Research Projects

Computer-based analyses are more extensively used to understand the behaviour of engineering systems. In recent years, statistical approximation models have been used to minimize the time and computational resources necessary for analysis. The development of an approximation model involves data collection, learning from the collected data and construction of Artificial Intelligence (AI) Models based on the learning. The main theme of this work is to construct AI models like LSTM, ConvLSTM, Bi-LSTM and SVM along with the incorporation of optimizers like Adam, Nadam, Rmsprop, GA, etc., to predict the ionospheric TEC and also to correct the range errors. This model is based on input parameters namely, hour, Sunspot Number, Solar wind, Solar flux index (F10.7), F 245 MHZ Values, Magnetic index (Kp & Ap) and DsT data. Ionospheric TEC will be obtained from the Ionolab and other GNSS network stations. F 245 MHz data will be collected from the experimental setup. Also, to cover all regions of our country/world and also to evaluate the performance of proposed AI based models, ionospheric TEC will be predicted by using the AI models for different latitudes, during different seasons and during different solar activity conditions and also, TEC will be predicted during solar flare, storm days based on the previous days of TEC data obtained from IGS network stations. Apart from that, the predicted TEC will be analyzed to find out the possible occurrence of earthquake in advance. Various implementations of these four algorithms from state-of-the-art AI libraries such as Keras, Tensorflow, PyTorch, etc. will be investigated as part of this project. This study can be used for characterizing the ionosphere and also it can be used for possible detection of earthquakes in advance. This proposed study can be extended to multiple receiver data from reference stations to construct an ionospheric model for the prediction of TEC over the Indian Region and different parts of the world. In the context of a fully realized GNSS Receiver and GNSS Simulator installation, the proposed TEC Prediction and ionospheric study will independently assess the performance of several navigation systems.The input parameters such as F10.7, SSN, Ap, Kp, Dst, Solar wind will be collected from NASA website, TEC data for different stations will be obtained from GNSS receivers and Ionolab. F 245 MHz data will be collected from the experimental setup. This data can be used for forecasting the space weather. By using these data sample databases will be created, which in turn used for constructing the AI models incorporated with optimizers for the prediction of TEC. The predicted TEC from AI models will be compared with other models.