Refined Path Loss Prediction: Solving the ITU-R P.1546 Gap in Rural Australia

TL;DR

Accurate radio coverage prediction is the backbone of mobile network planning. This study reveals that the international standard Recommendation ITU-R P.1546-2 fails significantly in rural Australian environments, underestimating field strength by more than 10 dB. The authors introduce Model A, a localized adaptation that fixes stability issues with antenna heights and incorporates vegetation density data, resulting in a more robust and accurate prediction tool for rural CDMA/LTE/5G-low-band deployments.

Contextualizing the Problem

While urban environments attract the bulk of propagation research, rural areas present unique challenges: irregular terrain, varying vegetation types, and long-range macrocells. Existing "gold standards" like the Okumura-Hata model are frequently too optimistic, leading to coverage gaps. Conversely, the ITU-R P.1546 series—intended to be a universal successor—has introduced complexity that, in certain versions, leads to massive errors. Specifically, the definitions of Effective Antenna Height ($h_1$) can turn negative on slopes, causing the mathematical model to "explode" and predict physically impossible path losses (>50 dB error).

The Core Insight: Why P.1546-2 Fails

The paper highlights a critical change in version P.1546-2: the Terrain Clearance Angle (TCA) correction.

• The Flaw: P.1546-2 excludes TCA corrections for angles below 0.55°, making it overly pessimistic for flat rural terrain where Line-of-Sight (LOS) or near-LOS conditions prevail.
• The Result: It treats these areas as more obstructed than they are, leading to an average error of 11.11 dB.

Methodology: Engineering a Better Model (Model A)

1. Reciprocal Antenna Height ($h_e$)

Traditional definitions are non-reciprocal, meaning the prediction changes if you swap the transmitter and receiver. The authors propose an alternative $h_e$ based on the average difference of terrain height relative to a line connecting the ground levels of both points.

Fig 1: Standard effective height definition which the authors found unstable at short distances.

2. Vegetation-Specific Attenuation

The authors didn't just use clutter categories; they utilized 25m grid Perennial Crown Density data. They mapped 10 distinct vegetation types into two categories:

• Woodland: Attenuation = $(Densityimes20)/100$
• Shrubland: Attenuation = $(Densityimes10)/100$

Experimental Validation

Using a CDMA pilot scanner across 400km of Western Australia, the authors compared the different P.1546 versions against their proposed Model A.

Fig 2: A measurement track showing the "overshoot" of previous models (black/gray lines) vs the more accurate Model A (dashed line).

Quantitative Results

Critical Analysis & Conclusion

Model A’s success stems from fixing the "negative $h_1$" instability and re-introducing TCA logic that respects LOS conditions in flatlands. However, the study identifies a remaining challenge: Distance-Dependent Corrections. Both TCA and Mobile Antenna Height corrections appear to behave differently within 10 km of the base station than they do at 50 km.

Takeaway for Engineers: When using ITU-R P.1546 for rural planning, ensure your implementation handles negative effective heights gracefully and consider local vegetation density rather than just generic "clutter" categories. The "latest" version (P.1546-2 in this context) isn't always the best for every geography.