KOCHI: Cochin University of Science and Technology (Cusat) has been awarded a patent for developing a new machine learning-based technology to extract rainfall information from atmospheric radar signals.
This method integrates distribution-based signal processing algorithms and advanced machine learning techniques to enhance the performance of a 205 MHz stratospheric-to-tropospheric (ST) radar. Radar data often contains echoes from both clear skies and raindrops, making it difficult to accurately obtain atmospheric parameters. To address this challenge, the research team developed a hybrid adaptive bigaussian fitting algorithm that can effectively separate rain and clear-sky echoes within the troposphere below 5 km, even at different rainfall intensities.
Furthermore, by applying a Gaussian mixture model (GMM), this technique demonstrated high efficiency in accurately identifying spectral components even under complex atmospheric conditions. The stochastic nature of the GMM allows accurate identification of rain and air echoes.
This patented research is the result of a collaboration between ToCH Institute of Science and Technology Assistant Professor Dhanya R (former Research Fellow, Department of Electrical Engineering), Faculty of Engineering Professor Anju Pradeep, Radar Center Director Dr Abhilash S, and Radar Center Research Fellow Abhiram Nirmal.
