Urban vs Suburban Distance-Dependent Shadowing Analysis

A MATLAB-based wireless propagation case study that models the shadowing parameter as a function of both distance and environment. The project compares urban and suburban scenarios using a modified log-distance path-loss model and evaluates shadowing behaviour, path loss, received power, and outage probability.

Research question

Traditional log-normal shadowing models commonly use a fixed shadowing parameter for an entire environment. This case study investigates whether a distance-dependent model can better represent the increasing uncertainty experienced as a signal travels through additional obstacles.

The proposed model is:

sigma_urban(d)    = sigma_0 + k_urban log10(d)
sigma_suburban(d) = sigma_0 + k_suburban log10(d)

where k_urban > k_suburban because dense urban environments generally introduce stronger blockage and greater signal variability.

Technical note: in standard log-normal shadowing notation, sigma is the standard deviation in dB; the variance is sigma^2. The submitted report's original wording is retained in the report PDF, while the repository documentation uses the more precise term.

Simulation parameters

Parameter	Urban	Suburban
Operating frequency	900 MHz	900 MHz
Distance range	100–2000 m	100–2000 m
Path-loss exponent	3.5	2.7
Base shadowing standard deviation	4 dB	4 dB
Growth factor	2	1
Transmit power	0 dBm	0 dBm
Outage threshold	-100 dBm	-100 dBm

Methodology

1. Define a 900 MHz cellular communication scenario over 100 m to 2 km.
2. Calculate mean path loss using the log-distance path-loss model.
3. Increase the shadowing standard deviation logarithmically with distance.
4. Generate Gaussian shadowing samples for each environment.
5. Compute received power and outage probability.
6. Compare urban and suburban behaviour through four plots.

Key observations

• Shadowing standard deviation increases faster in the urban scenario.
• Urban path loss is generally higher due to its larger path-loss exponent and stronger shadowing.
• Received power decreases with distance and fluctuates because of random shadowing.
• The simulation reported approximately 97.91% urban outage and 79.58% suburban outage for a threshold of -100 dBm.
• The high outage values are strongly influenced by the low transmit-power assumption, the 2 km range, and the absence of power control, diversity, or adaptive modulation.

Because shadowing is randomly generated, exact outage percentages can change between runs. The cleaned script uses a fixed random seed to make its output reproducible.

Repository structure

distance-dependent-shadowing-case-study/
├── README.md
├── .gitignore
├── results/
│   └── simulation_result.png
└── src/
    └── shadowing_analysis.m

How to run

1. Open MATLAB.
2. Open src/shadowing_analysis.m.
3. Click Run, or execute:

run('src/shadowing_analysis.m')

The script prints outage probabilities in the Command Window and saves the generated figure inside the results folder.

Scope and limitations

This is a proof-of-concept analytical simulation rather than a measurement-validated propagation model. The distance-growth coefficients are assumed values, the reference path loss is simplified, and the current implementation does not calibrate results against field measurements. A direct constant-sigma baseline comparison is described in the report methodology but is not separately implemented in the original submitted code.

Future work

• Validate the model using real received-signal measurements.
• Compare constant and distance-dependent shadowing under identical conditions.
• Add Rayleigh or Rician small-scale fading.
• Study power control, diversity, and adaptive modulation.
• Extend the analysis to LTE and 5G deployment scenarios.
• Perform Monte Carlo trials with confidence intervals instead of relying on one random realisation.

References

The report draws on standard wireless-propagation literature, including work by T. S. Rappaport, Y. Okumura, M. Hata, A. Goldsmith, H. Hashemi, and M. Gudmundson.