To get a answer on the question "how accurate is my prediction" you need to calibrate your prediction model or perform at least a verification of your simulation algorthm.

In Commercial cellular communication networks it is common to use a emperic model such as COST-231. But in our case Radio Mobile is using the ITM PTP model. This model has the number of parameters reduced to a minimum.

Verfication of the prdeiction model in Radio Mobile can be done in many ways. Roger Coud? performed his calibration in 2003 (Radio Mobile Calibration). In his measurement a proper selection fo various terrains was done. Also the terrain types measured was documented. Because I want this website to be as complete as possible I have added his verification measurement to this website.

The verification or calibration of a RF prediction model starts with the definition of what that the model should archive. A model can not be accurate in all types terrain and over all distances. Some models perform good in urban environment (Mobile cellular) and some perfom better over bigger distances (broadcast). Other models are better suitable for mountain areas. There is no general model that can do everything with the highest possible accuracy.

The model that Radio Mobile has is the ITM PTP model (irregular terrain Model in Point to Point mode). This model is based on the model known as the Longley Rice model. This model is a general model suitable for a genaral environment. The model will not produce excellent performance in dense urban areas as it is not suited for that porpose. Also the model will not be able to accurate predictions for hand held radios wich are below the surface of the clutter.

The model in Radio Mobile will give good results for radios and antennes installed on mobiles and radios and antennas installed in masts. To use Radio Mobile for calculations in situations where a hand held radio is under the surface of the clutter other precautions have to be taken. These will be taken in the Link budget for example.

A RF propagation model 'predicts' the fieldstrength that probably might be measured at a specific point. It is never telling you the truth. As the model is a theoretic approach, where assumptions are catched in a formula, It is always possible that mother nature fools you. Even when the used terrain data in both clutter and DEM is as accurate as possible ther is a chance that there is a error in the predicted value with respect to the real measured value. Please remeber that even the measured fieledstrength is a snapshot in time and will change constantly.

The acuracy of a calculation model predicting RF propagation is displayed in 2 main KPI:

- Average error of predicted to the measured fieldstrength
- Standard deviation
- Correlation of the predicted to the measured fieldstregth

Based on these KPI we can identify 2 methodes of determining the accuracy of the model:

- Mathematical way, Calculating the required KPI
- Optical way, Look if the prediction result matches with the measured values.