EXAMINING THE FEASIBILITY OF PAIRED CLOSELY-SPACED PARALLEL APPROACHES

EXAMINING THE FEASIBILITY OF PAIRED CLOSELY-SPACED PARALLEL APPROACHES Seven J. Landry and Amy R. Priche Georgia Insiue of Technology Alana GA 30332-0205 ABSTRACT Paired closely-spaced parallel approaches may enable closely-spaced runways o operae a nearvisual capaciy in all weaher condiions. This concep requires each rail aircraf in he pair o fly in a safe zone in fron of he lead aircraf s wake vorex and in back of he lead aircraf s poenial blunder rajecories. This paper examines he posiion and lengh of his safe zone o undersand he srucure of his operaing environmen. The resuls highligh he need for an alering sysem, and idenify faces of procedures needed for his operaion. INTRODUCTION Many airpors have parallel runways, allowing for simulaneous operaions wihou concern for crossing raffic flows. For aircraf o land simulaneously on hese runways requires hem o fly close ogeher for several minues on parallel approach pahs o he runways. This demands a mechanism for ensuring safe separaion beween aircraf, and beween aircraf and oher aircraf s wake vorices. When weaher condiions allow aircraf o visually acquire each oher for all of he approach, he responsibiliy for separaion is given o he pilos. However, weaher condiions frequenly do no allow visual approaches; in hese cases, oher mechanisms are required. Currenly, air raffic conrollers supervise he approaches and ensure separaion. However, several facors including concerns abou wake vorex separaion, he resoluion of conrollers radar sysems, and he ime required for conroller inervenion limi parallel operaions o runways wih a leas 4300 fee separaion for independen operaions, 3000 fee for independen operaions wih he specialized Precision Runway Monior, or 2500 fee for dependen operaions where he conroller saggers he aircraf along he approaches o keep a large separaion beween hem. 1 Several sudies have examined airborne separaion assurance, in which pilos and fligh deck sysems mainain safe separaion, wih he hope of allowing all weaher use of closely spaced parallel approaches. One mehod of parallel approaches is reacive; i.e. i relies upon pilo monioring and/or alering sysems o deec any possible loss of separaion. 2,3,4,5,6,7 Anoher mehod is paired approaches, where wo aircraf on closely spaced parallel approaches are paired, and he rail aircraf mainains posiion wihin a safe zone relaive o he lead aircraf. 8,9 As shown schemaically in Figure 1, his mehod is prevenive in naure his posiioning guaranees neiher aircraf will be in danger of loss of separaion should he oher depar is approach pah, and ha neiher aircraf will be affeced by he oher s wake. Lead aircraf curren Wake curren Safe zone curren Range of poenial lead blunders Trail aircraf Wake fuure Trail aircraf curren fuure

Figure 1. Schemaic of he safe zone a he curren ime o guaranee collision and wake avoidance a fuure imes This operaion needs o be clearly undersood before beween he aircraf in a urn lasing for a duraion T urn implemenaion, as he aircraf are closer ogeher in can hen be found: reduced visibiliy han allowable in any oher airborne 2 2 D? x civilian operaion. Several preliminary sudies have 2 0? K1? K 2? K3? K 4 examined paired approaches, suggesing possible v procedures, 10 and assessing cockpi raffic displays. 11 1 K1? sin?? v1turn cos?? However, sudies o dae have ypically examined specific implemenaions, such as ess of a specific K 2? v1 cos?? v2 raffic display or procedure. v The novely and complexiy of his operaion meris a comprehensive examinaion of he underlying sysem dynamics; wih such knowledge, we can more conclusively deermine he pilos informaion, procedure, and echnology needs, and poenially applied more-srucured design mehodologies o reduce he ime required o design and es new echnologies and procedures for his operaion. In his case, he sysem dynamics are largely governed by he size and posiion of he safe zone, which is deermined by several facors including pilo acions o say wihin he safe zone, procedures, aircraf dynamic behavior, and alering sysem look-ahead ime. In his paper we firs ouline how he safe zone is calculaed, and hen we idenify he impac of procedural assumpions and echnologies on sysem dynamics. Finally we discuss he insighs provided ino he srucure of he environmen, and heir implicaions for his air raffic conrol operaion. CALCULATING SAFE ZONE LOCATION The locaion of he safe zone can be found analyically from examinaion of curren posiions and poenial fuure rajecories of he wo aircraf. Since verical posiion can vary quickly and unpredicably, separaion can no be guaraneed in he verical plane. Insead, he fron of he safe zone provides separaion beween he wo aircraf in he horizonal plane, which is limied by procedural resricions on allowable behavior and on aircraf dynamic behavior namely, ranspor aircraf can no fly sideways, and can no quickly urn o a new heading. In calculaing he safe zone, we use he wors case model of a poenial loss of separaion, i.e. a blunder by he lead aircraf where i changes heading. We assume he blunder heading change rae is achieved insanly, an assumpion ha was esed and found o have negligible consequences. The horizonal disance K K 3 4 1? 1? cos?? y2 0?? v sin? 1 The fron of safe zone is he curren locaion of he rail aircraf ha can guaranee no loss of separaion can occur wihin a given proecion ime; i.e. his equaion can be solved for he minimal value for x 2 guaraneeing separaion in all allowable fuure condiions wihin he proecion ime. The disance he rail aircraf can fly while he wake o crosses beween he aircraf is he maximum allowable in-rail disance. We assume ha he wake is ranspored a crosswind velociy, a reasonable approximaion of recen research. 12 The back of he safe zone is herefore deermined by he laeral separaion beween aircraf, he crosswind, and he speed of he rail aircraf. In summary, he size and locaion of he safe zone is deermined by he variables lised in Table 1. The calculaions used here do no consider uncerainy abou parameers; raher, soluions can be found for wors-case values given knowledge of heir saisical properies. x2 0 Iniial longiudinal separaion y2 0 Iniial laeral separaion beween aircraf v, v Lead and rail aircraf speeds 1 2? Blunder heading change? Blunder heading change rae pro Proecion ime cap of safe zone Crosswind V c Table 1. Fron of safe zone variables RESULTS

This sudy explores how he fron and back of he safe zone change wih he variables lised in Table 1. The locaion of he fron and back of he safe zone was calculaed for laeral separaions of 750 fee o 3000 fee, lead and rail speeds of 100 o 230 knos, heading changes of up-o 45 degrees, and up o infinie heading change raes. The following secions highligh some of he mos imporan resuls of his analysis.

Impac of Aircraf Ground Speed The speed of boh he lead and rail aircraf can have a significan impac on he posiion of he fron of he safe zone. For example, Figure 2 shows is locaion as he minimum allowable disance in rail for a range of aircraf speeds. Two addiional insighs should be noed beyond is overall sensiiviy o speed. Firs, here may be some condiions where he rail aircraf can safely be in fron of he lead aircraf specifically, when he lead aircraf is faser as shown by negaive allowed disances in rail. Second, addiional analysis idenified condiions where he fron of he safe zone needs o proec agains blunders of he rail aircraf owards he lead specifically, when he rail aircraf is faser. Impac of Blunder Severiy The safe zone is defined as prevening a blunder from causing loss of separaion. Therefore, we examined he impac of requiring i o proec agains blunders of greaer and lower severiy, as esablished by wors-case blunder heading changes and change raes. As shown in Figure 3, he mos resricive safe zone i.e. one requiring a greaer minimum in-rail disance is deermined by he lowes blunder heading change, and mus be applied in siuaions where he rail aircraf is or may be faser han he lead aircraf, a siuaion corresponding o siuaions where he lead aircraf may slowly drif over-and-back ino he rail aircraf. Large heading changes, conversely, do no have a significan impac on safe zone locaion as he lead aircraf may no have he ime and space o enac a large urn before crossing over he rail aircraf s approach pah. A similar effec was found for blunder heading change rae. Impac of Proecion Time The previous secion s resuls sugges ha he fron of he safe zone, if required o proec agains all possible inrusions, is posiioned conservaively by he lowes possible blunder heading changes and change raes. However, such blunders require a long ime o cause a separaion hazard. Therefore, we analyzed he impac of esablishing a proecion ime which he safe zone needs o provide; i.e. he safe zone will guaranee separaion agains blunders aking less han he proecion ime o evolve, whils anoher mechanism such as an alering sysem is required for blunders aking longer. As shown in Figure 4, laxer proecion ime limis move he safe zone forward subsanially from he wors-case locaions shown in Figure 3. Minimum Allowed Disance In Trail f 3000 2000 1000 0-1000 Proecion agains 30-degree blunder wih laeral spacing of 1000 f 110 130 150 170 190 Lead Speed 120 140 160 180 200 Trail Speed ks

Minimum Allowed Disance In Trail f Figure 2. Impac of rail and lead aircraf ground speed on locaion of he fron of he safe zone 3500 laeral separaion = 1000 f, rail speed = 150 ks 3000 heading change rae = 6 deg per sec 2500 Lead 2000 Speed ks 1500-20 130 1000-10 140 150 0 500 160 10 170 20 0-500 -1000 5 10 15 20 25 30 35 40 45 50 55 60 Blunder Heading Change deg Figure 3. Locaion of fron of safe zone as a funcion of maximum possible blunder heading change, wihou a proecion ime limi 3000 lead speed = 150 ks, rail speed= 170 ks Minimum Allowed Disance In Trail f 2500 2000 1500 1000 500 No blunders can cause separaion violaions wihin 20 secs a greaer han 1800' laeral separaion Proecion Time sec 20 30 40 0 750 900 10501200135015001650180019502100225024002550270028503000 Laeral Separaion f Figure 4. Locaion of he fron of he safe zone wih differen proecion imes and laeral separaions

200 Trail speed knos 180 160 140 120 Achievable for for 40 40 sec sec proecion ime, ime, 1000 1000 f f laeral laeral separaion Achievable for 30 Achievable sec proecion for 30 sec ime, proecion 750 f laeral ime, separaion 750 f laeral separaion Achievable for 40 sec proecion ime, 750 f laeral separaion 100 100 110 120 130 140 150 160 170 180 190 200 Lead speed knos Figure 5. Range of condiions in which safe zone exiss and paired approaches is feasible Is There Always a Safe Zone? The previous secions discussed he locaion of he fron of he safe zone. As noed earlier, he back of he safe zone is a simple linear funcion of laeral separaion, crosswind speed, and rail aircraf speed. The safe zone may be said o exis when he back is behind he fron. There are some cases where his is no rue; however, hese combinaions are confined o low laeral separaions, large proecion imes, low lead aircraf speeds, and high rail aircraf speeds, as shown in Figure 5. The safe zone lenghens rapidly away from hese combinaions. CONCLUSIONS This sudy provides a comprehensive examinaion of he safe zone during paired approaches. These resuls show ha he safe zone exiss for all bu he wors combinaions of high proecion imes, low runway spacings, high crosswinds, and low aircraf speeds. These resuls also highligh he imporan facors in calculaing he safe zone locaion and size aircraf speeds, proecion ime, laeral separaion, and crosswind speed. Addiionally, hese resuls highligh he complexiy of calculaing he safe zone heir require compuaions can poenially be performed in near-real-ime by a compuerized cockpi display given knowledge of he immediae siuaion, bu are likely oo complex and inricae o require of a pilo during he high-empo operaion of final approach. In addiion, oher sudies of reacion mehods of closely spaced parallel approaches sugges pilos may no be accusomed o, and no very good a, he ask of spacing and collision deecion in his phase of fligh. 6 The resuls also highligh ha he safe zone can only reasonably provide a limied proecion ime, capable of guarding agains rapid losses of separaion bu no agains very slow maneuvers of one aircraf owards anoher. This suggess ha supplemenal mechanisms of deecing slow inrusions will be beneficial in lenghening he safe zone. Reliance solely on conroller inervenion is unlikely due o he specialized ground based sysems his would require and due o he exra ime needed for a conroller o effecively inervene hrough voice

commands. Likewise, relying solely on pilo monioring of cockpi displays may reasonably be hypohesized o be problemaic given heir workload and compeing asks during approach. Therefore, hese resuls sugges he need for a cockpi alering sysem; design of such a sysem does no need o be overly complex, however, as i only o deec slow inrusions. Procedural concerns are raised by hese resuls. For example, he concep of paired approaches has ypically assumed ha he rail aircraf boh should be behind he lead aircraf, and ha he rail aircraf would be requiring proecion from he lead aircraf s acions. Insead, hese resuls idenified siuaions where he rail aircraf may safely be in fron of he lead aircraf when he rail aircraf is slower, and siuaions where he rail aircraf could poenially be he insigaor of loss of separaion when i is faser. As such, he disincion beween rail and lead aircraf may insead be viewed as a procedural maer where each aircraf in a pair are assigned a role one o fly independenly, and he oher o fly wihin he safe zone hey joinly define. Of course, while providing guidance ino procedural and echnical needs, his numerical analysis can no compleely specify he bes implemenaion. However, he knowledge gained here may also help subsequen design processes by enabling more srucured design mehods. For example, Ecological Inerface Design EID serves o creae displays clearly porraying informaion abou all levels of absracion relevan o he ask a hand, including funcional relaionships and physical properies. EID is has been widely sudied in siuaions where he requisie means-end relaionships are clearly defined by physical properies; he analysis here may serve as an example of idenifying means-end relaionships based on complex inerplay beween physical sysems in his case, aircraf dynamics, inelligen compuerized sysems in his case, alering sysems, and procedural definiions of normaive behavior. ACKNOWLEDGEMENTS This sudy is suppored by NASA and he FAA, Gran NAG 2-1314, wih Vernol Baise and Gene Wong serving as Technical Moniors. The inpu of Anand Mundra, Jonahan Hammer, Oscar Olmos and Randy Bone of MITRE/CAASD is also appreciaed. REFERENCES 1 Owen, M.R. The Memphis Precision Runway Monior Program Insrumen Landing Sysem Final Approach Sudy, Repor DOT/FAA/NR- 92/11, MIT Lincoln Labs, May 1993. 2 Priche, A.R. Carpener, B.D., Kuchar, J.K., Hansman, R.J. & Asari, K. Issues in Airborne Sysems for Closely-Spaced Parallel Runway Operaions, Proceedings of he 14 h AIAA/IEEE Digial Avionics Sysems Conference, 1995 3 Waller, M.C. and Scanlon, C.H. A Simulaion Sudy of Insrumen Meeorological Condiion Approaches o Dual Parallel Runways Spaced 3400 and 2500 Fee Apar Using Fligh-Deck- Cenered Technology, NASA TM-1999-208743, NASA Langley Research Cener, March 1999. 4 Carpener, B.D. and Kuchar, J.K. A Probabiliy- Based Alering Logic for Aircraf on Parallel Approach, NASA CR 201685, April 1997. 5 Teo, R. and Tomlin, C.J. 2000 Compuing Provably Safe Aircraf To Aircraf Spacing For Closely Spaced Parallel Approaches. Presened a IEEE/AIAA Digial Avionics Sysems Conference, Philadelphia PA, 2001. 6 A.R. Priche Display Effecs on Shaping Apparen Sraegy: A Case Sudy in Collision Deecion and Avoidance, Inernaional Journal of Aviaion Psychology, 101, 59-83, 2000. 7 A.R. Priche and B. Vandor Designing Siuaion Displays o Promoe Conformance o Auomaic Alers, o be Presened a The Annual Meeing of he Human Facors and Ergonomics Sociey, Minneapolis, MN, Ocober 2001 8 Sone, R. Paired Approach Concep, Proceedings of he NASA Workshop on Fligh Deck Cenered Parallel Runway Approaches in Insrumen Meeorological Condiions, NASA Langley Research Cener, Ocober 1996. 9 Priche, A. R. Pilo Performance a Collision Avoidance During Closely Spaced Parallel Approaches, Air Traffic Conrol Quarerly, 71, 47-75, 1999. 10 Hammer, J. Sudy of he Geomery of a Dependen Approach Procedure o Closely Spaced Parallel Runways, 18h IEEE/AIAA Digial Avionics Sysems Conference, S. Louis, MO, 1999.

11 Bone, R. Olmos, O. and Mundra, A. Paired Approach: A Closely-Spaced Parallel Runway Approach Concep, presened a he Inernaional Symposium on Aviaion Psychology, 2001. 12 Hamilon, D.W. and Procor, F. Wake Vorex Transpor in Proximiy o he Ground, Presened a IEEE/AIAA Digial Avionics Sysems Conference, Philadelphia PA, 2001.