Implementing the Optimal CDMA Multiuser Detector with Hopfield Neural Networks

We propose a novel implementation of the CDMA optimal multiuser detector (OMD) based on the Hopfield Neural Networks (HNN). We show how the OMD’s objective function can be mapped to a HNN energy function and we evaluate the proposed detector via extensive simulations and theoretical analysis. Its is shown that the HNN detector outperforms by far the conventional detector and has a performance comparable to that of the OMD at a much lower computational cost.