the preset time period number.

[0041] For each cell grid in each time period, the occupancy probability of all flights within that cell grid in that time period is extracted. The occupancy probability is obtained by calculating the difference between the fractal dimension and a preset dimension threshold and determining the probability distribution. The occupancy probability of each flight in each cell grid is then truncated according to the cell grid and time period. All flights within that cell grid in that time period are paired up in pairs. The number of pairings is N×(N-1) / 2, where N represents the total number of flights within that cell grid in that time period.

[0042] The conflict probability C is calculated as: C = Σ(Pi × Pj); where Pi represents the probability of the i-th flight in a pair being occupied, Pj represents the probability of the j-th flight in a pair being occupied, Σ represents the summation over all pairs, and i and j are the flight numbers when traversing all pairs, with i not equal to j. The conflict probability is a comprehensive measure of the probability that any two flights in the same cell will simultaneously occupy that cell within the same time period, and the dimension of the conflict probability is the square of the probability.

[0043] A spatiotemporal conflict matrix is ​​constructed with each unit grid as a row, each time period as a column, and each conflict probability as a matrix element value. The number of rows in the spatiotemporal conflict matrix is ​​the total number of unit grids, the number of columns in the spatiotemporal conflict matrix is ​​the preset number of time periods, the row order corresponds to the row index and column index of the unit grid, and the column order corresponds to the time period number.

[0044] After grouping the occupancy probabilities of each flight in each cell grid at each time period according to cell grid and time period, the conflict probability is calculated and a spatiotemporal conflict matrix is ​​constructed. Compared with the method of directly superimposing the occupancy probabilities of each flight in the cell grid as the conflict metric, the process of multiplying and accumulating the probabilities in pairs can distinguish the difference in conflict risk between two high occupancy probability flights and multiple low occupancy probability flights in the cell grid at the same time. The contribution of the pairwise multiplication result of high occupancy probability flights to the accumulated result is greater than that of the pairwise multiplication result of low occupancy probability flights, which is consistent with the collaborative risk principle of multiple aircraft approaching at the same time in air traffic management.

[0045] The cell grids in the spatiotemporal conflict matrix whose conflict probability exceeds a preset probability threshold and their corresponding time periods are determined as the route conflict prediction results. The specific implementation method is as follows.

[0046] After constructing the spatiotemporal conflict matrix, the conflict probabilities of each cell grid in the spatiotemporal conflict matrix are traversed for each time period. The traversal order is to obtain the conflict probabilities of each cell grid in the order of the time period number. The conflict probabilities are derived from the conflict probabilities calculated during the construction of the spatiotemporal conflict matrix based on the occupancy probabilities of each flight in each cell grid for each time period.

[0047] The preset probability threshold is determined comprehensively based on the sensitivity requirements and false alarm tolerance requirements of air traffic management for conflict alarms. Specifically, it is determined as follows: The cell grids and corresponding time periods recorded by the air traffic management system as having route conflict events within a preset statistical period are obtained historically. The conflict probability of the corresponding cell grid and time period in the spatiotemporal conflict matrix at the time of each historical route conflict event is extracted. The distribution interval of the conflict probability for all historical route conflict events is calculated, and the lower quartile of this distribution interval is taken as the preset probability threshold. The lower quartile refers to the conflict probability value at the quarter-th quartile position after arranging the conflict probabilities of all historical route conflict events in ascending order. The dimension of the preset probability threshold is consistent with that of the conflict probability; the dimension of the preset probability threshold is the square of the probability. The preset statistical period is determined based on the periodicity of airspace operations; for example, the preset statistical period is 30 consecutive days of operation.

[0048] The conflict probability is compared with a preset probability threshold. When the conflict probability exceeds the preset threshold, it indicates that the cumulative sum of the probabilities of any two flights simultaneously occupying the corresponding cell grid within the corresponding time period reaches the conflict probability level corresponding to historical route conflict events, thus determining that a route conflict exists within the corresponding cell grid during the corresponding time period. Cell grids with conflict probabilities exceeding the preset probability threshold and their corresponding time periods are extracted, and these extracted cell grids and corresponding time periods are used as the route conflict prediction results. The route conflict prediction results include the row index of the cell grid, the column index of the cell grid, and the time period sequence number, which are used to determine the geographical location and time interval of the route conflict that needs attention in the global aeronautical chart airspace.

[0049] When the probability of conflict does not exceed the preset probability threshold, it means that the probability of conflict of the corresponding cell grid in the corresponding time period has not reached the level of the probability of conflict corresponding to the historical route conflict events, and the corresponding cell grid and the corresponding time period are not included in the route conflict prediction results.

[0050] This step uses the lower quartile of the distribution range of conflict probabilities corresponding to historical route conflict events as the preset probability threshold. Compared to using fixed empirical values ​​as the preset probability threshold, this method determines the threshold based on the statistical characteristics of conflict probabilities in actual operational data, ensuring that the sensitivity of route conflict prediction matches the actual airspace operational safety level. The selection of the lower quartile balances the sensitivity of route conflict prediction with the control of false alarms, capturing the conflict probability levels corresponding to most historical route conflict events while eliminating the downward skew of the preset probability threshold caused by extremely low values ​​in the conflict probability distribution.

[0051] This embodiment forms a logically interconnected technical chain around the technical objective of ensuring the reliability of conflict prediction in the transition zone of airway rules. Fractal dimension here is not simply a mathematical measure of the geometric shape of a flight track, but rather establishes a quantitative mapping relationship between the degree of clustering of spatial distribution of track points and the flight's route operation mode. This transforms the determination of operational status from relying on flight plan declaration data to autonomous identification based on track geometric characteristics. The application of this cross-domain metric in this technical scenario and its correspondence with operational status are not documented in existing air traffic management fields. The introduction of a preset transition interval and a mixed probability distribution further solves the problem of discontinuous probability distribution switching at the boundary of operational status determination. Using the state difference as a bridge, the determination result of fractal dimension is converted into a weighted allocation between normal and uniform distributions, ensuring that the calculation method of occupancy probability changes continuously rather than abruptly with the operational status tendency. The mapping relationship between fractal dimension and mixing coefficient is concise and has a clear physical meaning. Thresholds and parameters can be determined based on statistical analysis of track data. The implementation process does not rely on external systems or complex models, achieving a balance between the uniqueness of the technical concept and the feasibility of engineering implementation.

[0052] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0053] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0054] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

Claims

1. A method for predicting global aeronautical chart route conflicts based on an airspace conflict matrix, characterized in that, The steps include the following: S1: Obtain flight plans and track data of each flight in the global aeronautical chart airspace within a preset time window, and divide the airspace into cell grids; S2: For any cell grid that any flight passes through, extract the historical flight path points before the flight enters the cell grid and calculate the fractal dimension of the historical flight path points. When the fractal dimension is lower than the preset dimension threshold, it is determined that the flight is in a fixed route operation state within the cell grid; otherwise, it is determined that the flight is in a free route operation state. S3: Calculate the difference between the fractal dimension and the preset dimension threshold. When the difference does not fall into the preset transition range, the occupancy probability of flights in fixed route operation state is calculated by normal distribution along the center line of the fixed route, and the occupancy probability of flights in free route operation state is calculated by uniform distribution within the cell grid. When the difference falls into the preset transition range, the normal distribution and the uniform distribution are combined into a mixed probability distribution based on the difference to calculate the occupancy probability using a mixing coefficient. S4: Construct a spatiotemporal conflict matrix based on the occupancy probability of each flight in each cell grid at each time period; S5: The cell grids in the spatiotemporal conflict matrix whose conflict probability exceeds the preset probability threshold and their corresponding time periods are determined as the route conflict prediction results.

2. The global aeronautical chart route conflict prediction method based on the airspace conflict matrix according to claim 1, characterized in that, Acquire flight plans and track data for each flight within a preset time window in the global aeronautical chart airspace, and divide the airspace into a grid of cells, including: The flight plans of all flights entering the global chart airspace within a preset time window are extracted from the flight plan database of the air traffic management system. The track data corresponding to the flight plans of all flights are extracted from the track database of the air traffic management system. The start point of the preset time window is the current time, and the end point of the preset time window is the current time plus a preset duration. The size of the cell grid along the direction perpendicular to the fixed route centerline is determined by the average spacing of the fixed route centerline in the global chart airspace, and the size of the cell grid along the fixed route centerline is determined by the average distance between adjacent waypoints on the fixed route centerline.

3. The global aeronautical chart route conflict prediction method based on the airspace conflict matrix according to claim 2, characterized in that, The preset duration of the preset time window is determined based on the average flight time of all flights flying over the unit grid covered by the global aeronautical chart airspace, and is an integer multiple of the average flight time.

4. The global aeronautical chart route conflict prediction method based on the airspace conflict matrix according to claim 1, characterized in that, For any cell grid that any flight passes through, extract the historical flight path points before the flight enters the cell grid and calculate the fractal dimension of the historical flight path points. If the fractal dimension is lower than a preset dimension threshold, the flight is determined to be in a fixed-path operation state within the cell grid; otherwise, it is determined to be in a free-path operation state, including: All flight points within a preset flight duration before a flight enters a cell grid are extracted from the flight data as historical flight points. The plane containing the historical flight points is divided into a coverage grid with proportionally increasing side lengths using box counting. The number of historical flight points occupying each coverage grid under each side length is counted. The absolute value of the slope of the straight line fitted by the logarithm of the coverage grid side length and the logarithm of the corresponding number of occupied points is used as the fractal dimension. The fractal dimension is compared with a preset dimension threshold. When the fractal dimension is lower than the preset dimension threshold, the flight is determined to be in a fixed-path operation state within the cell grid. When the fractal dimension is not lower than the preset dimension threshold, the flight is determined to be in a free-path operation state within the cell grid.

5. The global aeronautical chart route conflict prediction method based on the airspace conflict matrix according to claim 4, characterized in that, Scattered points in a logarithmic coordinate system are constructed with the logarithm of the covered grid side length as the x-axis and the logarithm of the corresponding occupied quantity as the y-axis. The least squares method is used to perform linear fitting on the scattered points to obtain the fitted line, and the absolute value of the slope of the fitted line is taken as the fractal dimension.

6. The global aeronautical chart route conflict prediction method based on the airspace conflict matrix according to claim 1, characterized in that, Calculate the difference between the fractal dimension and the preset dimension threshold. When the difference does not fall into the preset transition range, the occupancy probability of flights operating on fixed routes is calculated using a normal distribution along the center line of the fixed route, and the occupancy probability of flights operating on free routes is calculated using a uniform distribution within the cell grid. When the difference falls within a preset transition interval, a mixing coefficient based on the difference is used to combine the normal distribution and the uniform distribution into a mixed probability distribution to calculate its occupancy probability, including: The absolute value of the difference between the fractal dimension and the preset dimension threshold is used as the state difference. The state difference is compared with half the width of the preset transition interval. When the state difference is greater than half the width of the preset transition interval, the difference does not fall into the preset transition interval. When the state difference is less than or equal to half the width of the preset transition interval, the difference falls into the preset transition interval. When the difference falls into the preset transition interval, the ratio of the state difference to half the width of the preset transition interval is used as the mixing coefficient. The probability distribution obtained by weighting the normal distribution along the fixed flight path centerline with the mixing coefficient as the weight and the uniform distribution within the cell grid with the complement of the mixing coefficient as the weight is used as the mixed probability distribution. The occupancy probability of the flight within the cell grid is calculated using the mixed probability distribution.

7. The global aeronautical chart route conflict prediction method based on the airspace conflict matrix according to claim 6, characterized in that, The mean of the lateral deviations of the historical flight path points before entering the cell grid relative to the centerline of the fixed route are used as the mean of the normal distribution, and the standard deviation of the lateral deviations is used as the standard deviation of the normal distribution to construct a normal distribution along the centerline of the fixed route. When the standard deviation is lower than the preset lower limit of the standard deviation, the preset lower limit of the standard deviation is used to replace the standard deviation for construction.

8. The global aeronautical chart route conflict prediction method based on the airspace conflict matrix according to claim 1, characterized in that, Based on the occupancy probability of each flight in each cell grid at each time period, a spatiotemporal conflict matrix is ​​constructed, including: For each time period within a preset time window, the occupancy probabilities of all flights within the same time period of the same unit grid are multiplied in pairs. The sum of all pairwise multiplication results is taken as the conflict probability of the unit grid in that time period. A spatiotemporal conflict matrix is ​​constructed with each unit grid as the row, each time period as the column, and each conflict probability as the element value.

9. The global aeronautical chart route conflict prediction method based on the airspace conflict matrix according to claim 8, characterized in that, The preset time window is divided into consecutive and non-overlapping time periods according to the preset number of time periods. The duration of each time period is equal to the total duration of the preset time window divided by the preset number of time periods.

10. The global aeronautical chart route conflict prediction method based on the airspace conflict matrix according to claim 1, characterized in that, The cell grids in the spatiotemporal conflict matrix whose conflict probability exceeds a preset probability threshold and their corresponding time periods are identified as route conflict prediction results, including: The conflict probability of each cell in the spatiotemporal conflict matrix is ​​traversed for each time period. The conflict probability is compared with a preset probability threshold. Cells with conflict probabilities exceeding the preset probability threshold and their corresponding time periods are extracted. The extracted cell and corresponding time periods are used as the route conflict prediction results.

Images