Impact of realistic amplification models on dynamic VBLAST optimization

The impact of realistic transmit amplifier models on MIMO systems´ capacity is investigated under dynamic VBLAST optimization. Classical optimization algorithms perform bit rate adaptation by constellation change as a function of the channel´s SNR under constant constellation RF powers. We consider variable constellation powers in optimizing a MIMO system´s capacity to reflect the more realistic conditions where different constellations require different amplifier back-offs. In addition, a realistic predistortion model and different system parameters (filtering, target BER, etc) are considered and their impact on the overall optimization is assessed. We show that, with such realistic amplification models, the VBLAST performances that can be obtained are significantly changed. The proposed model may also be used to optimize a transmitter´s power efficiency by reducing its energy consumption.