Low Computational Enhancement of STFT-Based Parameter Estimation

Abstract

Short-time Fourier transform (STFT) is one of the most widely used tools to analyze frequency and phase of local sections of time-varying signals using a time window function. Since the fixed window size of STFT causes limitation in time-frequency representation (TFR), adaptive STFT (ASTFT) techniques have been studied to adjust the window size depending on the local signal characteristics. However, ASTFT techniques suffer from heavy computational complexity in general. In this paper, we propose low computational enhanced temporal and spectral parameter estimation techniques for STFT output of radar signals; forward and backward STFT enhancement techniques. The proposed forward STFT enhancement technique is to enhance the spectral resolution of a received radar signal by multiple times using a number of successive STFT outputs without applying additional STFT with wider time window. On the other hand, the proposed backward STFT enhancement technique is to improve the temporal resolution of a received radar signal using an STFT output with a wide time window, without applying multiple STFTs with smaller time window. The performance of the proposed techniques are theoretically analyzed, and the advantage of the proposed techniques to the conventional techniques in terms of computational complexity is demonstrated with numerous simulations.