Estimation of Window Frequency Around 30 GHz at Various Tropical Locations in India Using Machine Learning Techniques

The millimeter-wave frequency range between the two absorption maxima at 22.235 GHz and 60 GHz is critical for communication and radar applications. Atmospheric parameters such as temperature, water vapor, and oxygen play major roles in determining latitude-dependent window frequencies. Surface meteorological data from Weather Underground across Indian latitudes (8.48°N to 33.99°N) were analyzed to identify the frequency at which minimum signal attenuation occurs in the millimeter-wave band, between 22.235 GHz and 60 GHz, at various locations in India. Seasonal variations (both winter and rainy seasons) were considered, and shifts in attenuation minima due to water vapor and temperature were studied. Machine learning models were employed to improve the accuracy of analytical results. The highest value of the window frequency ("fwindow") during the winter season was observed at 31.75 GHz in Dibrugarh (27.47°N), Assam, India. In contrast, the lowest "fwindow" value in the same season was recorded at 30.76 GHz in Budgam (33.99°N), Jammu & Kashmir, India. During the rainy season, the maximum and minimum "fwindow" values were 31.43 GHz (Budgam, 33.99°N) and 30.98 GHz (Dibrugarh, 27.47°N), respectively. Notably, in Budgam, the "fwindow" frequency gradually shifts toward higher frequencies from winter to the rainy season, whereas the opposite trend is observed in Dibrugarh. It was also observed that during the winter season, the effect of water vapor density on the shift in window frequency is more significant than that of temperature. However, in the rainy season, water vapor has no notable impact on the frequency shift, with temperature changes playing the dominant role. These findings highlight the significant influence of water vapor and temperature in determining the window frequencies for microwave and millimeter-wave applications across different climatic zones in India. Therefore, these climatological parameters are essential for identifying precise values of the window frequency for specific locations throughout the year.