Evaluating Radar Performance Under Complex Electromagnetic Environment Using Supervised Machine Learning Methods: A Case Study
Yujian Pan, Jingke Zhang, Guo Qing Luo, Bo Yuan · 2018
Evaluating radar performance under complex electromagnetic environment is important for modern warfare. Traditional experimental method is expensive due to large number of experimental parameters. This paper presents a new machine learning based method via a case study. In this case, only a small number of experimental samples are required. After the data preprocessing and feature selection, the model for predicting the radar performance is learned by the machine learning algorithm. We compare six machine learning algorithms via cross-validation and find the multiple layers perceptron (MLP) possesses the highest prediction accuracy with a satisfied root-mean-square error (RMSE) of 1.77. The results of the paper exhibit the effectiveness of the machine learning based radar performance evaluation method.