Interference Mitigation in Multi-radar Environment Using LSTM-Based Recurrent Neural Network

External disturbances, such as interference, have a significant impact on the functionality of radio detection and ranging (radar) systems, which are employed for the identification, ranging, and imaging of target objects. As radar systems are increasingly adopted across various sectors for different applications, it is essential to handle interference issues appropriately to mitigate false detections, poor signal-to-noise ratio (SNR), and reduced resolution. In the current paper, we introduce a Long Short-Term Memory (LSTM)—based multi-layer recurrent neural network (RNN) to tackle interference problems in a multi-radar setting. In the simulation, a 4-layered LSTM-RNN is trained with 50 different chirp rate interference signals. The efficiency of the introduced interference mitigation technique is evaluated by testing the randomly selected coherently interfered signals, non-coherently interfered signals, and a combination of both on the trained model. The LSTM-RNN effectively suppresses ghost targets in the range profile in the case of coherently interfered signals. Furthermore, the LSTM-RNN enhances the signal-to-interference noise ratio (SINR) by > 18dB in all cases. Thus, the proposed LSTM-RNN offers a promising solution to improve the accuracy and reliability of radar operation in multi-radar environments.