This presentation will give an overview of the capabilities available in FME that allow NAV CANADA to automate the instrument procedure chart process from source data to publication.
FME streamlines the complex process of translating and transforming AIXM 4.5 data following strict guidelines to create charts that conform to ICAO standards.
WHY:
Higher Quality - To enforce Standards and Specifications built into the process.
Productivity Boost – Increased speed of chart creation enables greater throughput.
Automation – Minimize human error.
9. ● Current inventory:4500 Procedures
○ 300/year new or overhauled
■ Approximately 70 attributes/procedure
○ 3000 Life cycle changes/year
■ Subset of attributes
● 56 day publication cycle
The Challenges
Procedure Data
Transcribed into
Documents
Interpreted into Chart 3
2
1
11. Aeronautical Chart
The FME Solution
CAD Procedure Design Load to the Oracle
Master Database
Step One Step Two
12. Old Workflow – Step One
Procedure Forms
Procedure Design
Microsoft Excel
Master Database
Manual Data
Entry in Master
Database
Manual Intervention
Excel content
manually entered
in master
database
Table
for Data
Entry
Manual Entry in
Excel and
Word/PDF
Manual Intervention
and interpretation
Design elements
copied to Excel
Design
13. Improved Workflow – FME – Step One
Table
For
Validation
PDF
AIXM
Validated
Table AIXM 4.5
Master DatabaseMicrosoft Excel
Only specific data
elements are
exported to Excel
FME Workbench
ETL process from
Excel to AIXM 4.5
FME Workbench
ETL process from
PDF and AIXM 4.5
to Excel
Procedure Form
Procedure Design
14. CAD Procedure Design Load to the Oracle
Master Database
Step One Step Two
Aeronautical Chart
The FME Solution
15. Master Database
Data User
Report Export
Microsoft Excel Manual Intervention GeoMedia
Old Workflow – Step Two
Creating SHP
Data Export
Data Analysis
Results Data
Export
MicroStation
User Manual
Review &
Adjustments
PDF Final
Results
Major processing
is done here
Only specific data
elements are
exported to Excel
Major drafting of
aeronautical charts
Formating
Data
Export
17. Master Database
AIXM 4.5
Snapshot Export
FME Workbench
Apply standardization
MicroStation
Make minor edits &
adjusments
PDF Final Results
Export completed
charts as PDF
Data Analysis &
Drafting Export
User Manual Review
& Adjustments
Improved Workflow – FME – Step Two
Imagine you are flying an airplane, and you want to land somewhere in there;
Where’s the airport?
Are there mountains we can’t see?
How do I know I won’t hit anything
Pilots call this “the soup”
Once you’re in there, you can’t see outside, and you trust your instruments.
Still, how do they know where to go?
Rights for picture: CC0 Creative Commons
Free for commercial use No attribution required
“The soup”
Give you an idea of the challenge
What does Nav do, bigger picture, context, flight safety
Volume of data changes, designs, chart changes, LoE, time, 56 day cycle
Permission: This file is licensed under the Creative Commons Attribution-Share Alike 4.0 International license
Hi, my name is Alex and I work for Nav Canada.
What Do we Do:
“We manage Canada’s airspace
Major Business ATC
But also:
Infrastructure
RADARNavigational Aids
Outer Space!
Aeronautical Information
That’s us
Pictures are all part of Nav Canada’s Branding Toolkit
Tracks are planned based on
Terrain
Obstacles
ATC coordination
Picture Source: Nav Canada licenced software screenshot (FPDAM- all shown information is public)
Then we protect you and decide
“How low can you go?”
Sophisticated software
Picture Source: Nav Canada licenced software screenshot (FPDAM- all shown information is public)
Lots of manual work
Transcription error risk
Requires A TON of verification
Thanks Alex for the introduction, my name is Philippe, I am a Data Analyst in the aeronautical information management group at NAV CANADA and work alongside Alex and Rob
Alex, that sophisticated output is called XML, its looks complex, but to the trained eye, its actually structure aeronautical data, but in this case, its not in the structure we want, so why not use FME to fix that
Here is the Big Picture of what we are trying to achieve with FME
The end goal is the chart, but to get there, we first need to populate our master database
Our master database sits in the middle of our business, its an oracle database with over 300,000 aeronautical features, from runways and airspaces to procedures and obstacles
Data needs to follow schema and business rules, must pass several checks before its committed
Once committed, its feeds applications and various systems across NAV CANADA as well as Flight management system in airplanes; and its also clients synchronized with euro database, so data most perfectly mesh with cross border information
My objective was to automate the input of procedure data to that database using FME, the next few slides will focus on that first step
Into master database first
Lines up with business objectives
Before we dive in on how we used FME, lets have a look at the old workflow, it looks simple, but the goal was to replace those manual process with FME
We started with a procedure design, the designer would transcribe certain data elements from the design software to a form (PDF) and excel table, introduce risk of typos,
The PDF and excel document is transferred to another group where that data is manually inserted in the master database, again introducing risk of typos, adding time to validate
Currently, to maintain quality, we need a large amount of verification which sometimes is a long process which can involved several people and email exchanges
The objective, minimize the risk of errors by eliminating the manual steps
With FME, its looks very similar, we were able to eliminate the two manual steps, therefore greatly reducing the risk of errors and decreasing the time its takes for the data to reach the master database
This is how we did it, we started with exporting the procedure design to xml:
The xml data follows the AIXM 4.5 schema, which is an exchange model for aeronautical data, designed to enable the management and distribution of aeronautical information data in digital format
A single Procedure Export in AIXM4.5 containing approx 7500 features with over 45000 attributes, that a lot of data for a single procedure
The aixm file is supplemented with a PDF form, not ideal format to work with, but by converting to text and using regular expressions, I was able to extract what I needed to complete my dataset
We only need approx 20 features, representing around 70 attributes changes, since the procedure we exported doesn’t follow our database structure, we needed to select what we needed, recreate some associations, clean the data, generate data for the empty fields, apply some schema and business rules,
Sounds simple, but the broke the 10 transformer rule doing so
The second workspace: transforms the excel table into an AIXM 4.5 ready for upload to our master database but before it loaded, we compare the aixm file with a data snapshot in order to only introduce and/or modify the required information
We started with a process that took 3-4 hours to complete, and with FME, we were able to the load the information to our master database in roughly 20-30, workspace processing time is about 2-3mins total, must of the 20 minutes is the validation process and the export and upload process
Once the data in the Master Database, it ready to be charted, my co-worker Rob will focus on that process
The Importancy of Phil’s Process
In order to create this pretty picture (aeronautical chart)
Information needs to be stored (where “Master Database”
Because information stored here is what drives to create this pretty picture.
So before we go further into the FME Solution, lets look at the OLD WORKFLOW
Built 10 Years ago! The process works (upgrade, spruce it up a bit)
Requires multiple applications and lengthy manual operation.
A process can take between 4 to 6 hours depending of the difficulty.
I see a DUNGEON. With different Stages…Just like a Video Game
So the purpose of this video game is to create a map to help me out of the dungeon.
Which in this case to create an Aeronautical Chart from information found on the Master Database.
We begin
Stage 1: Gather information from the Wizard (Information are extracted from the authoritative source into this excel sheet)
Stage 2: Then using our wand (microsoft Excel) we decipher thru the information
Stage 3: With the information we gather, we go out on an adventure to collect the material required by beating monsters and going thru traps to create (Manual intervention)
Stage 4: Dropping all Materials into this magic cauldron (Geomedia) , stir with to synthesize the PAPER which is required to create the map..
Stage 5: Now using all of the power from the magic wand (microstation) for the final time!... Abracadabra… Poof the MAP has been created (Aeronautical Chart)
Stage 6: The map has now been created and you can now find your way out of the dungeon.
So essentially,
We have a pile of documents (such as PDF, Word Doc, Excel)
That are manually processed by the user
To create this Aeronautical Charts
Anyone of you here remember the game Super Mario Brothers, where we get the flute to warp to different stages.
Well….this improved workflow, that’s what happened… we warped using jumping stages because of the new tool that has been offered the Flute (FME)
I look at this like a video game dungeon.
You notice now there is only 4 stages to beat to get out of the dungeon, a lot faster then the previous dungeon.
Stages (video game) Finding and Creating the map out of the dungeon (Whole process)
Stage 1: Gather information from the Wizard (Information are extracted from the authoritative source)
Stage 2: Simply just throw the information into the Magic Cauldron and poof
The paper needed to create the map has been synthesized.
Stage 3: Now using all of the power from the magic wand for the final time!... Abracadabra… Poof the MAP has been created (Aeronautical Chart)
Stage 4: The map has now been created and you can now find your way out of the dungeon.
This process takes between 1 to 2 hours depending of the difficulty.
The new improved workflow consist of:
Translating AIXM 4.5 XML snapshot and convert to Microstation DGN for Final Output.
The idea is to create a process to instance create an Aeronautical Chart
By essentially eliminating all the other application by having FME doing all the heavy lifting.
Reducing user’s intervention
Implementation of Speed
Productivity boost.
Based on the illustration shown here
An AIXM4.5 snapshot is captured here contain attributes such as:
Airspace, Runways, Obstacles, Procedure Legs, etc.
Keep in mind that attribute coming from AIXM4.5 contains mostly of point vector datas.
It is ingested by FME where all the magic happens.
Like I mentioned earlier, FME does the heavy lifting.
It will take the attribute information and transforms it into:
point (waypoints, obstacles, Other aerodromes),
line (procedure tracks, missed approach tracks) and
Polygons (special use airspace)
It will also implement our drafting specifications and business rules. Again reducing user’s intervention.
Once the FME major processing is complete normally taking about 5 to 10 mins to draft a whole chart.
It export the results into a Microstation DGN file format.
Another thing that is great with FME is, it export the results with the appropriate line size, line weight, levels, cells.
Now with the exported DGN file, the user will performed their manual review and adjustments.
Then Print to PDF…
Based on the illustration shown here
An AIXM4.5 snapshot is captured here contain attributes such as:
Airspace, Runways, Obstacles, Procedure Legs, etc.
Keep in mind that attribute coming from AIXM4.5 contains mostly of point vector datas.
It is ingested by FME where all the magic happens.
Like I mentioned earlier, FME does the heavy lifting.
It will take the attribute information and transforms it into…….
It will also implement our drafting specifications and business rules. Again reducing user’s intervention.
Once the FME major processing is complete normally taking about 5 to 10 mins to draft a whole chart.
Now with the exported DGN file, the user will performed their manual review and adjustments.
Then Print to PDF…
Picture:
This work is in the public domain in the United States because it was published in the United States between 1923 and 1977 without a copyright notice. See Commons:Hirtle chart for further explanation. Note that it may still be copyrighted in jurisdictions that do not apply the rule of the shorter term for US works (depending on the date of the author's death), such as Canada (50 p.m.a.), Mainland China (50 p.m.a., not Hong Kong or Macao), Germany (70 p.m.a.), Mexico (100 p.m.a.), Switzerland (70 p.m.a.), and other countries with individual treaties.
https://commons.wikimedia.org/wiki/File:Einstein-formal_portrait-35.jpg
Essentially the new improved workflow consist of:
Rob’s Lesson Learned:
With all of these automation processes, we will never get away with user intervention. There will always be a need of us to do some of the thinking.
To do some of the manual adjustment, that the application simply can’t do.
The business rules applied are more consistent, faster and efficient.
Automation reduces the risk of human errors. (it can be spelling mistake, wrong symbologies, line types) since the data base drives the final output.
Errors only occurs if the information inputed into the database is incorrect.
Phil’s Lesson Learned:
DOCUMENTATION is Important!!
Past workspaces were developed and run only once
Run on a regular basis by users not familiar with FME
Entire process required documentation, not only for user not familiar with FME but for other FME technical users in order to maintain and troubleshoot any issues
User Friendly process and interface
Structure workspace to facilitate understanding, we broke the 10 transformer rule, therefore workspace structure was crucial
All the workspaces are run on local copies of FME, a non technical, user friendly prompt was required
Alex, would you like to add anything from your end??
Alex’s Lesson Learned: Culture vs automation, stats/metrics to add
PROS
Minimize manual intervention
Automation reduce human error
Higher Quality
Fast and Efficient
User Friendly Interface
CONS.
Complexity of the workbench
Workbench Maintenance
Volume of change
Temporality
Culture vs automation