The Combination of Flight Count and Control Time as a New Metric of Air Traffic Control Activity

Transcription

1 DOT/FAA/AM-98/15 Office f Aviatin Medicine Washingtn, D.C The Cmbinatin f Flight Cunt and Cntrl Time as a New Metric f Air Traffic Cntrl Activity Sctt H. Mills Civil Aermedical Institute Federal Aviatin Administratin Oklahma City, OK May 1998 Final Reprt This dcument is available t the public thrugh the Natinal Technical Infrmatin Service, Springfield, Virginia ^OQUAirryiNSPcnBDi U.S. Department f Transprtatin Federal Aviatin Administratin

2 NOTIC This dcument is disseminated under the spnsrship f the U.S. Department f Transprtatin in the interest f infrmatin exchange. The United States Gvernment assumes n liability fr the cntents r use theref.

3 1. Reprt N. DOT/FAA/AM-98/15 Technical Reprt Dcumentatin Page 2. Gvernment Accessin N. 3. Recipient's Catalg N. 4. Title and Subtitle The Cmbinatin f Flight Cunt and Cntrl Time as a New Metric f Air Traffic Cntrl Activity 5. Reprt Date May Perfrming Organizatin Cde 7. Authr(s) Mills, S.H. 8. Perfrming Organizatin Reprt N. 9. Perfrming Organizatin Name and Address FAA Civil Aermedical Institute P.O. Bx Oklahma City, OK Wrk Unit N. (TRAIS) 11. Cntract r Grant N. 12. Spnsring Agency name and Address Office f Aviatin Medicine Federal Aviatin Administratin 800 Independence Ave., S.W. Washingtn, DC Supplemental Ntes 13. Type f Reprt and Perid Cvered 14. Spnsring Agency Cde 16. Abstract The explratin f measures f airspace activity is useful in a number f significant ways, including the establishment f baseline air traffic cntrl (ATC) measures and the develpment f tls and prcedures fr airspace management. This reprt intrduces a new metric f ATC activity that cmbines tw existing measures (flight cunt and the time aircraft are under cntrl). The Aircraft Activity Index (AAI) is sensitive t changes in bth flight cunt and flight length, and therefre is a superir measure fr cmparing aircraft activity between tw epchs f time. The AAI was applied t data frm 10 days f System Analysis Recrdings btained frm the Seattle Air Rute Cntrl Center. The advantages f the AAI were mst apparent when different aircraft types cnsistently had different mean flight lengths. Pssible uses f the AAI and ther ATC measures fr the evaluatin f new systems and prcedures are discussed. 17. Key Wrds Air Traffic Cntrl, Cmplexity, Wrklad, Aircraft Activity Index, Flight Cunt 19. Security Classif. (f this reprt) Unclassified Frm DOT F (8-72) 20. Security Classif. (f this page) Unclassified Reprductin f cmpleted page authrized 18. Distributin Statement Dcument is available t the public thrugh the Natinal Technical Infrmatin Service, Springfield, Virginia N. f Pages Price

4 TH MBINATION OF FLIGHT UNT AND NTROL TIM AS A NW MTRIC OF AIR TRAFFIC NTROL ACTIVITY INTRODUCTION An imprtant aspect f the evaluatin f new air traffic cntrl (ATC) systems is the cmparisn f current systems and prcedures with prpsed nes. A relevant factr in making these cmparisns is the level f varius activities ccurring in a specified unit f airspace during a defined perid f time. This prcess has been cnceptualized in different ways, with the use f such terms as cntrller wrklad, airspace cmplexity, and dynamic density. In general, hwever, the studies cnducted in this area have included measures f cntrller actins, aircraft dynamics, and sectr characteristics. The explratin f measures f airspace activity is useful in a number f significant ways. These include the establishment f baseline ATC measures, as well as the develpment f tls and prcedures fr airspace management (See Galushka, Frederick, Mgfrd, & Kris, 1995, and Pawlak, Bwles, Gel & Brintn, 1997, as examples). The establishment f baseline measures is necessary fr evaluating any changes resulting frm the intrductin f new ATC systems r prcedures. Significant mdificatins have been prpsed fr the United States' Natinal Airspace System (NAS; FAA, 1997). The NAS Architecture is the Federal Aviatin Administratin's (FAA's) cmprehensive plan fr mdernizing the infrastructure f the NAS. Versin 2.5 f this plan calls fr significant changes in the interactins between air traffic cntrllers and pilts, incrprating the use f updated navigatin, surveillance, and cmmunicatin equipment and crrespnding prcedures. The impact f these mdificatins n the NAS will need t be assessed by researchers. Tward this gal, the establishment f baseline measures will allw airspace activity, measured under the current system, t be cmpared with activity assessed after new systems and prcedures have been intrduced. These cmparisns will quantify any changes resulting frm the implementatin f new systems. As a simplified example, if a new set f prcedures designed t reduce handff activity were intrduced t an ATC facility, the effectiveness f the prcedures culd be assessed by cmparing measures f handff activity cllected befre and thse cllected after the implementatin f the changes. The amunt f reductin, if any, f handff activity wuld indicate the effectiveness f the new set f prcedures. Anther imprtant use f airspace activity measures is fr the develpment f tls fr mre effective airspace management. Currently the FAA perfrms an annual review f all en rute sectrs by using a cmplexity wrklad frmula (FAA, 1984). This frmula is designed t be applied t data frm peak traffic time perids and t be used t establish r adjust sectr cnfiguratin. Hwever, the need fr dynamic activity assessment, i.e., real-time measurement, is emerging with the aviatin cmmunity's mve tward the peratinal cncepts cllectively knwn as "Free Flight." Free Flight refers t the prpsed revisin f the NAS frm a centralized system, in which air traffic cntrllers assign aircraft t rutes, t a distributed system that allws pilts mre freedm t select their wn rutes and altitudes. Under Free Flight, pilts presumably will assume mre respnsibility fr maintaining separatin frm ther aircraft than they have under the current system. The changes included in Free Flight are cnsidered t be necessary because f the limitatins f the current system, the significant increases prjected fr air traffic, and desired imprvements in efficiency fr airlines. RTCA, Inc., an rganizatin that addresses requirements and technical cncepts fr aviatin and functins as a Federal Advisry Cmmittee, has perfrmed much f the planning fr Free Flight. The Final Reprt f RTCA Task Frce 3: Free Flight Implementatin (RTCA, 1995), emphasizes the need fr the FAA t develp a methd f assessing dynamic density, r "the prjected wrklad and safety f future traffic in an peratinal envirnment." Under Free Flight, this assessment prcess culd drive the dynamic recnfiguratin f sectrs, thus increasing the capacity and peratinal

5 efficiency f the NAS. The RTCA reprt cntends that this type f capacity management is essential t the Free Flight cncept. Several studies (cited belw) have explred airspace activity in varius ways. The methdlgy mst ften used assumes that many variables affecting activity in an airspace als influence the perceived wrklad and the bjective task perfrmance f the cntrller. Fr example, as the flight cunt in an airspace increases frm an average number t a high number, ne might expect the cntrller t perceive a higher wrklad level and t perfrm a higher number f bjectively measurable activities t maintain effective cntrl f the airspace. The variables used in such studies have been described with the terms wrklad, cmplexity, and dynamic density. Therefre, measures f subjective wrklad and f individual perfrmance may be related t airspace activity, including wrklad, cmplexity, and dynamic density. Buckley, DeBaryshe, Hitchner, and Khn (1983) cnducted a study cnsisting f a series f experiments in an ATC simulatin envirnment and, as a result, identified a set f fur general factrs (Cnflict, Occupancy, Cmmunicatins, and Delay) and tw auxiliary measures (Number f Aircraft Handled and Fuel Cnsumptin) that appeared t adequately represent all ther ATC measures. The authrs recmmended the use f these measures fr subsequent air traffic simulatin studies. Stein (1985) expsed cntrllers t different levels f airspace activity in anther simulatin experiment and cncluded that three variables, Aircraft Cunt, Clustering, and Restricted Airspace, significantly influenced wrklad. Mre recently, Mgfrd, Murphy, & Guttman (1994) used verbal reprts frm air traffic cntrl specialists and multidimensinal scaling t identify a list f 16 factrs that cntribute t airspace cmplexity. Pawlak, Brintn, Cruch, and Lancaster (1996) fcused n cntrllers' strategies and decisin-making activities; prpsing a list f 15 factrs that may influence perceived air traffic cmplexity. Althugh the studies cited abve used different cnceptualizatins and methds, mst included methds f cunting flights and/r the time aircraft were under cntrl. While these measures are useful fr assessing cntrller activity, it may be mre infrmative t emply a measure that cmbines infrmatin abut bth the number f flights and the duratin f cntrl. In additin, it may be mre useful t cmpute the measure separately fr different types f aircraft in rder t btain a measure f airspace activity that is sensitive t differences in specific circumstances. Althugh the measure prpsed here is nly ne f many that influence airspace activity, refinements such as this can add t ur verall ability t quantify peratinal aspects f the NAS. An ATC Aircraft Activity Index This paper intrduces a new metric f ATC activity that cmbines tw existing measures (flight cunt and the time aircraft are under cntrl) t prduce a mre infrmative metric than either measure prvides when used alne. Given an epch f airspace activity (e.g., all air traffic cntrlled by a certain ATC psitin during a specific perid f time), tw simple measures, (1) flight cunt, and (2) aggregate cntrl time, can be cmbined t prduce a measure f activity, the Aircraft Activity Index (AAI). By cmbining the tw measures, the AAI accunts fr sme f the limitatins f using each measure independently. Specifically, the prduct f the rati f number f flights t ttal epch time and the rati f aggregate cntrl time t ttal epch time is a measure that is sensitive t bth number f flights and length f flights within the epch (See quatin 1). Ratis are used in the equatin s that the measure is sensitive t the length f the epch being evaluated. quatin 1. Aircraft Activity Index Flight Cunt Aircraft Activity Index = pch Time Cntrl Time pch Time Table 1 summarizes the relative advantages and disadvantages f the different measurement techniques. Flight cunt is sensitive t the number f flights cntrlled during a certain time perid; hwever, it is nt sensitive t the length f flights. Therefre, a simple flight cunt cannt distinguish between perids when mst flights are relatively shrt and perids when they are relatively lng. Althugh the aggregate cntrl time measure is sensitive t the length f flights, it is nt sensitive t the number f flights. Cnsequently, it can nly distinguish between perids f shrt flight lengths and lng flight lengths when the flight cunts f bth perids are similar. It cannt distinguish between a relatively large number f shrt flights and a relatively small number f lng flights.

6 Table 1. Cmparisn f ATC Measures ATC Measure Advantages Disadvantages Flight Cunt / Ttal Time (F/T) Sensitive t number f flights. Nt sensitive t length f flights. Shrt flights (less cntrl activity) are nt distinguished frm lng flights (mre cntrl activity). Aggregate Cntrl Time / Ttal Time (C/T) Sensitive t length f flights. Nt sensitive t number f flights. Cannt distinguish between several shrt flights (mre cntrl activity) and fewer lng flights (less cntrl activity). Aircraft Activity Index (F/T x C/T) Sensitive t bth number f flights and length f flights. The AAI cmbines infrmatin frm flight cunt, flight length, and length f the time perid being analyzed, and therefre reflects changes in all three metrics. Figure 1 graphically demnstrates the different sensitivities f the measures when applied theretically t different flight cunt-flight length cmbinatins. Flight cunt cannt distinguish between tw shrt flights and tw lng flights, even thugh the aggregate cntrl time fr the shrt flights is nly half as lng as that fr the lng flights. Flight cunt can, hwever, distinguish between tw shrt flights and ne lng flight, even when their aggregate cntrl times are equal. Aggregate cntrl time can distinguish between tw shrt flights and tw lng flights but is nt sensitive t the difference between tw shrt flights and ne lng flight. Unlike the simpler measures used alne, the AAI can distinguish between tw shrt flights and tw lng flights and between tw shrt flights and ne lng flight. MTHOD Applicatin f the AAI Airspace activity shuld vary as a functin f the type f aircraft cntrlled because different aircraft types have different characteristics that may influence their interactin with ATC. Aircraft perfrmance characteristics are an imprtant factr that must be cnsidered by the cntrller when assessing ptential cnflicts r sequencing aircraft. One methd f classifying aircraft accrding t perfrmance characteristics is t distinguish between three types f engines: jet, turbprp, and pistn prpeller. Different types f aircraft als may require varius levels f interactin with ATC, depending n their flight purpse and the types f navigatin and cmmunicatin equipment installed. These flights can be distinguished by classifying them int three categries: cmmercial, general aviatin (GA), and military. Fr purpses f demnstrating the AAI, flights included in this study

7 Figure 1. Sensitivity f ATC Measures Flight Cntrl Aircraft Cunt/ Time/ Activity Time (F/T) Time (CAT) Index 1 Time Units: Flight I Flight Cannt Distinguish Can Distinguish 0.25 I Can Distinguish Flight Flight 0.5 _l_ Can Distinguish T Flight Cannt Distinguish 1.0 Can Distinguish T 0.25 Aircraft Activity Index = (Flight Cunt / pch Time) X (Aggregate Cntrl Time / pch Time) were analyzed by aircraft type, with type being distinguished as ne f the nine pssible cmbinatins f the abve classificatins. This resulted in the fllwing nine types: Cmmercial-Jet, Cmmercial-Turbprp, Cmmercial-Pistn, GA-Jet, GA-Turbprp, GA-Pistn, Military-Jet, Military-Turbprp, and Military-Pistn. The AAI was applied t data frm 10 days f System Analysis Recrdings (SAR) btained frm the Seattle Air Rute Traffic Cntrl Center (ZS). Selected data were extracted frm the SAR recrdings by thenas Data Analysis and Reductin Tl (DART). A cmputer prgram, the NAS Data Management System, was develped t encde the text-based DART utput reprts int relatinal databases. An additinal cmputer prgram analyzed the databases and identified and categrized aircraft types assciated with each cntrlled flight. Flight cunt was calculated fr each defined epch by cunting all flights cntrlled by Seattle Center fr at least 6 secnds (i.e., ne update f the Track file prduced by DART) during that epch. Aggregate cntrl time was btained by adding the elapsed time f all such updates fr every aircraft cntrlled by the Center during the epch. RSULTS The Aircraft Activity Index (AAI) was calculated fr each aircraft type fr each f the 11 mst active hurs f the day ( hurs, lcal time). Fr the data frm all 10 days, mean flight lengths are shwn in Figure 2. Mean flight cunts (per day) are shwn in Figure 3. Mre detailed data frm a representative day are displayed fr illustratin purpses in Figures 4-6. Flight cunt, aggregate cntrl time, and the AAI are displayed fr Friday, Octber 22, The AAI fr all aircraft types cmbined is displayed in Figure 7. The advantages f the AAI becme apparent when different aircraft types cnsistently have different mean flight lengths, as was the case with the Seattle data (see Figure 2). An example is the cmparisn between tw aircraft types, Cmmercial-Turbprps and GA-Pistn, which shwed similar hur-lng flight cunts between 1000 and 1200 hurs thrughut the ten days f data (see Figure 4). If flight cunt were cnsidered alne as a measure f activity, the tw aircraft types culd be cnsidered t require similar amunts f a cntrller's attentin during that time perid. Hwever, the types had different mean flight

8 L n imsma^f'''"" ^WKnWH SsSÄliMiß-'S^ >?4?&>tZ:&:it v.*< * " *. ~3 >. re a Iff.! Ji^lF»»)» "'.""»«Wfeii* W«J«H»«««'!!!!' )! '.».KM,; im.«i jmew-1- " i.!.»8!!.!!!ü!! >J<y- «"!.w. lr ^v./^v>'^-^ :: ' : /.-"' i-..'-' ::/;' <' '' ' >;. :. >:.. ;. W I < c TO c re v 0) l_ 3 D) 1 fcav.4v?'v:.< mmm. l! a-;^fp^^^w^^:,^!w " ^:: - l g^ rrm Wi'W'i < < CL I Ü?* l "2^"* : ^.'-w^.q t- Ü } i Ü O in sajnu!i/\

9 >» ra T3 O c 3 O Ü +rf JZ at c re a> 0) i_ 3 Ö) lun JMBIU

10 9 ~ m \- a. w i i i H Q- < < < n= ü ü Ü (DUO 222 HI P 9 * «I I Jet Turb Pist m 0) C 3 O Ü 3 D) il ^ en 0) Ü»un iqbiid

11 Cm-Jet Cm-Turb Cm-Pist GA-Jet GA-Turb Hl GA-Pist Mil-Jet Mil-Turb Mil-Pist 1 <? «* * i : : : > CVI O c ü d) w O) 0) l_ D) < LÖ (1) i_ 3 D) 'll ra u spuas

12 Cm-Jet Cm-Turb Cm-Pist m 1 GA-Jet GA-Turb GA-Pist Mil-Jet Mil-Turb Mil-Pist 1 * * «i : : : en cv x a> TO c <-» >» > U < (0 u IM xapu AiiAjiv uejjjv

13 O in > 0) Q. n x c > u < 9) t u V u 3 O) 8 in in in y T xapu AjjAjjDV tjejjjv 10

14 lengths with GA-Pistn flights (A/=44.83, S=0.59) being abut 7 minutes lnger than Cmmercial-Turbprp flights (Af=37.58, =0.25), and pssibly requiring mre f the cntrller's attentin. A mre dramatic example is the cmparisn between Cmmercial-Pistn flights and Military-Turbprp flights. These grups had similar hur-lng flight cunts thrughut the data, but the average duratin f the military flights (M=52.28, S =2.46) was apprximately 20 minutes lnger than the cmmercial flights (A/=31.52,Sf=0.77). DISCUSSION Sme f the disadvantages f using flight cunt and cntrl time by themselves as activity measures are apparent in the Seattle data. Because Cmmercial- Pistn flights were usually the shrtest, the amunt f attentin required t cntrl them may be verrepresented in the flight cunt measure, as this measure is insensitive t flight lengths. Cnversely, the cntrller attentin assciated with Cmmercial-Pistn flights may be under-represented in the aggregate cntrl time measure mre flights were cntrlled per unit f cntrl time. Similarly, the cntrller attentin required by Cmmercial-Jets, which usually had lnger flight lengths, may be under-represented in the flight cunt measure and ver-represented in the aggregate cntrl time measure. The AAI is a mre infrmative measure than either f the simpler measures alne because it is sensitive t bth the number and length f flights. The value f calculating separate index values fr different aircraft types can be illustrated by cmparing Figures 6 and 7. Fr example, the activity assciated with hurs 1200 and 1500 appears t be the same when aircraft types are cmbined (Figure 7). Hwever, when aircraft types are cnsidered separately as in Figure 6, the activity appears t be nticeably different, with mre Cmmercial-Jet activity at 1200 than at In the future, this index f activity can be used in cnjunctin with ther ATC measures t establish baseline data fr specific sectrs r grups f sectrs. This use will facilitate evaluatin f new systems and prcedures. As new technlgies are applied t bth ATC systems and t aircraft, resulting changes in airspace activity, including cntrller tasklad, must be evaluated. Objective measures such as the AAI will cntribute t such evaluatins. Other measures that may reflect cntrller tasklad are als being investigated, including aircraft activity (numbers and amunts f changes in heading, speed, and altitude), cntrller activities (numbers and types f keybard entries), and sectr characteristics. The refinement and develpment f these measures will enhance capabilities fr evaluating ATC systems and effectively managing airspace. RFRNCS Buckley,. P., DeBaryshe, B. D., Hitchner, N., & Khn, P. (1983). Methds and measurements in real-time air traffic cntrl system simulatin. (Reprt N. DOT/FAA/CT-83/26). Atlantic City, NJ: Federal Aviatin Administratin Technical Center. Federal Aviatin Administratin. (1984). stablishment and validatin f en rute sectrs. (FAA Order N ). Washingtn, DC: Authr. Federal Aviatin Administratin. (1997). Natinal Airspace System (NAS) Architecture, Versin 2.5. Office f System Architecture and Prgram valuatin (ASD-1). Washingtn, DC: Authr. Galushka, J., Frederick, J., Mgfrd, R., & Kris, P. (1995). Plan view display baseline research reprt. (Reprt N. DOT/FAA/CT-TN95/45). Atlantic City, NJ: Federal Administratin Technical Center. Mgfrd, R. H., Murphy,. D., & Guttman, J. A. (1994). Using knwledge explratin tls t study airspace cmplexity in air traffic cntrl. The Internatinal Jurnal f Aviatin Psychlgy, 4(1), Pawlak, W. S., Bwles, A., Gel, V., & Brintn, C. R. (1997). Initial evaluatin f the Dynamic Resectrizatin and Rute Crdinatin (DI- RCT) System. NASA Final Cntract Reprt N. NAS Pawlak, W. S., Brintn, C. R., Cruch, K., & Lancaster, M. (1996). A framewrk fr the evaluatin f air traffic cntrl cmplexity. Prceedings f the AIAA Guidance Navigatin and Cntrl Cnference, San Dieg, CA. 11

15 Rbertsn, A., Grssberg, M., & Richards, J. (1979). Stein, arl S. (1985). Air traffic cntrller wrklad: An Validatin f air Traffic cntrller wrklad mdels. examinatin f wrklad prbe. (Reprt N. DOT/ (Reprt N. FAA-RD-79-83). Cambridge, MA: FAA/CT-TN84/24). Atlantic City, NJ: Federal U.S. Department f Transprtatin Research and Aviatin Administratin Technical Center. Special Prgrams Administratin - Vlpe Natinal Transprtatin System Center (VNTSC). RTCA. (1995). Final Reprt f RTCA Task Frce 3: Free Flight Implementatin, Washingtn, DC; RTCA Incrprated. 19 ÜTU.S. GOVRNMNT PRINTING OFFIC: IMS - M048S/MM3