This paper presents a novel approach for frequency estimation of acoustic Doppler signals for use in speed and distance estimation of acoustic sources. The basis of the proposed approach is a nonlinear adaptive algorithm capable of tracking the amplitude, phase and frequency of nonstationary sinusoids. The approach provides a real time estimation of instantaneous frequency with very high time-frequency resolution and has a high level of noise immunity. This offers advantages over traditional Fourier-based methods as well as current Wigner-Ville based approaches seen in the literature. Examples using simulated and real signals will be shown to demonstrate the capabilities and strengths of this approach; comparisons to existing techniques are presented.