A neural network for estimating the parameters of multiple sinusoids

The authors present an approach, using a circuit based on the concept of neural networks, for estimating the parameters of multiple sinusoids. The computation of the parameters in the case of analog circuit realization is extremely fast because the network's speed is limited only by the time constants of the operational amplifiers. Several examples are used to test the efficiency of the method by simulation. This is done for the noiseless case as well as for the case where noise is present. The proposed artificial neural network (ANN) exhibits good performance for near continuous frequencies and for the case where only a small part of the signal is covered by the sampling points.<>