CN106023655A - Sector air traffic congestion state monitoring method - Google Patents

Abstract

The invention belongs to the field of the air traffic congestion identification technology and discloses a sector air traffic congestion state monitoring method. The method is used for monitoring the air traffic congestion state of the space and the time at each dimension of a point-line-surface-body within a sector. The method comprises the steps of analyzing the structural data of a spatial unit and the data of an operating state to establish a set of air traffic congestion monitoring index system for each dimensional space geared to the needs of sector important points, route segments, flying levels, all sectors and the like, including engine-driven wait times, engine-driven wait behavior ratio, flow ratio, average speed, density, delay, volume flow ratio, the workload of the controlled staff and the like; and providing a sector air traffic congestion state synthetical fuzzy clustering identification method through analyzing the correlation and the difference among air traffic congestion monitoring indexes. According to the technical scheme of the invention, the air traffic congestion monitoring index system is improved, and a sector air traffic congestion state monitoring method is provided. The sector air traffic congestion state monitoring method has the advantages of innovativeness, systematicness and practicability, and closely conforms to the actual requirements of air traffic operation characteristics and air traffic control.

Description

Traffic congestion state monitoring method in a kind of sector null

Technical field

The present invention relates to air traffic planning and management domain, particularly to traffic congestion state monitoring in a kind of sector null Method.

Background technology

Along with air transportation demand quickly increases, air traffic congestion problems becomes increasingly conspicuous, and job of air traffic control is born Lotus sharply increases.Analysing scientifically, accurately identifying air traffic congestion status is effectively to dredge air traffic to block up, reduce aerial friendship The key foundation of logical control workload.Air traffic control generally region or terminal (entering near) regulatory area be divided into two or Person's plural control unit, each unit is referred to as a sector, in order to alleviate regulatory area workload, to improve air traffic Safety and operational efficiency.Therefore, sector is the ultimate unit of air traffic control, and in sector null, traffic congestion state identification is also The important basic work of air traffic control, can be air traffic control, spatial domain systems organization etc. provide technical support and Reference frame, has important practical significance.

In sector null, traffic congestion refers to cause airborne vehicle owing to sector service ability can not meet follow-on mission demand Queue up, aerial wait, the motor-driven air traffic phenomenon being diversion etc..Currently for the research of traffic congestion recognition methods in sector null Less, mainly simply pass judgment on sector congestion state with working capacity threshold value by comparing sector flow, it is impossible to analyze fan The room and times such as Qu Dian, line, body block up feature and airborne vehicle microcosmic run produced by block up feature, largely shadow Its specific aim, science and practicality are rung.Therefore, need badly set up a set of for air traffic operation characteristic, meet control work Make the air traffic that is actually needed to block up monitored data analysis and traffic congestion state monitoring method in sector null, can not only monitor The air traffic congestion status of whole sector, also can the air traffic congestion status of careful assurance point, line, surface various dimensions, for fan District's air traffic safety control, spatial domain programming and planning etc. provide technical support.

Summary of the invention

For overcome above-mentioned existing research deficiency, the present invention establish a set of for air traffic operation characteristic, meet control The air traffic congestion status monitoring index of reality of work needs and computational methods thereof, it is proposed that traffic congestion in a kind of sector null State monitoring method.

Traffic congestion state monitoring method in a kind of sector null that the present invention provides, extracts air traffic running status feature Data, calculate air traffic congestion status monitoring index, use Fuzzy C-Means Cluster Algorithm (FCM) to gather around sector air traffic Stifled state is identified.

Described air traffic congestion status monitoring index, including motor-driven wait number of times, motor-driven wait behavior ratio, flow Ratio, average speed, speed ratio, density, it is delayed, holds flow ratio, ATC controller workload.

In described monitoring index, motor-driven wait number of timesParticularly as follows: control sector t₀-t₁In monitoring period of time, spatial domain unit Airborne vehicle there is the airborne vehicle quantity of motor-driven behavior of being diversion, wait in the air, in order to the directly perceived crowded impact weighing spatial domain unit Degree

$n^{t_0-t_1}=n_{waiting}^{t_0-t_1}+n_{maneuver}^{t_0-t_1}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, airborne vehicle that spatial domain unit monitors occurs aerial etc. Treat the airborne vehicle quantity of behavior,Represent control sector t₀-t₁In monitoring period of time, the generation that spatial domain unit monitors is motor-driven It is diversion the airborne vehicle quantity of behavior；

Described motor-driven wait behavior ratioParticularly as follows: control sector t₀-t₁In monitoring period of time, the airborne vehicle of spatial domain unit Occur motor-driven to be diversion, the quantity ratio of the behavior such as wait in the air, in order to weigh such airborne vehicle by crowded influence degree

${\mathrm{\λ}}^{t_0-t_1}=\frac{n^{t_0-t_1}}{N^{t_0-t_1}}$

Wherein,Represent control sector t₀-t₁The row such as in monitoring period of time, spatial domain unit occurs motor-driven to be diversion, wait in the air For airborne vehicle quantity,Represent control sector t₀-t₁In monitoring period of time, all airborne vehicle numbers that spatial domain unit monitors Amount；

Described flow-rate ratioParticularly as follows: control sector t₀-t₁In monitoring period of time, the flow of spatial domain unit and equal duration The ratio of interior history average discharge, in order to reflect the smooth and easy state of aircraft

${\mathrm{\σ}}^{t_0-t_1}=\frac{N^{t_0-t_1}}{{\stackrel{\‾}{N}}^{\mathrm{\Δ}t}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the flow that spatial domain unit monitors,Represent all In Historical Monitoring data, duration Δ t (Δ t=t₁-t₀) the history average discharge of interior spatial domain unit；

Described average speedParticularly as follows: control sector t₀-t₁In monitoring period of time, putting down of all airborne vehicles of spatial domain unit All flight speeds, average speed is the biggest, and the transport air flow of reflection is more for unobstructed

${\stackrel{\‾}{V}}^{t_0-t_1}=\frac{\underset{i=1}{\overset{N^{t_0-t_1}}{\Σ}}{\stackrel{\‾}{v}}_i}{N^{t_0-t_1}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the airborne vehicle quantity that spatial domain unit monitors,Represent Airborne vehicle i is at monitoring period of time t₀-t₁Interior average speed；

Speed ratioParticularly as follows: control sector t₀-t₁In monitoring period of time, the unit spatial domain all airborne vehicles of unit average The ratio of flight speed and the historical average speeds in equal duration, in order to reflect the degree of mobility of transport air flow

${\mathrm{\α}}^{t_0-t_1}=\frac{{\stackrel{\‾}{V}}^{t_0-t_1}}{{\stackrel{\‾}{V}}^{\′\mathrm{\Δ}t}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the average flight speed of all airborne vehicles of spatial domain unit,Represent in all Historical Monitoring data, duration Δ t (Δ t=t₁-t₀) historical average speeds of interior spatial domain unit；

Described densityParticularly as follows: control sector t₀-t₁In monitoring period of time, the airborne vehicle comprised in the unit of unit spatial domain Par, the dense degree of airborne vehicle in the unit of expression spatial domain

$k^{t_1-t_0}=\frac{\underset{t=t_0}{\overset{t_1}{\Σ}}\frac{N^t}{P}}{t_1-t_0}$

Wherein, N^tRepresenting control sector instantaneous moment t, the airborne vehicle quantity that spatial domain unit monitors, P represents a certain monitoring Interval, can be course line segment length, sector area or volume, t₀、t₁Respectively monitor initial time；

Described delayControl sector t₀-t₁In monitoring period of time, the airborne vehicle actually used time by spatial domain unit With the difference of planned time, in order to reflect airborne vehicle performance in operation situation, mean delay is the longest, illustrates that air traffic is more for crowded

$D^{t_0-t_1}=t_i-t_i^p$

Wherein, t_iRepresent control sector t₀-t₁In monitoring period of time, the airborne vehicle actually used time by spatial domain unit, Represent control sector t₀-t₁In monitoring period of time, the airborne vehicle planned time by spatial domain unit, ifThen

Described appearance flow ratioParticularly as follows: control sector t₀-t₁In monitoring period of time, the actual flow of spatial domain unit and operation The ratio of capacity, in order to reflect the degree of saturation of spatial domain unit

$R^{t_0-t_1}=\frac{N^{t_0-t_1}}{C}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the actual flow that spatial domain unit monitors, C represents empty The working capacity value of territory unit；

Described ATC controller workload W (t₀-t₁) particularly as follows: controller for alleviate the pressure that affords and complete objective The time length that the requirement of task is consumed

$W(t_0-t_1)=W^{t_0-t_1,com}+W^{t_0-t_1,incom}+W^{t_0-t_1,thk}$

Wherein, W (t₀-t₁) represent monitoring period of time t₀-t₁Interior total working load；Represent monitoring period of time t₀-t₁In The communication work load of controller；Represent monitoring period of time t₀-t₁The non-communicating workload of interior controller, as filled in The loads such as electrical steps list, screen operator；Represent monitoring period of time t₀-t₁The thinking load of interior controller.

Described extraction air traffic running status characteristic particularly as follows:

Step one, extracts spatial domain cellular construction data and running status basic data, including sector area, air route structure, Each leg length, vital point position and coordinate, the distribution of corridor, sector mouth and coordinate, in sector the position of any time every frame airborne vehicle Put, speed, course, highly, original base, land airport, airborne vehicle and estimated the data such as point, set up basic database；

Step 2, analyzes unitary space, spatial domain structure, extracts each vital point, course line section, height layer and whole sector respectively Various types of data needed for crowded monitoring, in calculating monitoring timeslice, sector " point-line-surface body " air traffic is blocked up monitoring index value, Form air traffic congestion status monitored data analysis.

It is right that traffic congestion state in sector null is identified referring specifically to by described employing Fuzzy C-Means Cluster Algorithm (FCM) The air traffic congestion status of " point-line-surface body " each Spatial Dimension of whole sector is identified, and refers specifically to:

Step 3, uses correlation analysis method respectively " point-line-surface body " each monitoring index to be carried out correlation analysis, sieve Except the index that dependency is bigger, calculate the correlation coefficient r between each index_jk:

$r_{jk}=\frac{\underset{i=1}{\overset{I}{\Σ}}(x_{ij}-\stackrel{\‾}{x_j})(x_{ik}-\stackrel{\‾}{x_k})}{\sqrt{\underset{i=1}{\overset{I}{\Σ}}{(x_i-\stackrel{\‾}{x_j})}^2}\sqrt{\underset{i=1}{\overset{I}{\Σ}}{(x_{ik}-\stackrel{\‾}{x_k})}^2}}$

Wherein, x_ij、x_ikIt is respectively jth, the i-th timed sample sequence value of kth index,It is respectively jth, k The sample average of individual index, I is number of samples；

Step 4, degree of membership based on fuzzy mathematics is theoretical, uses Fuzzy C-Means Cluster Algorithm (FCM) important to sector The air traffic of point, course line section, height layer and the full sector monitoring index that blocks up clusters respectively, and show that corresponding monitoring refers to Mark parameter；

Step 5, the aerial friendship in each vital point of Real-time Collection, each course line section, each height layer and full sector unit interval sheet The logical monitoring index that blocks up, uses Euclidean distance to differentiate the acquisition index distance size away from different congestion state index parameters, identifies This period each vital point, each course line section, each height layer and the air traffic congestion status of full sector.

Described step 4 includes:

Step 4.1: normalization achievement data

$x_{ij}^{*}=\frac{x_{ij}-x_j^{\min }}{x_j^{\max }-x_j^{\min }},$

Wherein, x_ijFor the i-th time sequential value of jth index,For the sample minimum of jth index,For The sample maximum of jth index；

Step 4.2: initialize subordinated-degree matrix U

Set up initial subordinated-degree matrix U, order

And

Wherein, u_niRepresenting that i-th index sample sequence is under the jurisdiction of the degree of the n-th safe class classification, N is congestion state Grade separation number, the present invention all takes N=4, i.e. vital point congestion state grade, course line section congestion state grade and the crowded shape in sector State grade be divided into freedom, pass unimpeded, four classes crowded, congested, it may be assumed that

Class={ " freely ", " passing unimpeded ", " crowded ", " congested " }

Step 4.3: calculate the cluster centre of N number of classification

$V_n^{*}=\frac{\underset{i=1}{\overset{I}{\Σ}}{u}_{ni}^m{X}_i}{\underset{i=1}{\overset{I}{\Σ}}{u}_{ni}^m},n=1,2,...,N$

Cluster centre matrixWherein, V₁ ^*,It is free state respectively, passes unimpeded The characteristic vector of state, congestion state and congestion state；M is Weighted Index；

Step 4.4: calculating FCM cost function J:

$J(U,c_1,c_2,...c_n)=\underset{n=1}{\overset{N}{\Σ}}{J}_n=\underset{n=1}{\overset{N}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}{u}_{ni}^m{d}_{ni}^2$

Wherein d_niBeing the Euclidean distance between the cluster centre of the n-th classification and i-th data, m is Weighted Index, J's Compactness in value reflection class, J is the least, shows to cluster and more compacts；

Cost function value J of front and back twice is compared, if cost function value knots modification Δ J is less than threshold epsilon, then turns Step 4.6, otherwise goes to step 4.5；

Step 4.5: renewal Subject Matrix U:

$u_{ni}=\frac{1}{\underset{k=1}{\overset{N}{\Σ}}{\left(\frac{d_{ni}}{d_{ki}}\right)}^{\frac{2}{m-1}}}$

Repeat step 4.3, step 4.4；

Step 4.6: output subordinated-degree matrix U, uses the principle of maximum membership degree that each data point is carried out congestion state and returns Class, and export cluster centre matrixCluster centre carrying out renormalization and obtains actual congestion indication parameter, renormalization is public Formula is as follows:

$V_{nj}=V_{nj}^{*}\×(x_j^{\max }-x_j^{\min })+x_j^{\min }$

Wherein,For the cluster centre of n congestion state jth index,For the sample minimum of jth index,Sample maximum for jth index.

The present invention uses above technical scheme compared with prior art, has following technical effect that

Traffic congestion state monitoring method in a kind of sector null provided by the present invention, (1) establishes a set of air traffic Congestion status monitoring index and computational methods thereof, can be used for monitoring " point-line-surface body " various dimensions space, sector and aerial in the time Traffic congestion state, it is possible to preferably embody feature and air traffic control actual demand that air traffic runs, have stronger Novelty, comprehensive and practicality.

(2) propose traffic congestion state monitoring method in a kind of sector null, cluster based on fuzzy mathematics theory and FCM and calculate Method carries out classification process and comprehensive quantification analysis to traffic congestion state in sector null, has appraisal procedure practicality, appraisal procedure Simply, the feature such as assessment result quantization.

Accompanying drawing explanation

Fig. 1 is traffic congestion state monitoring method flow chart in sector null；

Fig. 2 is air traffic congestion status monitored data analysis figure.

Detailed description of the invention

The present invention provides traffic congestion state monitoring method in a kind of sector null, for making the purpose of the present invention, technical scheme And effect is clearer, clearly, and referring to the drawings and give an actual example that the present invention is described in more detail.Should be appreciated that herein Described being embodied as, only in order to explain the present invention, is not intended to limit the present invention.

Fig. 1 is the flow chart of the present invention.In this sector null, traffic congestion state recognition methods includes implementing in detail below Step:

Step one: extract air traffic running status characteristic.

Air traffic running status characteristic is divided into spatial domain cellular construction data and running status basic data, and spatial domain is single Meta structure data mainly include that sector area, air route structure, each leg length, vital point position and coordinate, corridor, sector mouth divide Cloth and coordinate, can obtain from " internal navigation compilation of data (NAIP) " and GIS-Geographic Information System (GIS).Spatial domain unit runs Status data mainly includes the position (longitude and latitude) of any time every frame airborne vehicle, speed in sector ATC controller workload, sector Degree, course, highly, original base, land airport, airborne vehicle and estimated the data such as some time, wherein, sector controller works negative Lotus can obtain corresponding data, the position (longitude and latitude) of any time every frame airborne vehicle, speed, boat in sector by investigation on the spot To, highly, original base, land airport, airborne vehicle and estimated the data such as some time, can be by obtaining the radar of corresponding time period Flight path, flight planning, navigator report data acquisition.

Step 2: analyze unitary space, spatial domain structure, extract each vital point, course line section, height layer and whole sector respectively Various types of data needed for crowded monitoring, calculates monitoring timeslice (such as 15min) interior sector " point-line-surface body " air traffic and blocks up prison Survey desired value, form traffic congestion state monitored data analysis, such as accompanying drawing 2.

It is respectively directed to vital point, course line section, height layer, full sector, calculates gathering around in corresponding unit interval sheet (15min) Stifled monitoring index:

(1) vital point congestion status monitoring index

The most motor-driven wait number of timesControl sector t₀-t₁In monitoring period of time, attached at the vital point i such as holding point, reporting point Near occur motor-driven to be diversion, the airborne vehicle quantity of the behavior such as wait in the air, in order to the crowded influence degree weighing spatial domain unit directly perceived.

$n_i^{t_0-t_1}=n_{i,waiting}^{t_0-t_1}+n_{i,maneuver}^{t_0-t_1}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, near vital point i, there is the aerial boat waiting behavior Pocket quantity,Represent control sector t₀-t₁In monitoring period of time, near vital point i, there is the aviation of motor-driven behavior of being diversion Device quantity, the motor-driven behavior of being diversion occurred between vital point i-1 and vital point i is designated as the motor-driven behavior of being diversion of vital point i.

The most motor-driven wait behavior ratioControl sector t₀-t₁In monitoring period of time, at the vital point i such as holding point, reporting point Near occur motor-driven to be diversion, the airborne vehicle quantity ratio of the behavior such as wait in the air, affected journey in order to weigh such airborne vehicle by crowded Degree.

${\mathrm{\λ}}_i^{t_0-t_1}=\frac{n_i^{t_0-t_1}}{N_i^{t_0-t_1}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, occur near the vital point i such as holding point, reporting point Motor-driven be diversion, the airborne vehicle quantity of the behavior such as wait in the air,Represent control sector t₀-t₁In monitoring period of time, pass through vital point The airborne vehicle quantity of i.

3. flow-rate ratioControl sector t₀-t₁In monitoring period of time, by the flow of vital point i and going through in equal duration The ratio of history average discharge, in order to reflect the smooth and easy state of the aircraft of vital point.

${\mathrm{\σ}}_i^{t_0-t_1}=\frac{N_i^{t_0-t_1}}{\underset{k=1}{\overset{M}{\Σ}}{N}_i^{{\mathrm{\Δt}}_k}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, by the flow of leg j,Represent all history In Monitoring Data, duration Δ t (Δ t=t₁-t₀) timeslice Δ t_kThe interior flow by leg j, M express time sheet number, i.e. The M Δ total duration of t=Historical Monitoring.

(2) course line section congestion status monitoring index

1. speed ratioParticularly as follows: control sector t₀-t₁In monitoring period of time, the average flight of all airborne vehicles on the j of leg The ratio of speed and the historical average speeds in equal duration, in order to reflect the degree of mobility of transport air flow

${\mathrm{\α}}_j^{t_0-t_1}=\frac{{\stackrel{\‾}{V}}_j^{t_0-t_1}}{\underset{k=1}{\overset{M}{\Σ}}{V}_j^{\′{\mathrm{\Δt}}_k}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the average flight speed of all airborne vehicles of spatial domain unit,Represent in all Historical Monitoring data, duration Δ t (Δ t=t₁-t₀) timeslice Δ t_kThe interior speed by leg j, M table Show timeslice number, i.e. the M Δ total duration of t=Historical Monitoring.

2. flow-rate ratioControl sector t₀-t₁In monitoring period of time, by the flow of leg j and the history in equal duration The ratio of average discharge, in order to reflect the smooth and easy state of the aircraft on leg.

${\mathrm{\σ}}_j^{t_0-t_1}=\frac{N_j^{t_0-t_1}}{\underset{k=1}{\overset{M}{\Σ}}{N}_j^{{\mathrm{\Δt}}_k}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, by the flow of leg j,Represent all history In Monitoring Data, duration Δ t (Δ t=t₁-t₀) timeslice Δ t_kThe interior flow by leg j, M express time sheet number, i.e. The M Δ total duration of t=Historical Monitoring.

3. densityControl sector t₀-t₁In monitoring period of time, the airborne vehicle par in the j unit length of leg, table Show airborne vehicle dense degree on the j of leg.

$k_j^{t_1-t_0}=\frac{\underset{t=t_0}{\overset{t_1}{\Σ}}\frac{N_j^t}{L_j}}{t_1-t_0}$

Wherein,Represent control sector instantaneous moment t, the airborne vehicle quantity of monitoring, L on the j of leg_jRepresent the length of leg j Degree (km), t₀、t₁Respectively monitor initial time.

4. it is delayedControl sector t₀-t₁In monitoring period of time, airborne vehicle passes through actually used time and the plan of leg j The difference of time, in order to reflect airborne vehicle performance in operation situation, mean delay is the longest, illustrates that air traffic is more for crowded.

$D_j^{t_0-t_1}=\frac{\underset{i=1}{\overset{N_j^{t_0-t_1}}{\Σ}}{t}_i-t_i^p}{N_j^{t_0-t_1}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the flow that leg j monitors, t_iRepresent control sector t₀-t₁In monitoring period of time, airborne vehicle i passes through the leg j actually used time,Represent control sector t₀-t₁In monitoring period of time, aviation The device i planned time by leg j, ifThen

(3) height layer congestion status monitoring index

The most motor-driven wait number of timesControl sector t₀-t₁In monitoring period of time, the airborne vehicle taking height layer l occurs motor-driven Be diversion, the airborne vehicle quantity of the behavior such as wait in the air, in order to the crowded influence degree weighing spatial domain unit directly perceived.

$n_l^{t_0-t_1}=n_{l,waiting}^{t_0-t_1}+n_{l,maneuver}^{t_0-t_1}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, there is aerial wait in the airborne vehicle taking height layer l The airborne vehicle quantity of behavior,Represent control sector t₀-t₁In monitoring period of time, take the airborne vehicle generation machine of height layer l The airborne vehicle quantity of dynamic behavior of being diversion.

The most motor-driven wait behavior ratioControl sector t₀-t₁In monitoring period of time, take the airborne vehicle generation machine of height layer l Dynamic be diversion, the airborne vehicle quantity ratio of the behavior such as wait in the air, in order to weigh such airborne vehicle by crowded influence degree.

${\mathrm{\λ}}_1^{t_0-t_1}=\frac{n_l^{t_0-t_1}}{N_l^{t_0-t_1}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, take the airborne vehicle of height layer l occur motor-driven to be diversion, The airborne vehicle quantity of the behaviors such as aerial wait,Represent control sector t₀-t₁In monitoring period of time, take height layer l and carry out flat flying Airborne vehicle quantity.

3. flow-rate ratioControl sector t₀-t₁In monitoring period of time, take height layer l carry out the flat flow flown with time equal The ratio of the history average discharge in length, in order to reflect the smooth and easy state of the aircraft of height layer.

${\mathrm{\σ}}_l^{t_0-t_1}=\frac{N_l^{t_0-t_1}}{\underset{k=1}{\overset{M}{\Σ}}{N}_l^{{\mathrm{\Δt}}_k}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, take height layer l and carry out the flat flow flown,Table Show in all Historical Monitoring data, duration Δ t (Δ t=t₁-t₀) timeslice Δ t_kInside take height layer l and carry out the flat stream flown Amount, M express time sheet number, i.e. the M Δ total duration of t=Historical Monitoring.

(4) sector congestion status monitoring index

The most motor-driven wait number of timesControl sector t₀-t₁In monitoring period of time, occur in control sector k motor-driven to be diversion, empty The airborne vehicle quantity of the behaviors such as middle wait, in order to the crowded influence degree weighing spatial domain unit directly perceived.

$n_k^{t_0-t_1}=n_{k,waiting}^{t_0-t_1}+n_{k,maneuver}^{t_0-t_1}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, in control sector k, there is the aerial boat waiting behavior Pocket quantity,Represent control sector t₀-t₁In monitoring period of time, in control sector k, there is the aviation of motor-driven behavior of being diversion Device quantity.

The most motor-driven wait behavior ratioControl sector t₀-t₁In monitoring period of time, occur in control sector k motor-driven to be diversion, The airborne vehicle quantity ratio of the behaviors such as aerial wait, in order to weigh such airborne vehicle by crowded influence degree.

${\mathrm{\λ}}_k^{t_0-t_1}=\frac{n_k^{t_0-t_1}}{N_k^{t_0-t_1}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, occur in control sector k motor-driven to be diversion, wait in the air Deng the airborne vehicle quantity of behavior,Represent control sector t₀-t₁In monitoring period of time, airborne vehicle quantity in control sector k.

3. average speedControl sector t₀-t₁In monitoring period of time, the average flight of all airborne vehicles in control sector k Speed, average speed is the biggest, and the transport air flow of reflection is more for unobstructed.

${\stackrel{\‾}{V}}_k^{t_0-t_1}=\frac{1}{N_k^{t_0-t_1}\·M_k^{t_0-t_1}}\underset{i=1}{\overset{N_k^{t_0-t_1}}{\Σ}}\underset{j=1}{\overset{M_i^{t_0-t_1}}{\Σ}}{v}_i^{j,t_0-t_1}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the ground of airborne vehicle i in the control sector k of radar record Speed,Represent control sector t₀-t₁In monitoring period of time, the radar record number of times of airborne vehicle i,Represent control sector t₀-t₁ In monitoring period of time, airborne vehicle quantity in the control sector k of radar record.

4. densityControl sector t₀-t₁In monitoring period of time, the airborne vehicle comprised in control sector k unit are is average Quantity, the dense degree of airborne vehicle in expression sector k.

$k_k^{t_1-t_0}=\frac{\underset{t=t_0}{\overset{t_1}{\Σ}}\frac{N_k^t}{S_k}}{t_1-t_0}$

Wherein,The airborne vehicle quantity monitored in representing control sector instantaneous moment t, control sector k, S_kRepresent sector k Area, t₀、t₁Respectively monitor initial time.

5. it is delayedControl sector t₀-t₁In monitoring period of time, airborne vehicle by actually used time of control sector k with The difference of planned time, in order to reflect airborne vehicle performance in operation situation, mean delay is the longest, illustrates that air traffic is more for crowded.

$D_k^{t_0-t_1}=\frac{\underset{i=1}{\overset{N_k^{t_0-t_1}}{\Σ}}{t}_i-t_i^p}{N_k^{t_0-t_1}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, airborne vehicle quantity, t in control sector k_iRepresent control Sector t₀-t₁In monitoring period of time, the airborne vehicle i actually used time by sector k,Represent control sector t₀-t₁Monitoring period of time In, the airborne vehicle i planned time by sector k, ifThen

6. flow ratio is heldControl sector t₀-t₁In monitoring period of time, the ratio of the actual flow of control sector k and working capacity In order to reflect the degree of saturation of spatial domain unit.

$R_k^{t_0-t_1}=\frac{N_k^{t_0-t_1}}{C_k}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, airborne vehicle quantity, C in control sector k_kRepresent control The working capacity value of sector k.

7. ATC controller workload W_k(t₀-t₁): control sector t₀-t₁In monitoring period of time, in control sector k, controller is used for Alleviate the pressure afforded and complete the time length that the requirement of objective task is consumed.

$W_k(t_0-t_1)=W_k^{t_0-t_1,com}+W_k^{t_0-t_1,incom}+W_k^{t_0-t_1,thk}$

Wherein, W_k(t₀-t₁) represent monitoring period of time t₀-t₁The total working load (unit: second) of interior sector k；Represent Monitoring period of time t₀-t₁The communication work load (unit: second) of interior sector k controller；Represent monitoring period of time t₀-t₁Interior wing The non-communicating workload (unit: second) of district k controller, as filled in the loads such as electrical steps list, screen operator；Represent Monitoring period of time t₀-t₁The thinking load (unit: second) of interior sector k controller.

Step 3: use correlation analysis method respectively " point-line-surface body " each monitoring index to be carried out correlation analysis, sieve Except the index that dependency is bigger.The correlation analysis method used is as follows:

Calculate air traffic to block up monitoring index, obtain desired value sequence samples X:

X=(X₁,X₂,…,X_I)^T

X_i=(x_i1,x_i2,…,x_ij), i=1,2 ..., I

Wherein, X_iFor i-th time samples sequence, x_ijFor the i-th timed sample sequence value of jth index, I is sample Number.

Calculate the correlation coefficient r between each index_jk:

$r_{jk}=\frac{\underset{i=1}{\overset{I}{\Σ}}(x_{ij}-\stackrel{\‾}{x_j})(x_{ik}-\stackrel{\‾}{x_k})}{\sqrt{\underset{i=1}{\overset{I}{\Σ}}{(x_i-\stackrel{\‾}{x_j})}^2}\sqrt{\underset{i=1}{\overset{I}{\Σ}}{(x_{ik}-\stackrel{\‾}{x_k})}^2}}$

Wherein, x_ij、x_ikIt is respectively jth, the i-th timed sample sequence value of kth index,It is respectively jth, k The sample average of individual index, I is number of samples.

Step 4: degree of membership based on fuzzy mathematics is theoretical, uses Fuzzy C-Means Cluster Algorithm (FCM) important to sector The air traffic of point, course line section, height layer and the full sector monitoring index that blocks up clusters respectively, and show that corresponding monitoring refers to Mark parameter.Fuzzy C-Means Cluster Algorithm (FCM) concrete steps include:

(1) normalization achievement data:

Monitoring index value of the air traffic calculated in step 3 being blocked up is normalized, order

$x_{ij}^{*}=\frac{x_{ij}-x_j^{\min }}{x_j^{\max }-x_j^{\min }}$

Wherein, x_ijFor the i-th time sequential value of jth index,For the sample minimum of jth index,For The sample maximum of jth index.

(2) subordinated-degree matrix U is initialized:

Set up initial subordinated-degree matrix U, order

And

Wherein, u_niRepresenting that i-th index sample sequence is under the jurisdiction of the degree of the n-th safe class classification, N is congestion state Grade separation number, the present invention all takes N=4, i.e. vital point congestion state grade, course line section congestion state grade and the crowded shape in sector State grade is divided into " freely ", " passing unimpeded ", " crowded ", " congested " four class, it may be assumed that

Class={ " freely ", " passing unimpeded ", " crowded ", " congested " }

(3) cluster centre of N number of classification is calculated

$V_n^{*}=\frac{\underset{i=1}{\overset{I}{\Σ}}{u}_{ni}^m{X}_i}{\underset{i=1}{\overset{I}{\Σ}}{u}_{ni}^m},n=1,2,...,N$

Cluster centre matrixWherein, V₁ ^*,It is free state respectively, passes unimpeded The characteristic vector of state, congestion state and congestion state；M is Weighted Index, and its value size is just affecting fuzzy clustering result Really property and clustering performance, the present invention takes m=2.

(4) FCM cost function J is calculated:

$J(U,c_1,c_2,...c_n)=\underset{n=1}{\overset{N}{\Σ}}{J}_n=\underset{n=1}{\overset{N}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}{u}_{ni}^m{d}_{ni}^2$

Wherein d_niBeing the Euclidean distance between the cluster centre of the n-th classification and i-th data, m is Weighted Index, J's Compactness in value reflection class, J is the least, shows to cluster and more compacts.

Cost function value J of front and back twice is compared, if cost function value knots modification Δ J is less than threshold epsilon, then turns Step (6), otherwise goes to step (5).

(5) Subject Matrix U is updated:

$u_{ni}=\frac{1}{\underset{k=1}{\overset{N}{\Σ}}{\left(\frac{d_{ni}}{d_{ki}}\right)}^{\frac{2}{m-1}}}$

Wherein, d_niBeing the Euclidean distance between the cluster centre of the n-th classification and i-th data, m is Weighted Index.Weight Multiple step (3), (4).

(6) output subordinated-degree matrix U, uses the principle of maximum membership degree that each data point is carried out congestion state classification, and Output cluster centre matrixCluster centre is carried out renormalization and obtains traffic congestion monitoring index parameter in sector null.Instead Normalization formula is as follows:

$V_{nj}=V_{nj}^{*}\×(x_j^{\max }-x_j^{\min })+x_j^{\min }$

Wherein,For the cluster centre of n congestion state jth index,For the sample minimum of jth index,Sample maximum for jth index.

Step 5: the aerial friendship in each vital point of Real-time Collection, each course line section, each height layer and full sector unit interval sheet The logical monitoring index that blocks up, uses Euclidean distance to differentiate the acquisition index distance size away from different congestion state index parameters, identifies This period each vital point, each course line section, each height layer and the air traffic congestion status of full sector.

Claims (6)

1. traffic congestion state monitoring method in a sector null, it is characterised in that extract air traffic running status characteristic number According to, calculate air traffic congestion status monitoring index, use Fuzzy C-Means Cluster Algorithm (FCM) to traffic congestion in sector null State is identified.

Traffic congestion state monitoring method in a kind of sector null the most according to claim 1, it is characterised in that described sky Middle traffic congestion state monitoring index, including motor-driven wait number of times, motor-driven wait behavior ratio, flow-rate ratio, average speed, speed Ratio, density, it is delayed, holds flow ratio, ATC controller workload.

Traffic congestion state monitoring method in a kind of sector null the most according to claim 1, it is characterised in that

In described monitoring index, motor-driven wait number of timesParticularly as follows: control sector t₀-t₁In monitoring period of time, the boat of spatial domain unit There is the airborne vehicle quantity of motor-driven behavior of being diversion, wait in the air in pocket, in order to the crowded influence degree weighing spatial domain unit directly perceived

$n^{t_0-t_1}=n_{waiting}^{t_0-t_1}+n_{maneuver}^{t_0-t_1}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the airborne vehicle that spatial domain unit monitors occurs to wait row in the air For airborne vehicle quantity,Represent control sector t₀-t₁In monitoring period of time, generation that spatial domain unit monitors is motor-driven is diversion The airborne vehicle quantity of behavior；

Described motor-driven wait behavior ratioParticularly as follows: control sector t₀-t₁In monitoring period of time, the airborne vehicle of spatial domain unit occurs Motor-driven be diversion, the quantity ratio of the behavior such as wait in the air, in order to weigh such airborne vehicle by crowded influence degree

${\mathrm{\λ}}^{t_0-t_1}=\frac{n^{t_0-t_1}}{N^{t_0-t_1}}$

Wherein,Represent control sector t₀-t₁The behaviors such as in monitoring period of time, spatial domain unit occurs motor-driven to be diversion, wait in the air Airborne vehicle quantity,Represent control sector t₀-t₁In monitoring period of time, all airborne vehicle quantity that spatial domain unit monitors；

Described flow-rate ratioParticularly as follows: control sector t₀-t₁In monitoring period of time, in the flow of spatial domain unit and equal duration The ratio of history average discharge, in order to reflect the smooth and easy state of aircraft

${\mathrm{\σ}}^{t_0-t_1}=\frac{N^{t_0-t_1}}{{\stackrel{\‾}{N}}^{\mathrm{\Δ}t}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the flow that spatial domain unit monitors,Represent all history In Monitoring Data, duration Δ t (Δ t=t₁-t₀) the history average discharge of interior spatial domain unit；

Described average speedParticularly as follows: control sector t₀-t₁In monitoring period of time, averagely flying of all airborne vehicles of spatial domain unit Line speed, average speed is the biggest, and the transport air flow of reflection is more for unobstructed

${\stackrel{\‾}{V}}^{t_0-t_1}=\frac{\underset{i=1}{\overset{N^{t_0-t_1}}{\Σ}}{\stackrel{\‾}{v}}_i}{N^{t_0-t_1}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the airborne vehicle quantity that spatial domain unit monitors,Represent aviation Device i is at monitoring period of time t₀-t₁Interior average speed；

Described speed ratioParticularly as follows: control sector t₀-t₁In monitoring period of time, the unit spatial domain all airborne vehicles of unit average The ratio of flight speed and the historical average speeds in equal duration, in order to reflect the degree of mobility of transport air flow

${\mathrm{\α}}^{t_0-t_1}=\frac{{\stackrel{\‾}{V}}^{t_0-t_1}}{{\stackrel{\‾}{V}}^{\′\mathrm{\Δ}t}}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the average flight speed of all airborne vehicles of spatial domain unit, Represent in all Historical Monitoring data, duration Δ t (Δ t=t₁-t₀) historical average speeds of interior spatial domain unit；

Described densityParticularly as follows: control sector t₀-t₁In monitoring period of time, the airborne vehicle average comprised in the unit of unit spatial domain Amount, the dense degree of airborne vehicle in the unit of expression spatial domain

$k^{t_1-t_0}=\frac{\underset{t=t_0}{\overset{t_1}{\Σ}}\frac{N^t}{P}}{t_1-t_0}$

Wherein, N^tRepresenting control sector instantaneous moment t, the airborne vehicle quantity that spatial domain unit monitors, P represents a certain monitoring interval, Can be course line segment length, sector area or volume, t₀、t₁Respectively monitor initial time；

Described delayControl sector t₀-t₁In monitoring period of time, airborne vehicle passes through actually used time and the meter of spatial domain unit The difference of the time of drawing, in order to reflect airborne vehicle performance in operation situation, mean delay is the longest, illustrates that air traffic is more for crowded

$D^{t_0-t_1}=t_i-t_i^p$

Wherein, t_iRepresent control sector t₀-t₁In monitoring period of time, the airborne vehicle actually used time by spatial domain unit,Represent Control sector t₀-t₁In monitoring period of time, the airborne vehicle planned time by spatial domain unit, ifThen

Described appearance flow ratioParticularly as follows: control sector t₀-t₁In monitoring period of time, the actual flow of spatial domain unit and working capacity Ratio, in order to reflect the degree of saturation of spatial domain unit

$R^{t_0-t_1}=\frac{N^{t_0-t_1}}{C}$

Wherein,Represent control sector t₀-t₁In monitoring period of time, the actual flow that spatial domain unit monitors, C represents that spatial domain is single The working capacity value of unit；

Described ATC controller workload W (t₀-t₁) particularly as follows: controller is for alleviating the pressure afforded and completing objective task The time length that consumed of requirement

$W(t_0-t_1)=W^{t_0-t_1,com}+W^{t_0-t_1,incom}+W^{t_0-t_1,thk}$

Wherein, W (t₀-t₁) represent monitoring period of time t₀-t₁Interior total working load；Represent monitoring period of time t₀-t₁Interior control The communication work load of member；Represent monitoring period of time t₀-t₁The non-communicating workload of interior controller, as filled in electronics The loads such as process list, screen operator；Represent monitoring period of time t₀-t₁The thinking load of interior controller.

Traffic congestion state monitoring method in a kind of sector null the most according to claim 1, it is characterised in that described extraction Air traffic running status characteristic particularly as follows:

Step one, extracts spatial domain cellular construction data and running status basic data, including sector area, air route structure, respectively navigates Segment length, vital point position and coordinate, the distribution of corridor, sector mouth and coordinate, the position of any time every frame airborne vehicle in sector, Speed, course, highly, original base, land airport, airborne vehicle and estimated the data such as point, set up basic database；

Step 2, analyzes unitary space, spatial domain structure, extracts each vital point, course line section, height layer and whole sector respectively crowded Various types of data needed for monitoring, in calculating monitoring timeslice, sector " point-line-surface body " air traffic is blocked up monitoring index value, is formed Air traffic congestion status monitored data analysis.

Traffic congestion state monitoring method in a kind of sector null the most according to claim 3, it is characterised in that described employing Traffic congestion state in sector null is identified referring specifically to " the point-line-surface to whole sector by Fuzzy C-Means Cluster Algorithm (FCM) Body " the air traffic congestion status of each Spatial Dimension is identified, refers specifically to:

Step 3, uses correlation analysis method respectively " point-line-surface body " each monitoring index to be carried out correlation analysis, screens out phase The index that closing property is bigger, calculates the correlation coefficient r between each index_jk:

$r_{jk}=\frac{\underset{i=1}{\overset{I}{\Σ}}(x_{ij}-\stackrel{\‾}{x_j})(x_{ik}-\stackrel{\‾}{x_k})}{\sqrt{\underset{i=1}{\overset{I}{\Σ}}{(x_i-\stackrel{\‾}{x_j})}^2}\sqrt{\underset{i=1}{\overset{I}{\Σ}}{(x_{ik}-\stackrel{\‾}{x_k})}^2}}$

Wherein, x_ij、x_ikIt is respectively jth, the i-th timed sample sequence value of kth index,It is respectively jth, k finger Target sample average, I is number of samples；

Step 4, degree of membership based on fuzzy mathematics is theoretical, uses Fuzzy C-Means Cluster Algorithm (FCM) to sector vital point, boat The air traffic of line segment, height layer and the full sector monitoring index that blocks up clusters respectively, and show that corresponding monitoring index is joined Number；

Step 5, the air traffic in each vital point of Real-time Collection, each course line section, each height layer and full sector unit interval sheet is gathered around Stifled monitoring index, uses Euclidean distance to differentiate the acquisition index distance size away from different congestion state index parameters, when identifying this The each vital point of section, each course line section, each height layer and the air traffic congestion status of full sector.

Traffic congestion state monitoring method in a kind of sector null the most according to claim 5, it is characterised in that described step Four include:

Step 4.1: normalization achievement data

$x_{ij}^{*}=\frac{x_{ij}-x_j^{\min }}{x_j^{\max }-x_j^{\min }},$

Wherein, x_ijFor the i-th time sequential value of jth index,For the sample minimum of jth index,For jth The sample maximum of individual index；

Step 4.2: initialize subordinated-degree matrix U

Set up initial subordinated-degree matrix U, order

And

Wherein, u_niRepresenting that i-th index sample sequence is under the jurisdiction of the degree of the n-th safe class classification, N is congestion state grade Classification number, the present invention all takes N=4, i.e. vital point congestion state grade, course line section congestion state grade and sector congestion state etc. Level be divided into freedom, pass unimpeded, four classes crowded, congested, it may be assumed that

Class={ " freely ", " passing unimpeded ", " crowded ", " congested " }

Step 4.3: calculate the cluster centre of N number of classification

$V_n^{*}=\frac{\underset{i=1}{\overset{I}{\Σ}}{u}_{ni}^m{X}_i}{\underset{i=1}{\overset{I}{\Σ}}{u}_{ni}^m},n=1,2,...,N$

Cluster centre matrixWherein,Be free state respectively, pass unimpeded shape The characteristic vector of state, congestion state and congestion state；M is Weighted Index；

Step 4.4: calculating FCM cost function J:

$J(U,c_1,c_2,...c_n)=\underset{n=1}{\overset{N}{\Σ}}{J}_n=\underset{n=1}{\overset{N}{\Σ}}\underset{i=1}{\overset{I}{\Σ}}{u}_{ni}^m{d}_{ni}^2$

Wherein d_niBeing the Euclidean distance between the cluster centre of the n-th classification and i-th data, m is Weighted Index, and the value of J is anti- Reflecting compactness in class, J is the least, shows to cluster and more compacts；

Cost function value J of front and back twice is compared, if cost function value knots modification Δ J is less than threshold epsilon, then goes to step 4.6, otherwise go to step 4.5；

Step 4.5: renewal Subject Matrix U:

$u_{ni}=\frac{1}{\underset{k=1}{\overset{N}{\Σ}}{\left(\frac{d_{ni}}{d_{ki}}\right)}^{\frac{2}{m-1}}}$

Repeat step 4.3, step 4.4；

Step 4.6: output subordinated-degree matrix U, uses the principle of maximum membership degree that each data point is carried out congestion state classification, and Output cluster centre matrixCluster centre carrying out renormalization and obtains actual congestion indication parameter, renormalization formula is such as Under:

$V_{nj}=V_{nj}^{*}\×(x_j^{\max }-x_j^{\min })+x_j^{\min }$

Wherein,For the cluster centre of n congestion state jth index,For the sample minimum of jth index, Sample maximum for jth index.