CN110135451B - Flight path clustering method based on distance from point to line segment set - Google Patents
Flight path clustering method based on distance from point to line segment set Download PDFInfo
- Publication number
- CN110135451B CN110135451B CN201910235711.XA CN201910235711A CN110135451B CN 110135451 B CN110135451 B CN 110135451B CN 201910235711 A CN201910235711 A CN 201910235711A CN 110135451 B CN110135451 B CN 110135451B
- Authority
- CN
- China
- Prior art keywords
- track
- distance
- jth
- ith
- line segment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G5/00—Traffic control systems for aircraft, e.g. air-traffic control [ATC]
- G08G5/0095—Aspects of air-traffic control not provided for in the other subgroups of this main group
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Image Analysis (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a flight path clustering method based on a point-to-line segment set distance, which measures the similarity degree between flight paths by using a user-defined distance; for each pair of tracks in all the track data, calculating the distance between the tracks, mainly comprising: calculating the distance from each track point to the other track, synthesizing the distances corresponding to all the track points to obtain the one-way distance from one track to the other track, and synthesizing the one-way distances of the two tracks to calculate the two-way distance between the tracks; calculating to obtain a similarity matrix according to the two-way distance matrix; clustering all tracks according to the similarity matrix; and obtaining the clustering label of each flight path. The invention can realize automatic clustering of the flight paths, the measurement of the flight path distance is more reasonable, and the clustering result is more in line with the actual situation. The method has wide application prospect in the fields of air traffic management and airspace planning.
Description
Technical Field
The invention relates to a flight path clustering method, in particular to a flight path clustering method based on a point-to-line segment set distance.
Background
In recent years, economic development promotes rapid increase of travel demands of people, so that traffic flow is more complex and congestion is increasingly serious. To solve this problem, in the intelligent transportation system, a large number of various positioning devices are put into use, such as sensor networks, GPS, mobile phones, etc., which makes it more and more convenient to acquire real-time position information of a mobile object. For air traffic, a large number of surveillance devices, such as radar, ADS-B, etc., provide track data for the aircraft's motion. In the air traffic management, track clustering analysis is carried out on track data describing the actual geographic position of the flight, valuable flight modes are mined, the airspace capacity is improved, and the number of flights is increased on the premise of ensuring the safety.
In the problem of track clustering, the measurement of the distance between tracks is a key point, and is the basis of a clustering method. The track data comprises a series of chronologically arranged points. A simple distance metric is to calculate the sum of the euclidean distances between all pairs of track points. Although the euclidean distance is simple to calculate, it requires that the two tracks have the same length. The requirement of the euclidean distance is difficult to meet because the tracks are usually not fixed in length, i.e. the number of tracks of two tracks will be different. The Hausdorff distance treats tracks as a set of track points, and converts the distance between tracks into the distance between the sets of points, but it does not consider the time sequence of the track points. Dynamic Time Warping (DTW) repeats track points by warping time to expand the track, thereby calculating the distance of two tracks. The repetition of this distance on track points destroys the geometric distribution of the track to some extent. The longest common subsequence (LCSS) distance measures the distance between two waypoints by finding their longest common part, but it requires a parameter to determine whether the two waypoints match, and the setting of the parameter is not simple and requires repeated debugging. For the above reasons, these track distances do not perform well in track clustering.
Disclosure of Invention
The invention aims to provide a track clustering method based on a point-to-line set distance, which fully considers the geometric relationship between a track point and a track line segment, thereby better measuring the track distance.
The technical solution for realizing the purpose of the invention is as follows: a flight path clustering method based on a point-to-line segment set distance comprises the following steps:
and 8, obtaining the clustering label of each track according to the similarity matrix S.
In the invention, the step 2 comprises the following steps:
for ith track TiThe set of track points contained inThe line segment set is obtained by integrationLine segment setThe k line segment ofWith the ith track TiThe kth course point of (1)And the (k + 1) th track pointAs an endpoint;
for jth track TjThe set of track points contained inThe line segment set is obtained by integrationLine segment setThe r-th line segment ofWith jth track TjThe r-th track point in (1)And the r +1 th track pointAs endpoints.
In the present invention, step 3 comprises:
for jth track TjEach track point ofCalculating the r track pointTo ith track TiSet of line segmentsThe distance betweenUsing the same method, for the ith track TiEach track point ofCalculating the kth track pointTo jth track TjSet of line segmentsIs a distance of
In step 3 of the invention, the jth track TjMiddle r track pointTo ith track TiSet of line segmentsThe distance betweenAnd ith track TiMiddle kth track pointTo jth track TjSet of line segmentsIs a distance ofThe calculation process comprises the following steps:
step 3-1, calculating the jth track TjMiddle r track pointTo ith track TiIs the k-th line segment in the line segment setIn a vertical projection ofSetting the jth track TjMiddle r track pointHas coordinates of (x, y), line segmentThe coordinates of the two end points are respectively (x)1,y1) And (x)2,y2),Has the coordinates of (x)p,yp) Then, there are:
Step 3-3, according to the vertical projectionWhether on line segmentAbove, two sets are defined, if at line segmentAbove, is defined asOtherwise defined as a set
Step 3-4, calculating the jth track TjMiddle r track pointTo ith track TiSet of line segmentsThe distance between
Step 3-5, calculating to obtain the ith track T by adopting the methods of the step 3-1 to the step 3-4iMiddle kth track pointTo jth track TjSet of line segmentsIs a distance of
In the invention, the step 4 comprises the following steps: according to distanceCalculate the ith TiTo jth TjOne-way distance D ofhseg(Ti,Tj) According to distanceCalculate the jth TjTo the ith TiOne-way distance D ofhseg(Tj,Ti):
In step 5 of the present invention, the bidirectional distance D is calculated according to the following formulaHseg(Ti,Tj):
DHseg(Ti,Tj)=max{Dhseg(Ti,Tj),Dhseg(Tj,Ti)}。
In step 6 of the present invention, the element D in the ith row and the jth column of the distance matrix DijIs set to DHseg(Ti,Tj)。
In step 7 of the invention, the ith row and jth column elements S of the similarity matrix S are calculated according to the following formulaij:
Sij=exp(-Dij/std(D)),
Where std (D) represents the standard deviation of all elements in matrix D.
In step 8 of the invention, a spectrum clustering algorithm is adopted for the similarity matrix S to obtain clustering labels of each flight path, and the specific steps comprise:
step 8-1, decomposing the S characteristic of the similarity matrix, wherein S is U Λ U-1Wherein U is a matrix composed of eigenvectors, Λ is a diagonal matrix composed of eigenvalues (ref: department of mathematics of university of the same society. linear algebra. fifth edition. Beijing: higher education Press, 2007: 117-;
step 8-2, k eigenvectors corresponding to the minimum k eigenvalues are taken from the matrix U to form a matrix U with N rows and k columnsmWhere k is the number of clusters;
step 8-3, adding UmRegarding the vectors as N k-dimensional vectors, a k-means clustering algorithm is adopted for the N k-dimensional vectors (reference: Zhou Shi machine learning. Beijing: Qing Hua university Press, 2016: 202-;
and 8-4, obtaining clustering labels of the N tracks.
Compared with the prior art, the invention has the following remarkable advantages: the distance between the flight path point and the flight path segment is fully considered, the distance between the point and the segment is expanded, the distance between the point on one flight path and the other flight path is measured, the distance between the flight paths is further calculated, and the geometric characteristics of the flight paths can be better reflected.
Drawings
The foregoing and other advantages of the invention will become more apparent from the following detailed description of the invention when taken in conjunction with the accompanying drawings.
FIG. 1 is a flow chart of the method for clustering tracks based on the distance from a point to a set of line segments according to the present invention.
FIG. 2 is a diagram of radar track data according to an embodiment of the present invention.
FIG. 3a shows the first type of result of the inventive method for track clustering.
FIG. 3b shows the second result of the inventive method for track clustering.
Detailed Description
The invention is further explained below with reference to the drawings and the embodiments.
With reference to fig. 1, the flight path clustering method based on the distance from a point to a line segment set of the present invention includes the following steps:
step 1: for a track data set TR to be clustered, TiI 1, …, N, from which a pair of tracks T is selectediAnd TjAnd ensure the track TiAnd TjWhere i ≠ j.
Step 2: for track TiThe set of track points contained inThe line segment set is obtained by integrationFor track pointAnd to be composed ofAndsegment as end point, for track TjThe same operation is done.
And step 3: for track TjEach track point ofCalculating pointsTo TiLine segment setThe distance betweenDistance between two adjacent platesThe calculation process of (2) is as follows:
(1) calculating pointsTo line segmentIn a vertical projection ofSuppose thatHas the coordinates of (x, y),the coordinates of the two end points are respectively (x)1,y1) And (x)2,y2),Has the coordinates of (x)p,yp) Then there is
And 5: calculating T from the one-way distanceiAnd TjOf (2) a bidirectional distance DHseg(Ti,Tj)
DHseg(Ti,Tj)=max{Dhseg(Ti,Tj),Dhseg(Tj,Ti)}
Step 6: obtaining a distance matrix D according to the two-way distance;
and 7: transforming the matrix D to obtain a similarity matrix S;
and 8: and (3) obtaining the clustering label of each flight path by adopting a spectrum clustering algorithm for the matrix S, wherein the spectrum clustering algorithm comprises the following specific steps:
(1) decomposing the characteristic of the matrix S, wherein S is U Λ U-1;
(2) Taking the least k (k is the number of clustering clusters) of eigenvectors corresponding to the eigenvalues,
(3) will UmConsidering N k-dimensional vectors, and adopting a k-means clustering algorithm for the N k-dimensional vectors;
(4) and obtaining clustering labels of the N tracks.
The present invention will be further explained by embodiments of simulation experiments and evaluation of effects thereof, with reference to fig. 2 to 3a and 3b (in the figures, positive directions of X and Y axes represent east and north, respectively, and distances are in meters).
In the present embodiment, as each curve corresponds to a flight path in fig. 2, the distribution of the flight path data represents two routes, one is straight flight and the other is broken flight. The experimental objective is to cluster the track data into two categories by a clustering method. Fig. 3a and 3b show the result of the flight path clustering of the inventive method, and the experimental data are clustered into two types, which are respectively plotted in fig. 3a and 3b, and the result shows that the two types of flight paths obtained by the inventive method are identical to the two routes of the flight paths, and the inventive method has good performance. In air traffic management, flight radar data between the same city pair sometimes show tracks as shown in fig. 2, however, actual operation data does not provide a clustering result of flight routes. Therefore, by adopting the flight path clustering method provided by the invention, flights flying in straight line and broken line can be distinguished in the post analysis, and the conditions of flight 'getting-off-bend and straight flight' are further analyzed by combining factors such as weather, control and the like at the time, so that the configuration of airspace resources is optimized or a more efficient flight directing strategy is excavated.
The present invention provides a flight path clustering method based on the distance from a point to a line segment set, and a plurality of methods and approaches for implementing the technical scheme are provided, the above description is only a preferred embodiment of the present invention, it should be noted that, for those skilled in the art, a plurality of improvements and embellishments can be made without departing from the principle of the present invention, and these improvements and embellishments should also be regarded as the protection scope of the present invention. All the components not specified in the present embodiment can be realized by the prior art.
Claims (5)
1. A flight path clustering method based on a point-to-line segment set distance is characterized by comprising the following steps:
step 1, acquiring a track data set TR (T) to be clusterediI-1, …, N }, from which a pair of tracks, i.e. the ith track T, is selectediAnd jth track TjAnd i is not equal to j, and N takes the value of a natural number;
step 2, respectively integrating to obtain the ith track TiSet of line segments of (1) and the jth track TjA set of line segments of (a);
step 3, calculating the ith track TiFrom each track point to the jth track TjThe distance of the line segment set of (1) is calculated to the jth track TjFrom each track point to the ith track TiThe distance of the line segment set of (a);
step 4, calculating to obtain the ith track T according to the distance obtained in the step 3iTo jth track TjAnd the jth track TjTo ith track TiThe unidirectional distance of (a);
step 5, calculating the ith track T according to the one-way distanceiAnd jth track TjOf (2) a bidirectional distance DHseg(Ti,Tj);
Step 6, obtaining a distance matrix D according to the two-way distance;
step 7, transforming the matrix D to obtain a similarity matrix S;
step 8, obtaining clustering labels of each track according to the similarity matrix S;
the step 2 comprises the following steps:
for ith track TiThe set of track points contained inThe line segment set is obtained by integrationLine segment setThe k line segment ofWith the ith track TiThe kth course point of (1)And the (k + 1) th track pointAs an endpoint;
for jth track TjThe set of track points contained inThe line segment set is obtained by integrationLine segment setThe r-th line segment ofWith jth track TjThe r-th track point in (1)And the r +1 th track pointAs an endpoint;
the step 3 comprises the following steps:
for jth track TjEach track point ofCalculating the r track pointTo ith track TiSet of line segmentsThe distance betweenUsing the same method, for the ith track TiEach track point ofCalculating the kth track pointTo jth track TjSet of line segmentsIs a distance of
In step 3, the jth track TjMiddle r track pointTo ith track TiSet of line segmentsThe distance betweenAnd ith track TiMiddle kth track pointTo jth track TjSet of line segmentsIs a distance ofThe calculation process comprises the following steps:
step 3-1, calculating the jth track TjMiddle r track pointTo ith track TiIs the k-th line segment in the line segment setIn a vertical projection ofSetting the jth track TjMiddle r track pointHas coordinates of (x, y), line segmentThe coordinates of the two end points are respectively (x)1,y1) And (x)2,y2),Has the coordinates of (x)p,yp) Then, there are:
Step 3-3, according to the vertical projectionWhether on line segmentAbove, two sets are defined, if at line segmentAbove, is defined asOtherwise defined as a set
Step 3-4, calculating the jth track TjMiddle r track pointTo ith track TiSet of line segmentsThe distance between
Step 3-5, calculating to obtain the ith track T by adopting the methods of the step 3-1 to the step 3-4iMiddle kth track pointTo jth track TjSet of line segmentsIs a distance of
Step 4 comprises the following steps: according to distanceCalculate the ith TiTo jth TjOne-way distance D ofhseg(Ti,Tj) According to distanceCalculate the jth TjTo the ith TiOne-way distance D ofhseg(Tj,Ti):
2. The method of claim 1, wherein in step 5, the two-way distance D is calculated according to the following formulaHseg(Ti,Tj):
DHseg(Ti,Tj)=max{Dhseg(Ti,Tj),Dhseg(Tj,Ti)}。
3. The method according to claim 2, wherein in step 6, the element D in the ith row and the jth column of the distance matrix DijIs set to DHseg(Ti,Tj)。
4. The method according to claim 3, wherein in step 7, the i row and j column elements S of the similarity matrix S are calculated according to the following formulaij:
Sij=exp(-Dij/std(D)),
Where std (D) represents the standard deviation of all elements in matrix D.
5. The method according to claim 4, wherein in step 8, a spectral clustering algorithm is applied to the similarity matrix S to obtain a clustering label of each track, and the specific steps include:
step 8-1, decomposing the characteristics of the similarity matrix S, wherein S is U Λ U-1Wherein U is a matrix composed of eigenvectors, and Λ is a diagonal matrix composed of eigenvalues;
step 8-2, k eigenvectors corresponding to the minimum k eigenvalues are taken from the matrix U to form a matrix U with N rows and k columnsmWhere k is the number of clusters;
step 8-3, adding UmRegarding the N k-dimensional vectors, and adopting a k-means clustering algorithm for the N k-dimensional vectors;
and 8-4, obtaining clustering labels of the N tracks.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910235711.XA CN110135451B (en) | 2019-03-27 | 2019-03-27 | Flight path clustering method based on distance from point to line segment set |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910235711.XA CN110135451B (en) | 2019-03-27 | 2019-03-27 | Flight path clustering method based on distance from point to line segment set |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110135451A CN110135451A (en) | 2019-08-16 |
CN110135451B true CN110135451B (en) | 2020-06-26 |
Family
ID=67568935
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910235711.XA Active CN110135451B (en) | 2019-03-27 | 2019-03-27 | Flight path clustering method based on distance from point to line segment set |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110135451B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111398947B (en) * | 2020-04-08 | 2022-01-11 | 成都汇蓉国科微系统技术有限公司 | Pulse Doppler radar clutter suppression method based on track clutter map |
CN113177097B (en) * | 2021-04-16 | 2023-07-25 | 江西航天鄱湖云科技有限公司 | Track start judging method based on attribute clustering and space-time constraint |
CN114065851B (en) * | 2021-11-11 | 2023-05-02 | 中国兵器装备集团自动化研究所有限公司 | Maneuvering target track clustering method and system |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108256285A (en) * | 2018-01-09 | 2018-07-06 | 上海交通大学 | Flight path exception detecting method and system based on density peaks fast search |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9060018B1 (en) * | 2014-02-05 | 2015-06-16 | Pivotal Software, Inc. | Finding command and control center computers by communication link tracking |
CN104898104A (en) * | 2015-06-10 | 2015-09-09 | 中国西安卫星测控中心 | Target combined positioning method based on Euler's distance means clustering |
CN105893621B (en) * | 2016-04-29 | 2019-11-05 | 中国人民解放军海军航空大学 | Goal behavior law mining method based on multidimensional track cluster |
CN106646450B (en) * | 2016-12-29 | 2019-04-12 | 中国人民解放军海军航空大学 | Radar track robust correlating method based on distance substep cluster |
-
2019
- 2019-03-27 CN CN201910235711.XA patent/CN110135451B/en active Active
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108256285A (en) * | 2018-01-09 | 2018-07-06 | 上海交通大学 | Flight path exception detecting method and system based on density peaks fast search |
Also Published As
Publication number | Publication date |
---|---|
CN110135451A (en) | 2019-08-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110135451B (en) | Flight path clustering method based on distance from point to line segment set | |
Mohamed et al. | Accurate real-time map matching for challenging environments | |
CN107679558B (en) | A kind of user trajectory method for measuring similarity based on metric learning | |
Muckell et al. | Algorithms for compressing GPS trajectory data: an empirical evaluation | |
US8825586B2 (en) | Vehicle type recognition at a checkpoint using PCA and BIC | |
CN101551809B (en) | Search method of SAR images classified based on Gauss hybrid model | |
CN110188093A (en) | A kind of data digging system being directed to AIS information source based on big data platform | |
CN111079602A (en) | Vehicle fine granularity identification method and device based on multi-scale regional feature constraint | |
CN109640269B (en) | Fingerprint positioning method based on CSI and time domain fusion algorithm | |
Yan et al. | SeTraStream: semantic-aware trajectory construction over streaming movement data | |
CN108806301B (en) | Automatic identification method for bus information | |
CN104915970A (en) | Multi-target tracking method based on track association | |
CN115049944B (en) | Small sample remote sensing image target detection method based on multitasking optimization | |
Markos et al. | Unsupervised deep learning for GPS-based transportation mode identification | |
CN104820718A (en) | Image classification and searching method based on geographic position characteristics and overall situation vision characteristics | |
Cheng et al. | Vitrack: Efficient tracking on the edge for commodity video surveillance systems | |
WO2023201938A1 (en) | Missing trajectory filling method and system | |
Li et al. | An improved traffic lights recognition algorithm for autonomous driving in complex scenarios | |
Guo et al. | Transportation mode recognition with deep forest based on GPS data | |
Jia et al. | TTSL: An indoor localization method based on Temporal Convolutional Network using time-series RSSI | |
CN111405461A (en) | Wireless indoor positioning method for optimizing equal-interval fingerprint sampling number | |
CN116543603B (en) | Flight path completion prediction method and device considering airspace situation and local optimization | |
CN112101051B (en) | Vehicle positioning method and system based on ultrahigh frequency RFID | |
Cao et al. | Map matching for sparse automatic vehicle identification data | |
Zygouras et al. | Corridor learning using individual trajectories |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |