The Loran Propagation Model: Development, Analysis, Test, and - - PowerPoint PPT Presentation
Avionics Engineering Center The Loran Propagation Model: Development, Analysis, Test, and Validation Janet Blazyk, Ohio University Dr. Chris Bartone, Ohio University Frank Alder, Ohio University Mitch Narins, Federal Aviation Administration
2
Accurate navigation using Loran requires precise timing
Mis-modeling or erroneous measurements of Additional
To support RNP 0.3 for non-precision approach and
This requirement can be met by providing accurately
For enroute navigation, error tolerances are more lenient,
Hence a large-scale ASF map of predicted ASF values
3
The Loran signal may propagate over a great distance,
Delays due to propagation through the atmosphere and
ASF delays are affected by:
Ground conductivity (the most significant factor) Changes in terrain elevation Receiver elevation Temporal changes (seasons, time-of-day, local weather)
Additionally, various other factors such as system timing
4
Computer program to predict
Formerly known as BALOR. Originally developed by Paul
Maintained and improved by
Models Loran groundwave
Performance needs to be
5
System to accurately measure
Developed by Reelektronica. Utilizes LORADD eLoran
A simulated Loran pulse is
Calibrated Loran H-field
A small timing offset is possible
The TMS rack-mounted in Ohio University’s King Air C90 Aircraft
6
Five days of flights over
Flights included:
Approaches at certain
Enroute legs between
Flights over ocean and
Altitude tests Calibration circles
Loran and GPS data
ASFs predicted by LPM
Airport Name ID Location Ohio University Airport UNI Albany, Ohio Norwalk-Huron County Airport 5A1 Norwalk, Ohio Craig Municipal Airport CRG Jacksonville, Florida Bay Bridge Airport W29 Stevensville, Maryland Atlantic City International Airport ACY Atlantic City, New Jersey Monmouth Executive Airport BLM Belmar/Farming- dale, New Jersey Portland International Jetport PWM Portland, Maine
Significant airports
7
Key airports
Background
12 separate
8
Approaches at CRG (racetrack between CRG and Point A) Inland to Point B Across coast to Point C (along radial from Malone) Back to land at CRG
9
10
Path from Xtm is long, but
The large central peak
Differences are in the
Other plot features are
11
Path from Xtm is relatively
Measured and modeled
Peak ~ 5800 corresponds
coastal crossing Closest to Xtm
12
Approaches at PWM Over ocean to point E Out to point F at 6000 m
Flight 11 – Altitude
Descend to 2000 m Return to point E Climb to 6000 m again Pass over Nantucket Continue on to touchdown
13
14
The path from the Xtm is
Large peak ~ 8500 and
Match between LPM and
altitude drop
15
Very long path from Xtm;
Larger ASFs over land LPM predicts a peak at ~
Differences are ~ 0 to 0.4 µs
island altitude drop
16
Long flight over land Enroute altitude around 5000 m Low mountains for first half of
Flight 1 – Altitude
17
18
Path from the Xtm is
Path is all over land
Up to 1 µs offset when
Good match where there
19
The path from the Xtm is
The path is longest at the
Modeled ASFs follow the
offset of about 1.5 µs
near the start
decreasing to about 0.6
µs near the end.
20
Comparison of modeled and
Good agreement when path
from Xtm is short or mostly
Modeled results always too
low for a long, land path.
All valid data points over the
The modeled ASF falls
ASF Offsets vs. Measured ASFs
21
ASF Offsets vs. Land Distance ASF offsets is related to
The slope of the line in this
Need to determine if bias
For example, bias can be
22
A complex factor is used
Correction is a function of
Height correction was
While this correction may
Flight 11 – Nantucket
Height correction improvements
23
To compensate for
Traditionally, αe = 4/3 for
What is best for Loran? LPM has used 4/3 and 1.14 in
Examining the ASFs over a
Flight 4 – Nantucket
Effect of αe over a long ocean path
24
An “ASF offset bias,” amounting to 1.1 ns per km of
After compensating for the offset bias, ~ 0.6 µs of
Factors that contribute to ASF modeling errors:
Ground conductivity map is not very detailed or accurate. Height correction could be improved. Terrain slope correction could be improved.
Note that system timing errors, measurement system
25
The Loran Propagation Model is efficient and robust, but
The TOA Measurement System produces precise and
Comparing LPM and TMS ASF results revealed a offset
The remaining error, from all causes, is about 0.6 µs. Additional field testing and model refinement should
26
27
28
29
Path from Xtm is relatively
Good overall match. Difference of 0 to 0.1 µs. Approaches at beginning Large central peak is due to
Low ASFs over the ocean.
30
Path from Xtm is relatively
Low ASFs on right
Differences are in the
Other plot features are
31
Path from Xtm is fairly long
Three approaches on left. ASFs get lower over the
Decrease in altitude seen
The small peak ~ 8500
Differences ~ 0.1 to 0.2 µs.
island altitude drop
32
The path from Xtm is
Measured and modeled
Shape mismatches ~ 0.4
TMS does register a small
island
33
The path from the Xtm is
As the flight progresses,
Differences are near 1
34
The length of the path
The path is mostly over
Measured and modeled
35
Comparison of
The plot shows
This radial skims the
Some of the
Flight 4 – Carolina Beach Radial to point at time 4920 s
36
Again, the plot shows
In this case, we are
The “noise” is due to
The ASF measured by
Flight 1 – Malone Radial to point at time 400 s