An electronic artificial neural network architecture is presented for estimating the parameters of a sine wave distorted by a DC exponential signal and corrupted by noise. In the proposed approach a finite set of sampled data from noisy measurements is used to implement Lp—norm (with 1 ≤p≤ ∞), The standard least-squares (L2 — norm) criterion is considered as a special case. Mathematical algorithms are presented in detail and associated architectures of analogue artificial neural networks are proposed. Extensive computer simulations are used to demonstrate the validity and performance of the proposed algorithms and neural network realizations. The proposed methods seems to be particularly useful for real-time, high-speed and low-cost estimations of parameters of sinusoidal signals.