Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 22

FUZZY MODEL FOR

THE EVALUATION OF FUTURE


AIR TRAFFIC MANAGEMENT
SCENARIOS SAFETY
PERFORMANCE INDICATORS
D U ŠA N CR NO G O R AC , F E ĐA N E TJ A S OV

BELG RA DE U NI VERSITY – FACULT Y OF T R ANSPORT AND


T R A FFIC E NG INE ER ING
Introduction to AirTrafficManagement (ATM)

• Backbone of prosperity – air transport industry


• EUROCONTROL - Performance Review Commission (1998)
• Key Performance Areas – safety, capacity, environment and cost-
efficiency
• Safety = highest priority
• Safety indicators

2
Introduction to APACHE

3
Introduction to conflict

4
RA model architecture

5
Safety indicators

SAF-1: Number of Traffic Alerts warnings # TAs Count of TAs

SAF-2: Number of Resolution Advisors issued # RAs Count of RAs

SAF-3: Number of Near Mid Air Collisions – NMACs # NMACs Count of NMACs

SAF-4: Number of separation violations # SVs Count of separation violations

6
Approximation boundaries

7
FUZZY model architecture

input variables output variables


Fuzzifier Defuzzifier

Rules Inference engine

8
Input variables membership functiones

9
Output variables membership functiones

10
FLS Rules

11
FLS Rules Surface

12
Output (after normalization)

• interval [0 - 0.001] - SAF3


• interval [0 - 0.125] - SAF2
• interval [0.125 - 0.555] - SAF1
• interval [0 - 1] - SAF4

13
Numerical example - Scenarios dates

• 2016/07/28 - summer, middle demand


• 2017/02/20 - winter, low demand
• 2023/07/21 - summer, high demand

14
Results
SAF-1 SAF-2 SAF-3
Scenario
FLS TAM % FLS TAM % FLS TAM %
S001 94 91 3,3 34 35 -2,9 26 24 8,3
S101 1084 1023 6,0 277 269 3,0 182 173 5,2
S103 1704 1603 6,3 466 455 2,4 328 322 1,9
S105 1704 1603 6,3 466 455 2,4 328 322 1,9
S201 487 454 7,3 40 42 -4,8 27 27 0,0
S203 275 262 5,0 48 48 0,0 39 39 0,0
S205 631 597 5,7 104 102 2,0 73 68 7,4
S301 346 317 9,1 49 47 4,3 29 29 0,0
S303 275 291 -5,5 31 26 19,2 20 18 11,1
S305 549 510 7,6 89 87 2,3 53 47 12,8

15
Results – better view (1)
Scenario 001: total separation violation = 602

TA

RA module
RA
FLS

NMAC

0 20 40 60 80 100

16
Results – better view (2)

Scenario 101: total separation violation = 4726

TA

RA module
RA
FLS

NMAC

0 200 400 600 800 1000 1200

17
Results – better view (3)
Scenario 201: total separation violation = 2755

TA

RA module
RA
FLS

NMAC

0 100 200 300 400 500

18
Results – better view (4)
Scenario 301: total separation violation = 1752

TA

RA module
RA
FLS

NMAC

0 50 100 150 200 250 300 350

19
Conclusion
• Pros:
• TAM requirements – complicated preparation of flight parameters (position,
horizontal and vertical speed, heading, ...)
• FLS requirements – direct input from Separation Violation Detection Module
• Acceptable precision

• Cons:
• FLS is dependant on Separation Violation Detection Module

• Proposal for further work:


• Further testing of FLS
• Validation of FLS against real-life safety data

20
21
22

You might also like