Akçapınar, KudretÖnhon, Naime ÖzbenGürbüz, Özgür2025-02-202025-02-2020231937-8718https://doi.org/10.2528/PIERC23012305https://hdl.handle.net/20.500.12846/1769—In this paper, an artificial neural network (ANN) based approach is proposed for the estimation of the target angle using Multiple Input Multiple Output (MIMO) radars operating in Frequency Modulated Continuous Wave (FMCW). The proposed technique operates in two stages, with the first stage being the formation of the range profile at each MIMO element via Discrete Fourier Transform (DFT) and the second stage being the estimation of the target azimuth angle via an artificial neural network. The range profile formed in the first stage is fed to the second stage as a single snapshot angle measurement. The performance of the proposed technique is apprised with other existing methods under different Signal-to-Noise Ratio (SNR) conditions and measurement model uncertainties. The simulations performed show that the learning capability of the model strongly hinges on SNR conditions, and the learning process is ameliorated as SNR in training data increases as anticipated. Under low SNR conditions, the proposed technique performs better than other techniques in terms of Mean Square Error (MSE). We have also shown that our solution remains unaffected by the model uncertainties as it fully relies on the calibration data, while the performance of the model-based angle estimation techniques dramatically degrades as the uncertainty in the underlying model grows. © 2023, Electromagnetics Academy. All rights reserved.eninfo:eu-repo/semantics/openAccessDiscrete Fourier transformsFrequency estimationFrequency modulationLearning systemsMean square errorMIMO radarMIMO systemsSignal to noise ratioUncertainty analysisAngle estimationConditionEstimation techniquesFrequency modulated continous wavesModeling uncertaintiesMultiple-input multiple-output radarsNetwork-basedPerformanceRange-profilesTarget angleNeural networksArticle13410.2528/PIERC23012305119130Q32-s2.0-85164273748