Propagation Model

A propagation model is how planners predict, before any hardware exists, how much a signal will fade between a proposed site and the users around it. You can't measure a network you haven't built, so you feed the model frequency, antenna heights, distance, and terrain type, and it returns a path-loss estimate that drops straight into the link budget and the coverage map.

The classic empirical models came from fitting curves to real measurement campaigns. Okumura-Hata is the venerable one, tuned for the lower bands and split into urban/suburban/rural variants. COST-231 Hata extended it upward to around 2 GHz. For modern work, the 3GPP 38.901 channel models are the reference — they're built for the higher frequencies 5G uses (up to mmWave) and model the channel in much more statistical detail. The thing to keep honest about is that all of these are estimates. Empirical models are only as good as the environments they were calibrated against, and serious deployments still validate predictions with drive tests, then tune the model to the local clutter before committing to a build.