Data analysis From Support to Illumination Paul Bosman

Data analysis – From Support to Illumination Paul Bosman SWIM/Enterprise Architecture Unit Manager Eurocontrol Complex. World – Data Science in Aviation workshop Madrid, 15 October 2013 The European Organisation for the Safety of Air Navigation

Agenda • Todays usage of ATM Data • Trying to apply Data science to ATM • Some further reflections 2

Lets start with a question How wise are we today in ATM ? 3

Usage of data in ATM … as a drunken man uses lamp-posts. . . for support rather than illumination

Data in ATM – NOTAM Example E) NDB HB U/S DUE WIP. IAP NDB(L)DME RWY 04 AND LLZ/DME NDB(L) RWY 22 ARE TEMP WITHDRAWN. ACFT WILL BE RADAR VECTORED TO ILS APCH RWY 22 OR LOC/DME RWY 04 OR VISUAL FINAL. IN THE EVENT OF A MISSED APCH ON EITHER RWY, CLIMB STRAIGHT AHEAD TO 3000 FT QNH OR AS INSTRUCTED BY RADAR. IN THE EVENT OF BELFAST RADAR BEING UNAVAILABLE, ALDERGROVE RADAR WILL VECTOR ACFT TO LOC RWY 04 SUBJECT TO WORKLOAD. AS PER ALTERNATIVE PUBLISHED APCH PROC, ACFT MAY WHEN APPROVED, SELF POSITION TO FINALS RWY 22/04 VIA DME ARC, ROUTEING TOWARDS VOR BEL FM EITHER NELBO, RINGA OR NEVRI UNTIL INTERCEPTING THE DME ARC. ACFT JOINING VIA BLACA SHOULD CONTINUE ON TRACK MAGEE THEN LEAVE MAGEE TRACK 255 AND ESTABLISH ON THE RWY 22 LOC AT DME 11. 5 I-BFH, THEN AS DIRECTED. FOR RWY 04 AFTER PASSING MAGEE ESTABLISH ON THE DME 11 ARC RWY 04 HB QDM 022 CAN BE SUBSTITUTED BY BEL DME 15. 5. RWY 22 HB QDM 245 CAN BE SUBSTITUTED BY BEL R 085. MISSED APPROACH RWY 22: - CLIMB TO 3000 FT, INITIALLY STRAIGHT AHEAD TO 3000 FT OR I-BFH D 9, WHICHEVER IS EARLIER, THEN LEFT TURN ON TRACK 005 TO INTERCEPT ON BEL R 075 OUTBOUND TO MAGEE, OR AS INSTRUCTED. MISSED APPROACH RWY 04: - CLIMB TO 3000 FT INITIALLY STRAIGHT AHEAD TO INTERCEPT AND ESTABLISH ON THE BEL R 075 OUTBOUND TO MAGEE, OR AS INSTRUCTED. THESE INSTRUCTIONS SHALL ALSO APPLY IN THE EVENT OF A RTF DURING A MISSED APCH. THEREAFTER COMPLY WITH STANDARD RTF PROCEDURE. FROM: 29 MAY 2009 11: 05 TO: 31 OCT 2009 21: 30 A 1656/09 5

Data usage in ATM Pre-flight Information Bulletin (PIB) International NOTAM trends (source EAD) 6

So again How wise are we today in ATM ? 7

How did we get into this situation ? – ly t n a t r o p im e And mor ? n io t a u it s is h t f o t u o t e g e w o d w o H 8

Can Data science get us out of this situation ? • From www. complexworld. eu • Data science is a multi-disciplinary endeavor combining a range of skills from a variety of fields. Specifically the field is at the intersection of different disciplines: complexity science, statistical physics, network theory, data mining, knowledge discovery, computer science, indexing techniques, stream data processing, data evaluation environments, scalable analytics, visualization and, very importantly, strong domain expertise. o t ls ia r t e m o s e r Following now a s t p e c n o c l a n io t a create new oper e c n ie c S a t a D n • The application baseofddatao science principles to the aviation sector can open the gate to significant improvements in many other key aspects of aviation such as safety management, flight efficiency, environmental-impact mitigation or delay reduction. 9

What if ? took over ATM ? Search Drive Translate Part of SWIM CS E. g. ICAO-IATA Codes ? 10

What if ? took over ATM ? So, lets put a monetary value on ATM services E. g. Airport / ATC slots 11

What if ? took over ATM ? What Other Customers Are Looking At Right Now … ’[EFB] … 10 Pilots like you took a wrong taxiway on this runway last week’ Customers Who Bought This ’… meteo induced costs Item Also Bought … are over 900 MEUR/year. How prepared are we ? ’ ’…applied to Pilot Briefing, Traffic/meteo, Forecast, ATC (ATIS/…) ? ’ 12

Data science helping ATM – A real life example Problem Can En-route Wake. Vortex be predicted ? ? is h t f o e r o m e v a Can we h Available Data • Flight Data • Weather Data • Wake Vortex Models Or e r o m ls p p a a t a d Validation Are these Big ? n io t u • Wake Turbulence Reports c e x e n a h t g in n n la p / M D C for Initial answer = YES 13

What is the problem that we are trying to solve ? • Many different actors o, … te e M … , x a P. … , ry a it orts, Mil p ir A , s e n li ir A , M N , C • AT ts in a tr s n o c g in d d /a s n o ti optimiza n w o ir e th g in v ri d ll A • 14

The future is bright … • From Draft SESAR 2 – Work Programme • ‘. . . resilience systems in a highly automated environment …’ • ‘…ensuring cyber security in an automated environment …’ • ‘… Data Science as a multi-disciplinary endeavour blending skills at the intersection of different scientific domains such as statistical physics, network theory or complexity theory, with different techniques from computer science, like data mining, data indexing and visualization, and scalable analytics…’ BUT • NB: Future of distributed consistent systems • Brewers theorem & SES: Consistency – Availability – Partition (CAP) 15

ATM is a system with limited ‘capacity’ Do we know where we are for the ATM system? 16

Are we reaching the ‘capacity’ limits of the ATM system? • If so, what are we supposed to do? • Exhaust all potential capacity increments with better data / information / knowledge / wisdom management? But, how much would such increments represent and at which cost? • We need to better understand the ATM as a system and its ‘limits’, the factors that influence these limits and determine different optimization solutions and their impact on the stakeholders. • => change in the status-quo, ‘political’ agreements may be required. 17

Data science v 2. 0 • Agree on Penalties • Simulate a stable situation : Traffic patterns, Meteo, Aircraft types h c u m e b l il w e c n Data scie required !! • Measure impact of parameters such as - # flights, Flight plan schedules, Cruising levels, City pairs, . . . - Aircraft type, Meteo, Fiff modes of transport, … - Start mixing them • Determine system response incl bottlenecks 18

Conclusion • The road to illumination is long & windy • Many questions, less answers • Ready for more pragmatic Data science applications • Long term solutions will require some real work ! 19

Epilogue Digital NOTAMs E) NDB HB U/S DUE WIP. IAP NDB(L)DME RWY 04 AND LLZ/DME NDB(L) RWY 22 ARE TEMP WITHDRAWN. ACFT WILL BE RADAR VECTORED TO ILS APCH RWY 22 OR LOC/DME RWY 04 OR VISUAL FINAL. IN THE EVENT OF A MISSED APCH ON EITHER RWY, CLIMB STRAIGHT AHEAD TO 3000 FT QNH OR AS INSTRUCTED BY RADAR. IN THE EVENT OF BELFAST RADAR BEING UNAVAILABLE, ALDERGROVE RADAR WILL VECTOR ACFT TO LOC RWY 04 SUBJECT TO WORKLOAD. AS PER ALTERNATIVE PUBLISHED APCH PROC, ACFT MAY WHEN APPROVED, SELF POSITION TO FINALS RWY 22/04 VIA DME ARC, ROUTEING TOWARDS VOR BEL FM EITHER NELBO, RINGA OR NEVRI UNTIL INTERCEPTING THE DME ARC. ACFT JOINING VIA BLACA SHOULD CONTINUE ON TRACK MAGEE THEN LEAVE MAGEE TRACK 255 AND ESTABLISH ON THE RWY 22 LOC AT DME 11. 5 I-BFH, THEN AS DIRECTED. FOR RWY 04 AFTER PASSING MAGEE ESTABLISH ON THE DME 11 ARC RWY 04 HB QDM 022 CAN BE SUBSTITUTED BY BEL DME 15. 5. RWY 22 HB QDM 245 CAN BE SUBSTITUTED BY BEL R 085. MISSED APPROACH RWY 22: - CLIMB TO 3000 FT, INITIALLY STRAIGHT AHEAD TO 3000 FT OR IBFH D 9, WHICHEVER IS EARLIER, THEN LEFT TURN ON TRACK 005 TO INTERCEPT ON BEL R 075 OUTBOUND TO MAGEE, OR AS INSTRUCTED. MISSED APPROACH RWY 04: - CLIMB TO 3000 FT INITIALLY STRAIGHT AHEAD TO INTERCEPT AND ESTABLISH ON THE BEL R 075 OUTBOUND TO MAGEE, OR AS INSTRUCTED. THESE INSTRUCTIONS SHALL ALSO APPLY IN THE EVENT OF A RTF DURING A MISSED APCH. THEREAFTER COMPLY WITH STANDARD RTF PROCEDURE. FROM: 29 MAY 2009 11: 05 TO: 31 OCT 2009 21: 30 A 1656/09 Digital Briefing Digital Business Rules Data Presentation Process flow Business knowledge 20

Contact Information • EUROCONTROL http: //www. eurocontrol. int/im http: //im. eurocontrol. int/wiki Click here to access the NOP B 2 B • SESAR Joint Undertaking http: //www. sesarju. eu/swim • Data Exchange models • • Aviation Information Data Exchange AIXM - Aeronautical Information Exchange Model WXXM – Weather Information Exchange Model FIXM - Flight Information Exchange Model www. aidx. aero www. aixm. aero www. wxxm. aero www. fixm. aero • E-mail paul. bosman@eurocontrol. int 21