An artificial neural network (ANN) based equalizer for indoor optical wireless communication systems is presented in this paper. Although pulse position modulation (PPM) scheme is power efficient compared with on-off keying (OOK) modulation in the line-of-sight links, its performance is unacceptable in diffused channels. So that a novel ANN equalizer has been proposed that not only outperforms other adaptive equalization, but also has fewer computational complexities compared with recent ANN equalizers. Furthermore, the proposed ANN equalizer makes the PPM scheme appropriate to diffused links. Since the simulation results indicates that there is about 10 dB performance improvement in neurally equalized PPM system compared to similar case in OOK scheme.

Keywords