CN114118578A

CN114118578A - Calculation method for predicting flight arrival time based on air trajectory and big data

Info

Publication number: CN114118578A
Application number: CN202111424205.9A
Authority: CN
Inventors: 王宇; 董震岭; 脱垄玉; 马文博; 张旭婧; 徐晓明
Original assignee: Shanghai Xingsha Technology Co ltd
Current assignee: Shanghai Xingsha Technology Co ltd
Priority date: 2021-11-26
Filing date: 2021-11-26
Publication date: 2022-03-01

Abstract

A calculation method for predicting flight arrival time based on air tracks and big data comprises the steps of floor runway prediction calculation, floor time prediction calculation and gear-loading time prediction calculation. The invention adopts flight history big data with highest simulation degree under all factors as a matching prediction sample, has higher goodness of fit with real operation, has low requirement on a data source required by calculation and low technical realization difficulty, can also quickly realize product conversion even in a medium and small airport with low data condition, has stronger reproduction and popularization, establishes a prediction mechanism and an algorithm of flight landing time and gear-on time in an airport belonging to a flight control area by carrying out staged modeling and state identification on the air operation track of an aircraft and taking big data statistical analysis as a main technical means, provides accurate reference for an airport ground guarantee system, can effectively reduce the virtual consumption waiting or delay in place of airport ground guarantee resources, and effectively improves the utilization rate of the ground guarantee resources.

Description

Calculation method for predicting flight arrival time based on air trajectory and big data

Technical Field

The invention relates to the technical field of aviation management application methods, in particular to a calculation method for predicting flight arrival time based on air tracks and big data.

Background

The landing time and the gear-adding time of civil aviation flights are very important operation reference data in the operation of airports. In practical application, all ground support personnel and equipment in an airport can make advance preparation work by referring to the predicted landing time of a flight before the inbound flight reaches a designated parking space; after the inbound flight is on the gear, the flight surrounding personnel and equipment can approach the aircraft to implement ground support work.

At present, in the ground guarantee work of civil aviation airport operation, because of lack of accurate prediction method, the predicted landing time of flight is usually referred to as reference, and ground guarantee personnel and equipment are required to arrive at the designated position for waiting guarantee 5-10 minutes before the predicted landing time of flight. The above approach suffers from one of the following disadvantages: the traditional method for calculating the predicted landing time of the flight usually adopts the historical average flight time of the whole flight from a takeoff airport to a landing airport, or relies on a computer flight planning system to carry out forward section-by-section reasoning calculation; the large deviation is easy to occur due to the influence of factors such as flight performance, aeronautical weather, flight procedures, air pipe allocation and the like. The second step is as follows: a period of aircraft ground taxi time exists between the actual landing of the aircraft and the completion of the parking of the aircraft; the estimated landing time and the time of the gear on the aircraft are greatly different due to the influences of factors such as the direction of the flight landing runway, the parking position, the ground sliding path, avoidance waiting in the ground sliding and the like. And thirdly: the starting point of the work development of most airport ground support personnel and equipment is usually the gear-in time of an aircraft, and is not the expected landing time of a flight; with reference to the traditional estimated landing time, the virtual consumption waiting or delay in the airport ground guarantee resources is easily caused.

Disclosure of Invention

In order to overcome the defect that the virtual consumption waiting or the delay in place of airport ground guarantee resources is easy to cause in the ground guarantee activities of the existing civil aviation airport operation due to the lack of an accurate prediction method, the invention provides a calculation method for predicting flight arrival time based on air tracks and big data, which can quickly realize product conversion under the combined action of related steps, has stronger reproduction and popularization performance, can provide accurate reference for an airport ground guarantee system, can effectively reduce the virtual consumption waiting or delay in place of airport ground guarantee resources, effectively improves the utilization rate of the ground guarantee resources, and is particularly suitable for accurate decision making when a flight approaches to landing emergency dispatching.

The technical scheme adopted by the invention for solving the technical problems is as follows:

a calculation method for predicting flight arrival time based on air track and big data is characterized by comprising the steps of prediction calculation of a landing runway, prediction calculation of landing time and prediction calculation of gear-loading time; the method comprises the following steps of establishing a route-runway corresponding rule table, obtaining data samples of a plurality of continuous actual landing flights, setting a current same-kind route set predicted landing runway, and replacing the predicted landing runway with the landing runway distributed by a system; recording the coordinates of waypoints related to daily inbound flights of an airport in the boundary of a flight control area, recording the coordinates of waypoints related to the daily inbound flights of the airport in the boundary of a terminal area, forming a set of initial inbound positioning points, acquiring real-time position data of an aircraft, judging waypoints flown by the aircraft, screening a plurality of continuous historical flight data samples which meet the latest occurrence condition, recording the time interval from the plurality of historical flying waypoints to actual landing, acquiring the average flight time of a flight segment, and acquiring the predicted landing time of the flight; the wheel-class loading time prediction calculation comprises the following steps of screening a plurality of recent continuous historical flight data samples meeting conditions, calculating the time interval between the actual landing of the screened plurality of historical flight data samples to the actual wheel class, calculating to obtain average ground sliding time, and calculating to obtain the predicted wheel-class loading time of the flight.

Further, in the floor runway prediction calculation, the established rule table is a many-to-many mapping relation; the obtained data sample is based on the data screened from the airline collection of the same generic type in the historical data of flights; the set predicted landing runway data is derived from three data samples; the data replacing the expected landing runway is derived from the landing runway data that the flight has already allocated by the system.

Furthermore, in the prediction calculation of the landing time, the coordinates of the route points of the daily inbound flights with the airport in the boundary of the flight control area comprise longitude and latitude; coordinates of route points of daily inbound flights to the airport in the boundary of the terminal area comprise longitude and latitude; the positioning point set data is derived from a flight program and a navigation chart; the acquired real-time location data includes longitude and latitude.

Furthermore, in the prediction calculation of the landing time, it is determined that the data of the aircraft flying over waypoints are derived from the data of the set of real-time positions of the aircraft and the boundary port waypoints of the flight control area, the data of the set of boundary port waypoints of the terminal area and the data of the set of starting point of approach positioning.

Further, in the floor time prediction calculation, the filtered historical flight data samples need to satisfy the following condition: the sample flight ever flies past the identified waypoint, condition two: the actual landing runway of the sample flight is the same as the expected landing runway of the target flight; the recorded time intervals are sorted from small to large; the average time of flight is obtained by averaging samples of time intervals; the landing time is obtained by accumulating the average flight time of the flight segment by the time of the aircraft flying over the waypoint.

Further, in the calculation of the upper gear time prediction, a first condition of screening is as follows: the sample flight is the same as the predicted flight arrival station, and the condition two is as follows: the actual landing runway of the sample flight is the same as the expected landing runway of the target flight; calculating the sorted time intervals from small to large; the average ground taxi time is derived from a plurality of time interval data samples after sequencing; the predicted time to shift on the flight is derived from the accumulated leg average flight time and average ground taxi time data samples.

The invention has the beneficial effects that: the method takes the flight program in the flight control area of the airport as the main modeling reference, adopts the flight historical big data with the highest simulation degree under all factors as the matching prediction sample, and has higher goodness of fit with the real operation. The algorithm system has simple design structure, low requirement on data sources required by calculation, low technical implementation difficulty, capability of quickly realizing product conversion even in small and medium airports with low data conditions, and strong replication popularization. The method has the advantages that the prediction mechanism and algorithm of the flight landing time and the wheel-gear-up time in the flight control area belonging to the airport are established by carrying out staged modeling and state recognition on the air operation track of the aircraft and taking big data statistical analysis as a main technical means, so that accurate reference is provided for the airport ground guarantee system, the virtual consumption waiting or delay in place of airport ground guarantee resources can be effectively reduced, the utilization rate of the ground guarantee resources is effectively improved, and the method is particularly suitable for accurate decision making when the flight approaches to landing and is emergently adjusted. Based on the above, the invention has good application prospect.

Drawings

The invention is further illustrated below with reference to the figures and examples.

FIG. 1 is a block diagram illustration of the overall architecture of the present invention.

FIG. 2 is a representation of the course-runway correspondence rule of the present invention.

FIG. 3 is a diagram illustrating a set of start approach anchor points according to the present invention.

FIG. 4 is a schematic representation of the time that an aircraft of the present invention is flying over waypoints.

FIG. 5 is a schematic diagram of the invention for obtaining average flight time of the flight.

Detailed Description

As shown in fig. 1, a calculation method for predicting flight arrival time based on air trajectory and big data includes landing runway prediction calculation, landing time prediction calculation, and upper gear time prediction calculation (the above calculation processes are all implemented by application software in a flight control area, based on a PC management system application unit [ innovation point of this embodiment ], and a collaborative flight program).

As shown in fig. 1, the landing runway prediction includes the following. The first step is as follows: establishing a route-runway corresponding rule table according to an airport runway operation rule, which is generally a many-to-many mapping relation, and is schematically shown in FIG. 2; illustrated in example 2: examples are divided into two types of route sets, each corresponding to a number of use runways. The second step is that: screening data samples of three recently-occurred continuous actual landing flights in a same-class airline set in the historical flight data; it should be noted that if data samples of three continuous actual landing flights are screened, the take-off airport is respectively: ZGGG, ZPP and ZWWW, the landing airport is the local airport, and three flights belong to the same type of airline set. The third step: in the three data samples, the landing runway with the highest occurrence probability is set as a predicted landing runway of the current similar air route set; if the runways of the three data samples are different, setting the nearest data sample runway as a landing runway predicted by the current similar route set (which usually cannot happen); it should be noted that if the actual landing runways of the three flights are 35L, 35R, and 35R, the current landing runway of the set of similar flights is estimated to be 35R; if the actual landing runways of the three flights are respectively 35L, 35R and 17R, and the landing runway of the latest flight is 17R, the current landing runway of the same type of flight set is estimated to be 17R. The fourth step: and if the flight has the data of the landing runway distributed by the system, replacing the predicted landing runway by the landing runway distributed by the system.

As shown in fig. 1, the calculation of the landing time prediction mainly includes identifying the time when the aircraft flies over a key waypoint according to a flight program and the motion trajectory of the aircraft, screening and sampling data samples closest to a predicted flight according to historical flight data, further screening the sample data according to time distribution and a confidence interval, calculating the average flight time of the flight segment, and accumulating the average flight time of the flight segment according to the time of flying over the waypoint to obtain the predicted landing time. The calculation process is as follows: the first step is as follows: and searching all route point sets related to daily inbound flights of the airport in the boundary of the flight control area according to the flight program and the chart, and recording the coordinates (longitude and latitude) of the route points to form a 'set of inbound route points of the boundary of the flight control area'. The second step is that: and searching all route point sets related to daily inbound flights of the airport in the boundary of the terminal area according to the flight program and the navigation map, and recording the coordinates (longitude and latitude) of the route points to form a terminal area boundary inbound route point set. The third step: according to the flight program and the navigation chart, all the initial approach positioning points related to daily inbound flights of the airport are searched, and the coordinates of the route points are recorded to form an initial approach positioning point set, wherein an example diagram is shown in fig. 3, and is illustrated in fig. 3: A1-A17 are the sets of the airport-entering waypoints at the boundaries of the flight control area; B1-B6 are terminal area boundary port-entering waypoint sets; C1-C4 are sets of starting approach anchor points. The fourth step: and acquiring real-time position data (longitude and latitude) of the aircraft through one of systems (namely one of the systems) such as an air traffic control secondary radar, an air traffic control 4D track, an ADS-B system and an ACARS. The fifth step: when the horizontal distance between the real-time position of the aircraft and any route point in a boundary access way point set of a flight control area, a boundary access way point set of a terminal area and an initial access positioning point set is smaller than a set threshold value for the first time (different threshold values are set for the three sets), judging that the aircraft flies over the route point, namely the triggering time of the real-time position of the aircraft is the time of the aircraft flying over the route point, and entering the next step of calculation, otherwise, neglecting, and the illustration chart is shown in figure 4. And a sixth step: screening continuous 30 historical flight data samples which meet the following two conditions and occur recently; the first condition is as follows: the sample flight was flying past the identified waypoint; and a second condition: the actual landing runway of the sample flight is the same as the expected landing runway of the target flight. The seventh step: and calculating the actual flight time of the flight segments of the screened 30 historical flight data samples, namely the time interval from the flying waypoint to the actual landing, and sequencing from small to large. Eighth step: the left and right 10% of the sorted 30 time interval data samples are removed, the time interval samples of the middle section (80% of the data samples) are averaged, and the average flight time of the flight section is obtained, and an example graph is shown in fig. 5. The ninth step: and accumulating the average flight time of the flight section by using the time of the aircraft flying over the waypoints to obtain the predicted landing time of the flight.

As shown in fig. 1, the calculation of the time for gear-in-gear prediction mainly screens and samples data samples closest to the predicted flight according to the historical big data of the flight, further screens the sample data according to the time distribution and the confidence interval, calculates the average ground sliding time, and accumulates the average flight time of the segment and the average ground sliding time to obtain the predicted time for gear-in-gear. The calculation process is as follows: the first step is as follows: screening meets the following two conditions, namely, a condition I of screening the latest 30 continuous historical flight data samples: the sample flight is the same as the predicted flight arrival station; and a second condition: the actual landing runway of the sample flight is the same as the expected landing runway of the target flight. The second step is that: and calculating the actual ground sliding time of the screened 30 historical flight data samples, namely the time interval from actual landing to actual wheel gear, and sequencing from small to large. The third step: and (3) eliminating 10% of data on the left side and the right side of the sequenced 30 time interval data samples, and averaging the time interval samples of the middle section (80% of data samples) to obtain the average ground sliding time. The fourth step: and accumulating the average flight time of the flight segment and the average ground sliding time by using the time of the aircraft flying over the waypoints to obtain the predicted gear-in time of the flight.

As shown in fig. 1, the invention aims to establish a prediction mechanism and an algorithm for the landing time and the gear-in time of a flight in a flight control area belonging to an airport by performing staged modeling and state recognition on the air operation track of an aircraft and taking big data statistical analysis as a main technical means, so as to provide accurate reference for an airport ground support system. The algorithm design of the invention takes the flight program in the flight control area of the airport as the main modeling reference, adopts the flight historical big data with the highest simulation degree under all factors as the matching prediction sample, and has higher goodness of fit with the real operation. The algorithm system has simple design structure, low requirement on data sources required by calculation, low technical implementation difficulty, capability of quickly realizing product conversion even in small and medium airports with low data conditions, and strong replication popularization. After the algorithm is applied, the arrival information data of the aircraft can be intuitively and effectively obtained through data obtained through the prediction calculation of the landing runway, the prediction calculation of the landing time, the prediction calculation of the gear-loading time and the like, all ground support personnel and equipment of the airport reasonably do advance preparation work according to the predicted landing time of the flight before the flight arrives at a designated parking place, the virtual consumption waiting or the delay in the ground support resources of the airport can be effectively reduced, the utilization rate of the ground support resources is effectively improved, and the method is particularly suitable for accurate decision making when the flight approaches the landing emergency dispatching.

The expert data used in the present invention, shown in FIG. 1, is explained below. Flight control zone boundary: in the application, the boundary of the flight control area refers to the horizontal boundary of an air area covered by a civil aviation air traffic control area-level unit; terminal area boundary: in the present application, the terminal area boundary refers to a horizontal boundary of a flight activity area where a civil aviation air traffic management unit controls and commands aircraft within a certain distance range (about 150 km) from an airport. Starting an approach positioning point: in the present application, the starting approach positioning point refers to a starting point of an approach procedure before the aircraft performs landing. Gear shifting time: the time is the time when the aircraft arrives at the stand and is stably stopped after finishing sliding to the ground.

While there have been shown and described what are at present considered the fundamental principles and essential features of the invention and its advantages, it will be apparent to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, but is capable of other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.

Claims

1. A calculation method for predicting flight arrival time based on air track and big data is characterized by comprising the steps of prediction calculation of a landing runway, prediction calculation of landing time and prediction calculation of gear-loading time; the method comprises the following steps of establishing a route-runway corresponding rule table, obtaining data samples of a plurality of continuous actual landing flights, setting a current same-kind route set predicted landing runway, and replacing the predicted landing runway with the landing runway distributed by a system; recording the coordinates of waypoints related to daily inbound flights of an airport in the boundary of a flight control area, recording the coordinates of waypoints related to the daily inbound flights of the airport in the boundary of a terminal area, forming a set of initial inbound positioning points, acquiring real-time position data of an aircraft, judging waypoints flown by the aircraft, screening a plurality of continuous historical flight data samples which meet the latest occurrence condition, recording the time interval from the plurality of historical flying waypoints to actual landing, acquiring the average flight time of a flight segment, and acquiring the predicted landing time of the flight; the wheel-class loading time prediction calculation comprises the following steps of screening a plurality of recent continuous historical flight data samples meeting conditions, calculating the time interval between the actual landing of the screened plurality of historical flight data samples to the actual wheel class, calculating to obtain average ground sliding time, and calculating to obtain the predicted wheel-class loading time of the flight.

2. The calculation method for predicting the arrival time of the flight based on the air track and the big data as claimed in claim 1, wherein in the floor runway prediction calculation, the established rule table is a many-to-many mapping relation; the obtained data sample is based on the data screened from the airline collection of the same generic type in the historical data of flights; the set predicted landing runway data is derived from three data samples; the data replacing the expected landing runway is derived from the landing runway data that the flight has already allocated by the system.

3. The method of claim 1, wherein in the calculation of the arrival time of the flight, the coordinates of the waypoints of the daily inbound flights to the airport in the boundary of the flight control area comprise longitude and latitude; coordinates of route points of daily inbound flights to the airport in the boundary of the terminal area comprise longitude and latitude; the positioning point set data is derived from a flight program and a navigation chart; the acquired real-time location data includes longitude and latitude.

4. The method of claim 1, wherein the data for determining the departure waypoints of the aircraft is derived from data of a set of real-time positions of the aircraft and the arrival waypoints at the boundary of the flight control area, data of a set of arrival waypoints at the boundary of the terminal area, and data of a set of starting arrival fixes.

5. The method for calculating the predicted flight arrival time based on the air trajectory and the big data as claimed in claim 1, wherein in the landing time prediction calculation, the filtered historical flight data samples need to satisfy a condition one: the sample flight ever flies past the identified waypoint, condition two: the actual landing runway of the sample flight is the same as the expected landing runway of the target flight; the recorded time intervals are sorted from small to large; the average time of flight is obtained by averaging samples of time intervals; the landing time is obtained by accumulating the average flight time of the flight segment by the time of the aircraft flying over the waypoint.

6. The method for calculating flight arrival time based on air tracks and big data as claimed in claim 1, wherein in the calculation of the upper-gear time prediction, the condition for screening is as follows: the sample flight is the same as the predicted flight arrival station, and the condition two is as follows: the actual landing runway of the sample flight is the same as the expected landing runway of the target flight; calculating the sorted time intervals from small to large; the average ground taxi time is derived from a plurality of time interval data samples after sequencing; the predicted time to shift on the flight is derived from the accumulated leg average flight time and average ground taxi time data samples.