抄録
Blind source separation (BSS) is a technique for recovering original source signals from mixed signals without the aid on information of the source signals. The system restores the original source signals using the probability of the distribution of the original signal. In this paper, we consider the case where the original source signals are nonlinearly mixed. In general, the separation of the nonlinear mixed signals is difficult. In order to solve this problem, we apply a radial basis function (RBF) network with the nonlinear BSS system. The RBF network can approximate the nonlinear mapping. Therefore, the inverse mapping of the nonlinear mixture system is approximated by the RBF network. For the system to approximate the inverse mapping, it is necessary to adjust the parameters of the RBF network. We assume the original source signals to be independent of each other. In this case, if the mixed signals can be separated, the higher-order cross-moment of the output signals is decreased. In order to adjust the parameters of the RBF network, particle swarm optimization is used. We confirm the separation performance by numerical simulations. Simulation results indicate that the proposed approach has good performance.
