RaSim: A Radar Data Simulator Using Machine Learning

Radar simulation refers to the creation of artificial radar data to model and analysis of the behaviour and performance of radar systems. This synthetic data is generated based on some mathematical models and algorithms that replicate the physical phenomena radar systems encountered in real-world scenarios. This paper proposes a novel radar simulator called RaSim using polynomial regression. We have used the two types of radar datasets to prepare the model. Our model employs a robust local regression smoothing technique to filter and smoothen the noisy data. The smoothed data is then fed into the polynomial regression model for training and simulated data is obtained. While training the model, we used techniques like hyperparameter tuning, incremental learning and segmented polynomial fitting to improve the accuracy of our proposed model. Using mean square error to update the gradient descent makes the model even more resilient.