Radar Data Reconstruction for FMCW Interference Mitigation

The patent applies preliminary action by performing interference mitigation through iterative reconstruction before final target detection. The method pre-processes the radar data by identifying interfered samples using a mask and reconstructing them through iterative optimization, thereby eliminating interference artifacts before they can affect detection accuracy. This preliminary processing step ensures that subsequent detection operates on clean data, resolving the contradiction between achieving high detection accuracy and managing computational complexity.

The patent applies the discarding and recovering principle by first discarding (setting to predetermined value) the interfered samples identified by the mask to eliminate interference, then recovering the discarded information through iterative reconstruction. The reconstruction process uses the masked data and applies optimization algorithms to estimate and restore the original signal values at the masked positions. This two-step process of discarding interference-affected samples and then recovering their true values resolves the contradiction between removing interference and preserving signal information.

The patent applies periodic action through iterative reconstruction, where the processing is divided into multiple discrete iterations that periodically transform data between time domain and range-Doppler domain. Each iteration performs a subset of operations (e.g., Fourier transform, thresholding, inverse transform) and then repeats with updated data. This periodic, modular approach allows the system to achieve high processing gain through multiple passes while managing computational load by breaking the processing into manageable periodic steps rather than requiring all operations to execute simultaneously, thus resolving the contradiction between processing gain and processing time.