Stochastic Resonance Pre-Processing for Estimating Doppler Frequency Shift under Low SNR Conditions

Abstract

According to the demand of the Doppler frequency shift estimation under low signal-noise-ratio (SNR) conditions, the stochastic resonance (SR) technique from the nonlinear science is applied to signal pre-processing and a novel algorithm with the level crossing rate (LCR) estimation is proposed in this paper. Theoretical derivation demonstrates the proposed algorithm can effectively improve SNR and bandwidth ratio of the received signal, thereby reducing estimated error of the LCR algorithm. Moreover, after analysis and discussion of the optimal sampling frequency in stochastic resonance pre-processing (SRP), the best relationship of the sampling frequency with input SNR is obtained. The Monte-Carlo simulation results show that comparing with current algorithms, the proposed algorithm improves estimated performance of 2~4dB without increasing computational complexity under low SNRs, it verifies the consistency with the theoretical conclusions.