A restricted spatial unit identification method based on DBSCAN clustering algorithm

Abstract

The present invention discloses a restricted airspace unit identification method based on the DBSCAN clustering algorithm. The method first defines the airspace unit; secondly, collects the actual flight operation data and the corresponding route information, matches the route information with the latitude and longitude coordinate data of the airspace unit to construct a route network model; then, calculates the flight delay data, downsamples the data, and associates it with the airspace unit to obtain the flight delay feature data; finally, combines the route network model with the flight delay characteristics, and constructs an identification model for the restricted airspace unit based on the DBSCAN clustering algorithm. The method of the present invention can identify restricted airspace units and restricted start and end times, which helps to solve the randomness caused by events such as bad weather and air traffic control. Determining the start and end time of the airspace restriction can separate the direct flight delays from the delays caused by propagation, and further explore the propagation mechanism of flight delays, which is of great significance to improving the accuracy of flight delay prediction.

Description

Technical Field

[0001] The invention relates to a restricted airspace unit identification method based on a DBSCAN clustering algorithm, and belongs to the technical field of air traffic management. Background Art

[0002] With the rapid development of civil aviation, flight delays have become an important issue in the air transportation system, and the aviation industry has been suffering economic losses. In actual operation, there are many factors that lead to flight delays, such as airline operation management factors, weather factors, air traffic control, military activities, etc. These factors will restrict airspace units to varying degrees, resulting in flight delays. At the same time, the emergence and occurrence of these influencing factors are highly random and uncertain. When these factors, such as bad weather, suddenly appear on the route, even if there are no restrictions within and near the airport, flights will still be delayed. Moreover, the delay of a previous flight may also cause delays in multiple subsequent flights. Therefore, flight delays show a chaotic or nonlinear development trend, and it is difficult to accurately predict flight delays.

[0003] At present, many scholars have done a lot of research on flight delay prediction. However, as mentioned above, flight delays are affected by many factors, which have certain suddenness and randomness. Therefore, flight delays appear chaotic or nonlinear and are difficult to predict accurately. At the same time, a flight delay may also cause subsequent flight delays. Although some scholars consciously separate direct delays from delays caused by propagation, no effective method has been found to quantitatively distinguish the two. When an event that causes flight delays occurs, the airspace unit is restricted, causing flight delays. Identifying restricted airspace units and restricted start and end times can help solve the randomness of delay events caused by bad weather, air traffic control, etc. Determining the restricted start and end times can separate direct delays from delays caused by propagation, and further explore the propagation mechanism of flight delays, which is of great significance to improving the accuracy of flight delay prediction. Summary of the invention

[0004] The technical problem to be solved by the present invention is: to provide a restricted airspace unit identification method based on the DBSCAN clustering algorithm, to count the cumulative delay time of each airspace unit in each time period, to match the cumulative delay time with the longitude and latitude coordinates of the airspace unit, to establish a feature matrix, and to perform DBSCAN clustering on each time period to identify the restricted airspace unit.

[0005] The present invention adopts the following technical solutions to solve the above technical problems:

[0006] A restricted airspace unit identification method based on DBSCAN clustering algorithm comprises the following steps:

[0007] Step 1: Define airspace units as airports and waypoints, and restricted airspace units as airspace units with reduced capacity due to weather, military activities, and air traffic control factors;

[0008] Step 2: Collect historical flight data and chart data, extract route data from the historical flight data, and build a route network model by combining the longitude and latitude coordinates of each airspace unit in the chart data;

[0009] Step 3: Divide a day into 96 time periods at intervals of 15 minutes. Extract the cumulative correlation features of airspace delays including time series from the historical flight data in the order of the time periods. Establish an airspace unit data set including time features, spatial features and cumulative correlation features of delays based on the cumulative correlation features of airspace delays.

[0010] Step 4: Use the density-based DBSCAN clustering algorithm to cluster the airspace unit data set including time features, spatial features, and delay accumulation correlation features, and combine historical flight data to identify airspace restricted units and delay characteristics from all airspace units, including: the number, location, and restricted start and end times of airspace restricted units.

[0011] As a preferred solution of the present invention, the specific process of step 2 is as follows:

[0012] Step 21, collecting historical flight data and chart data, wherein the historical flight data includes the flight number, the expected take-off and landing time of the flight, the actual take-off and landing time, the take-off and landing airports, the route used and the waypoints on the route, and the chart data includes the latitude and longitude coordinates of the airspace unit;

[0013] Step 22, extracting all route data from the historical flight data, each route data including the flight number, take-off and landing airports, the route used, and the waypoints on the route;

[0014] Step 23, associate the take-off airport, landing airport and waypoints on each route with their corresponding longitude and latitude coordinates, that is, associate each airspace unit on the route with its corresponding longitude and latitude coordinates, and construct a route network model.

[0015] As a preferred solution of the present invention, the specific process of step 3 is as follows:

[0016] Step 31, counting flight delays for the departure airport and waypoints on the route according to the departure delay time of the flight, and counting flight delays for the landing airport according to the landing delay time of the flight, and a flight delay is considered to be a flight delay when the departure delay time or the landing delay time is greater than 15 minutes;

[0017] Step 32, dividing a day into 96 time periods at intervals of 15 minutes, and counting the flight delays passing through each airspace unit in each time period in the order of the time periods, and then counting the cumulative flight delays of each airspace unit in each time period, which is the airspace delay cumulative correlation feature including the time series;

[0018] Step 33, taking the time period corresponding to the flight passing through each airspace unit as the time feature and the longitude and latitude coordinates of each airspace unit as the spatial feature, to establish an airspace unit data set including the time feature, spatial feature and delay accumulation correlation feature.

[0019] As a preferred solution of the present invention, in step 31, the flight delay is counted according to the departure delay time of the flight at the departure airport and the waypoints on the route, and the flight delay is counted according to the arrival delay time of the flight at the landing airport, specifically:

[0020] The flight delay at the departure airport is equal to the actual departure time of the departing flight minus the estimated departure time;

[0021] The flight delay at the waypoint is equal to the flight delay at the departure airport;

[0022] The flight delay at the landing airport is equal to the actual landing time of the incoming flight minus the estimated landing time.

[0023] As a preferred solution of the present invention, the accumulated flight delays in step 32 are specifically: the sum of flight delays of all flights passing through a certain airspace unit in a certain time period.

[0024] As a preferred solution of the present invention, the specific process of step 4 is as follows:

[0025] Step 41, based on the route network model established in step 2, the restricted airspace unit identification model established by combining the time characteristics, spatial characteristics and delay accumulation correlation characteristics is:

[0026]

[0027] Among them, D T represents the restricted spatial unit recognition model for the Tth time period, T = 1, ..., 96, are the latitude and longitude of the first airspace unit, are the latitude and longitude of the second airspace unit, are the latitude and longitude of the kth airspace unit, respectively, k is the number of airspace units on all routes, They are airspace units U1, U2, U k The cumulative flight delays in the Tth time period;

[0028] Step 42, using the DBSCAN clustering algorithm to cluster the restricted airspace unit identification models of each time period, identifying the restricted airspace units from all airspace units, and obtaining delay characteristics based on the restricted airspace units, including: the number, location and restricted start and end times of the airspace restricted units.

[0029] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0030] The present invention identifies restricted airspace units in the airspace and identifies the start and end times of the restricted airspace units, which helps to solve the randomness caused by events such as bad weather and air traffic control. It is of great significance to deeply explore the propagation mechanism of flight delays and analyze the impact of a flight delay on subsequent flights in detail, and it can also further improve the accuracy of flight delay prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic diagram of a restricted airspace unit identification method based on the DBSCAN clustering algorithm of the present invention.

[0032] Figure 2 It is a schematic diagram of the route network model.

[0033] Figure 3 This is a restricted airspace unit identification result diagram from 22:15:00 to 22:29:59 on July 17, 2017 in the embodiment.

[0034] Figure 4 This is the restricted airspace unit distribution map from 22:15:00 to 22:29:59 on July 17, 2017 in the embodiment.

[0035] Figure 5 This is the thermal distribution map of the severely restricted airspace unit on July 17, 2017 in the embodiment. DETAILED DESCRIPTION

[0036] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0037] like Figure 1As shown, the present invention calculates the departure delay and arrival delay of each flight by mining historical flight data, and corresponds each flight to the route information it uses. Then, the coordinate information of each airport and waypoint is collected, and the historical flight data is classified and grouped according to the time series and the departure airport. The cumulative delay time and delay time of each airspace unit are counted, and the average delay time of each time period is matched with the coordinates of each airspace unit to establish a feature matrix. Finally, DBSCAN clustering is performed on each time period to identify restricted airspace units. The specific implementation steps are as follows:

[0038] Step 1, define the airspace units mentioned in the present invention. In the present invention, airspace units refer to waypoints and airports; restricted airspace units refer to airspace units with reduced capacity due to factors such as weather, military activities, and air traffic control.

[0039] Step 2: Collect historical flight data and chart data. Extract the round-trip route information between each target airport from the historical flight data, and build a route network model based on the longitude and latitude coordinates of each waypoint in the chart. The specific steps include:

[0040] (2.1) Collect historical flight data and chart data. The historical flight data includes the flight's estimated take-off / landing time, actual take-off / landing time, take-off / landing airport, flight number, and route used;

[0041] (2.2) Grouping historical flight data according to take-off and landing airports, and extracting route data used by round-trip flights between airports;

[0042] (2.3) Associate the route data with the longitude and latitude coordinate data of the airspace unit to construct a route network model and visualize it. Taking the flight data of the round-trip routes of Beijing Capital Airport (ZBAA), Shanghai Pudong Airport (ZSPD), Guangzhou Baiyun Airport (ZGGG) and Chengdu Shuangliu Airport (ZUUU) as an example, group them according to the departure airport and landing airport, and extract the route information used by each group. The route is represented by the waypoints it passes through. Mark the longitude and latitude coordinates of each point on the 12 routes of the four target airports in a plane rectangular coordinate system, and then connect the path points (waypoints and airports) in the order in which the aircraft flies over to form the following. Figure 2 Schematic diagram of the route shown.

[0043] Step 3: Perform correlation analysis on flight delay characteristics and airspace units to provide support for identifying airspace restricted unit information from historical data. Clean and resample historical flight schedule data, and cumulatively correlate flight delay data at each time interval with the airspace units it is expected to use, extract the cumulative correlation features of airspace delays including time series, and create a new airspace unit data set based on time features, spatial features (longitude / latitude) and cumulative correlation features of delays. Specifically, the following steps are included:

[0044] (3.1) Calculate the take-off and landing delays of each flight. In the present invention, a take-off / landing delay of more than 15 minutes is considered a flight delay;

[0045] (3.2) The flight delay characteristics are associated with the airspace units. In the present invention, the accumulated delay time is selected as the characteristic for research. The departure airport, the route and the waypoints are counted according to the departure delay of the flight, and the landing airport is counted according to the landing delay of the flight;

[0046] (3.3) Clean and resample the historical flight schedule data, divide a day into 96 time periods at intervals of 15 minutes, cumulatively associate the flight delay data of each time interval with the airspace unit it is expected to use, and count the cumulative flight delays of each airspace unit in each time period, i.e., the cumulative association characteristics of delays;

[0047] (3.4) The time period corresponding to the flight passing through each airspace unit is taken as the time feature, and the longitude and latitude coordinates of each airspace unit are taken as the spatial feature. A new airspace unit dataset is created based on the time feature, spatial feature (longitude / latitude) and delay accumulation correlation feature.

[0048] Step 4: Use the density-based DBSCAN clustering algorithm to cluster the airspace unit data set including time features, spatial features (longitude / latitude) and delay accumulation correlation features, and combine the expected and actual take-off and landing information of historical flights to identify airspace restricted units and delay features from all airspace units, including: the number, location and restricted start and end times of airspace restricted units. Specifically, the following steps are included:

[0049] (4.1) Based on the route network model established in the second step, the restricted airspace unit identification model is established by combining the time characteristics, spatial characteristics (longitude / latitude) and delay accumulation correlation characteristics. The characteristic matrix D of the T period is T for:

[0050]

[0051] Where: is the airspace unit U k Latitude; is the airspace unit U kLongitude; T is the time period; is the airspace unit U k The accumulated delay time in time period T;

[0052] (4.2) Use the DBSCAN clustering algorithm to cluster the restricted airspace unit identification model and identify the restricted airspace units;

[0053] (4.3) Output the recognition results for each time period and visualize the results. Figure 3 The following is the result of the restricted airspace unit identification for the routes between the above four airports from 22:15:00 to 22:29:59 on July 17, 2017. In order to more intuitively see the distribution of restricted path points in the airspace, the Figure 4 The restricted airspace unit distribution diagram is shown in Figure 2. The identification results of all flights in all 96 time periods on July 17, 2017 are comprehensively analyzed, and the number of airspace unit restrictions is counted to draw the following Figure 5 The heat distribution diagram shown in the figure shows that 0 to 55 represents the number of times the airspace unit is restricted in 96 time periods in a day. On July 17, 2017, Chengdu Shuangliu Airport had the most restrictions, followed by Beijing Capital Airport, Guangzhou Baiyun Airport and Shanghai Pudong Airport; for waypoints, the top five waypoints with the most restrictions in a day are ONEBA, SUBUL, BOBAK, JTG and WFX.

[0054] The above embodiments are only for illustrating the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A restricted spatial unit identification method based on DBSCAN clustering algorithm, characterized in that: The steps include: Step 1: Define airspace units as airports and waypoints, and restricted airspace units as airspace units with reduced capacity due to weather, military activities, and air traffic control factors; Step 2: Collect historical flight data and chart data, extract route data from historical flight data, and build a route network model based on the longitude and latitude coordinates of each airspace unit in the chart data. The specific process is as follows: Step 21, collecting historical flight data and chart data, wherein the historical flight data includes the flight number, the expected take-off and landing time of the flight, the actual take-off and landing time, the take-off and landing airports, the route used and the waypoints on the route, and the chart data includes the latitude and longitude coordinates of the airspace unit; Step 22, extracting all route data from the historical flight data, each route data including the flight number, take-off and landing airports, the route used, and the waypoints on the route; Step 23, associating the take-off airport, landing airport and waypoints on each route with their corresponding longitude and latitude coordinates, that is, associating each airspace unit on the route with its corresponding longitude and latitude coordinates, and constructing a route network model; Step 3: Divide a day into 96 time periods at intervals of 15 minutes. Extract the cumulative correlation features of airspace delays including time series from the historical flight data in the order of the time periods. Establish an airspace unit data set including time features, spatial features and cumulative correlation features of delays based on the cumulative correlation features of airspace delays. The specific process is as follows: Step 31, counting flight delays for the departure airport and waypoints on the route according to the departure delay time of the flight, and counting flight delays for the landing airport according to the landing delay time of the flight, and a flight delay is considered to be a flight delay when the departure delay time or the landing delay time is greater than 15 minutes; The flight delays are calculated based on the departure delay time of the flight at the departure airport and the waypoints on the route, and the flight delays are calculated based on the arrival delay time of the flight at the landing airport, specifically: The flight delay at the departure airport is equal to the actual departure time of the departing flight minus the estimated departure time; The flight delay at the waypoint is equal to the flight delay at the departure airport; The flight delay at the landing airport is equal to the actual landing time of the incoming flight minus the estimated landing time; Step 32, dividing a day into 96 time periods at intervals of 15 minutes, and counting the flight delays passing through each airspace unit in each time period in the order of the time periods, and then counting the cumulative flight delays of each airspace unit in each time period, which is the airspace delay cumulative correlation feature including the time series; The cumulative flight delay is specifically: the sum of flight delays of all flights passing through a certain airspace unit during a certain time period; Step 33, taking the time period corresponding to the flight passing through each airspace unit as the time feature, and the longitude and latitude coordinates of each airspace unit as the spatial feature, to establish an airspace unit data set including the time feature, the spatial feature and the delay accumulation correlation feature; Step 4: Use the density-based DBSCAN clustering algorithm to cluster the airspace unit data set including time features, spatial features and delay accumulation correlation features, and combine historical flight data to identify airspace restricted units and delay features from all airspace units, including: the number, location and restricted start and end times of airspace restricted units; the specific process is as follows: Step 41, based on the route network model established in step 2, the restricted airspace unit identification model for each time period is established by combining the time characteristics, spatial characteristics and delay accumulation correlation characteristics: Among them, D T represents the restricted spatial unit recognition model for the Tth time period, T = 1, ..., 96, are the latitude and longitude of the first airspace unit, are the latitude and longitude of the second airspace unit, are the latitude and longitude of the kth airspace unit, respectively, k is the number of airspace units on all routes, They are airspace units U1, U2, U k The cumulative flight delays in the Tth time period; Step 42, using the DBSCAN clustering algorithm to cluster the restricted airspace unit identification models of each time period, identifying the restricted airspace units from all airspace units, and obtaining delay characteristics based on the restricted airspace units, including: the number, location and restricted start and end times of the airspace restricted units.

Images