Complex-Valued Neural Network Based Detector for MIMO-OFDM Systems

摘要

The MIMO-OFDM detector with near-optimum performance and low complexity is valuable. This paper uses the complex-valued neural network to implement a detector for MIMO-OFDM systems according to the channel matrix in frequency domain which is composed of complex number. The parameters related to the network stability are given. The output of proposed detector is just the output of all neurons. This detector has advantage of low complexity because neuro-computing is a computational process with low-complexity. The simulations show that the proposed detector is one which can obtain a near-optimum performance, but with a low-complexity.