Source-linked AI summary
Correlation Analysis between Terrain Features and eLoran Spatial ASF according to DEM Spatial Resolution
Jaewon Yu, Pyo-Woong Son
TL;DR
Spatial ASF is a major source of eLoran positioning error, and the effect of terrain-data resolution on terrain–ASF relationships has not been sufficiently investigated. This study compares path-based terrain features from 30 m and 90 m DEMs with measured ASF in Incheon and Pyeongtaek; both features correlate strongly with ASF, with stronger correlations generally from the higher-resolution DEM, particularly for slope.
Problem
The effect of terrain-data spatial resolution on terrain–spatial ASF relationships has not been sufficiently investigated, despite terrain inputs being used in data-driven ASF analysis.
Source: §I · pp. 1–2Method
The study extracts path mean elevation and slope from 30 m NASA SRTM and 90 m NGII DEMs along Gwangju-to-measurement paths and evaluates their correlations with measured ASF.
Source: §I · p. 2, B. Terrain Feature Extraction along… · p. 2Results
The NASA SRTM vs. NGII Pearson correlation coefficient difference was 0.207 for path mean slope, versus 0.043 for path mean elevation; both features showed strong correlations with spatial ASF.
Source: B. Path Mean Slope and Correlation… · pp. 3–4, A. Path Mean Elevation and Correlation… · p. 3Takeaways & Limitations
DEM resolution has greater influence on slope-related features, highlighting the importance of high-resolution terrain data for machine-learning-based spatial ASF prediction.
Source: §IV · p. 4
Abstract
from arXiv · showhide
The ASF is a major source of positioning error in eLoran system and is affected by terrain characteristics along the signal propagation path. This study analyzes the correlations between spatial ASF and path-based terrain features using NASA SRTM 30m and NGII 90m DEMs. Path mean elevation and path mean slope were extracted along the propagation paths from the Gwangju eLoran transmitter to measurement locations in the Incheon and Pyeongtaek port areas. The results show that both terrain features are strongly correlated with spatial ASF, with the higher-resolution NASA SRTM DEM generally exhibiting stronger correlations, particularly for path mean slope. These findings demonstrate the importance of DEM spatial resolution in terrain-based spatial ASF analysis and provide a basis for selecting terrain data for future machine-learning-based ASF prediction
I. INTRODUCTION
eLoran is considered a resilient alternative to vulnerable GNSS, but spatial ASF is a major source of positioning error. The study examines whether DEM resolution affects terrain–ASF correlations to inform future machine-learning-based prediction.
I. INTRODUCTION
- Spatial ASF is difficult to predict with simple theoretical models because ASF reflects terrain and surface variations along the propagation path and limits eLoran positioning accuracy.Unlike temporal ASF, spatial ASF varies by location and is relatively stable over time at a given location.Source: §I · p. 1
- Terrain-data resolution has not been sufficiently investigated as a factor in the relationship between terrain characteristics and spatial ASF.The question matters for data-driven ASF methods that use terrain information as input, since input-data characteristics and quality can affect training and prediction performance.Source: §I · p. 2
- The study compares path mean elevation and slope from 30 m and 90 m DEMs against measured ASF maps for Incheon and Pyeongtaek, aiming to inform future prediction-model terrain inputs.The maps are merged, and each propagation path runs from the Gwangju eLoran transmitter to an ASF measurement location.Source: §I · p. 2
A. Spatial ASF Map and Multi-Resolution DEM Dataset
The study uses measured spatial ASF maps for Incheon and Pyeongtaek referenced to the Gwangju transmitter, alongside two DEMs at different native resolutions.
A. Spatial ASF Map and Multi-Resolution DEM Dataset
- The measured ASF reference dataset merges the Incheon and Pyeongtaek maps generated with respect to the Gwangju (9930W) eLoran transmitter.The merged map is shown in Fig. 1.Source: A. Spatial ASF Map and… · p. 2, Figure 1 · p. 2
- The merged grid contains 18,995 points at approximately 500 m spacing, spanning 36.935–37.655°N and 126.200–126.850°E; ASF values are in µs.The grid has 145 latitude points and 131 longitude points.Source: A. Spatial ASF Map and… · p. 2
- Terrain analysis uses NASA SRTM at 30 m and the NGII DEM at 90 m, with both datasets at native resolution and the same analysis procedure.The NASA DEM is in GeoTIFF geographic coordinates; the NGII DEM uses projected coordinates based on EPSG:5186.Source: A. Spatial ASF Map and… · p. 2
B. Terrain Feature Extraction along the Propagation Path
For each ASF measurement, the study samples terrain along the propagation path from the Gwangju transmitter and derives mean elevation and mean slope features.
B. Terrain Feature Extraction along the Propagation Path
- A great-circle path is generated from the Gwangju transmitter to each ASF measurement point, and elevation and slope are extracted along the path.This accounts for terrain along the full path, not only at the receiver.Source: B. Terrain Feature Extraction along… · p. 2
- Bilinear interpolation estimates elevations where measurement points or path samples do not coincide with discrete DEM grid locations.The interpolation uses the four surrounding grid points.Source: B. Terrain Feature Extraction along… · p. 2
- Path mean elevation is the average of sampled path elevations, while path mean slope is the average of sampled terrain-slope values.Terrain slope is calculated from the numerical gradient of each DEM at native resolution and expressed in degrees.Source: B. Terrain Feature Extraction along… · pp. 2–3
A. Path Mean Elevation and Correlation with Spatial ASF
Path mean elevation is strongly correlated with spatial ASF, and its correlation is higher when calculated from the 30 m NASA SRTM DEM than from the 90 m NGII DEM.
A. Path Mean Elevation and Correlation with Spatial ASF
- NASA SRTM path mean elevation had Pearson correlation 0.870 and Spearman correlation 0.821 with spatial ASF; NGII values were 0.827 and 0.780, respectively.The study reports a strong correlation for both DEMs.Source: A. Path Mean Elevation and Correlation… · p. 3
- Fig. 3 and Table II are identified as presenting path mean slope versus spatial ASF and correlations for the two DEMs.These captions identify the visual scope, but the supplied passages do not report their comparison values.Source: Figure 3 · p. 3, Table II · p. 3
- The higher-resolution NASA SRTM path mean elevation showed a higher correlation with spatial ASF than the NGII result.The authors interpret the result as showing that terrain information along the entire propagation path is effective in explaining spatial ASF.Source: A. Path Mean Elevation and Correlation… · p. 3
B. Path Mean Slope and Correlation with Spatial ASF
Path mean slope is strongly correlated with spatial ASF for both DEMs, with a larger correlation advantage for the higher-resolution NASA SRTM DEM than for path mean elevation. The authors interpret this difference as reflecting slope’s sensitivity to fine-scale terrain variation.
B. Path Mean Slope and Correlation with Spatial ASF
- Path mean slope correlated strongly with spatial ASF: Pearson and Spearman coefficients were 0.858 and 0.814 for NASA SRTM, versus 0.651 and 0.641 for NGII.The higher-resolution NASA SRTM DEM produced the higher correlation for both measures.Source: B. Path Mean Slope and Correlation… · pp. 3–4
- For path mean slope, NASA SRTM’s Pearson correlation with spatial ASF was 0.858 versus NGII’s 0.651, a 0.207 gap, more than four times path mean elevation’s 0.043 gap (0.870−0.827).Source: B. Path Mean Slope and Correlation… · p. 4
- The authors attribute slope’s greater resolution sensitivity to coarser grid spacing smoothing fine-scale terrain undulations.Slope represents the spatial rate of elevation change, whereas path mean elevation reflects absolute elevation over a broader path section.Source: B. Path Mean Slope and Correlation… · p. 4
IV. CONCLUSION
Using measured spatial ASF maps and DEM data from Incheon and Pyeongtaek, the study found strong, statistically significant correlations with path mean elevation and slope. NASA SRTM 30m generally correlated more strongly than NGII 90m, particularly for slope.
IV. CONCLUSION
- The analysis used measured spatial ASF maps and DEM data from Incheon and Pyeongtaek.Source: §IV · p. 4
- Path mean elevation and path mean slope showed strong, statistically significant correlations with spatial ASF, with NASA SRTM 30m generally stronger than NGII 90m, particularly for slope.The conclusion links slope’s stronger resolution sensitivity to its sensitivity to local terrain variations.Source: §IV · p. 4
- Future work will combine effective ground conductivity derived from land-cover data with elevation and slope features to develop a machine-learning model for Spatial ASF map prediction.Source: §IV · p. 4