A modified artificial neural network based prediction technique for tropospheric radio refractivity

Area of Science:

• Radio science and atmospheric physics.
• Artificial intelligence and machine learning applications.
• Telecommunications engineering.

Background:

• Radio refractivity significantly impacts radio system performance and can cause communication interruptions.
• Accurate prediction of radio refractivity is essential for reliable electromagnetic wave propagation.
• Existing methods require enhancement for precise refractivity forecasting.

Purpose of the Study:

• To develop and apply a modified artificial neural network (ANN) model for predicting radio refractivity.
• To assess the accuracy of the ANN model using historical meteorological data.
• To establish the relationship between atmospheric parameters and refractivity for improved radio system design.

Main Methods:

• A modified artificial neural network (ANN) model with data preparation, feature selection, and forecasting modules was developed.
• The ANN model utilized a sigmoid activation function and a multi-variate auto-regressive model for weight updates.
• Ten years of meteorological data (2002-2011) from Pakistan Meteorological Department (PMD) were used for training and prediction.
• Refractivity was estimated using the International Telecommunication Union (ITU) method and predicted for 2012.

Main Results:

• The proposed ANN model demonstrated high accuracy in predicting radio refractivity, with predicted and actual atmospheric parameters showing good agreement.
• Temperature and humidity were found to have a strong relationship with refractivity.
• Higher refractivity values were observed during the rainy season due to increased relative humidity.

Conclusions:

• The developed ANN model is effective for predicting radio refractivity and can be used to build a crucial database for radio communication systems.
• Understanding the influence of atmospheric conditions, particularly temperature and humidity, is vital for designing robust signal communication systems.
• The findings highlight the importance of accounting for refractivity variations in radio system planning, especially during adverse weather conditions.