Flight data

Question 1

How is flight data recorded?

Answer 1

• FDR flight data recorder -> many different types(old models still in use)
• CVR cockpit voice recorder

Question 2

What devices are used on board?

Answer 2

• measurement
• aquisition/processing (FDIU Flight data interface unit)
• recording (FDR, CVR, QAR quick accsess recorder)
• output (wireless, disk, pc card)

Question 3

What are some requirements for CVRs / FDRs?

Answer 3

FDR:
- must be on board to record parameters + voice related to crash
- data is continiously overwritten
- located in the rear of the aircraft
- painted yellow/orange
- at least for the last 25h
ICAO: flight path, speed altitude, engine power..
CVR:
- records pilot microphones+earphones
- records all ATC+intercom conversations
- Records background using area mic
- at leasr the last 30 min, since 2016 2h

Question 4

How is data structured in ARINC717?

Answer 4

• records in 64/128/256/1024/2048 words per second
• each word is 12bits (stored in 16bit blocks)
• each second is calles a subframe-> 4subframes= 1frame->16 frames=1 superframe
• 4 different subframes, identified by unique sync words

Question 5

How are different sampling rates realized? ARINC717

Answer 5

depends in Bit number: 64/128..

??

Question 6

How to read a parameter in ARINC717 given word & subframe position?

Answer 6

• sync words identifies subframe 1, 2, 3, 4

- unique decoding table which shows the position of the parameters in the subframes

Question 7

What other sources of information are there?

Answer 7

• ADS-B (automatic dependend surveillance broadcast)
• NOTAM, METAR, SIGMET
• terrain data
• airspace configuration
• radar tracks
• flight plan

Question 8

Why is QAR data Big Data?

Answer 8

• file size: 2-200MB, 39 Mio flight 2019
• many different informationfrom all stakeholders
  Big Data
• volume (Data at Rest): terabytes-exabytes of data
• velocity (Data in Motion): streaming data, milliseconds to seconds to respond
• variety (Data in many forms): un/structured, text, multimedia
• veracity (Data in Doubt): uncertainties due to inconsistencies, incomplet, latency, deception, model approximation

Question 9

How to work with such data on a large scale?

Answer 9

• setup depends on the volume/costs/…

- online requires de-identification