Optimized Training for Net Energy Maximization in Multi-Antenna Wireless Energy Transfer Over Frequency-Selective Channel

This paper studies the training design problem for multiple-input single-output (MISO) wireless energy transfer (WET) systems in frequency-selective channels, where the frequency-diversity and energy-beamforming gains can be both reaped to maximize the transferred energy by efficiently learning the channel state information (CSI) at the energy transmitter (ET). By exploiting channel reciprocity, a new two-phase channel training scheme is proposed to achieve the diversity and beamforming gains, respectively. In the first phase, pilot signals are sent from the energy receiver (ER) over a selected subset of the available frequency sub-bands, through which the ET determines a certain number of “strongest” sub-bands with largest antenna sum-power gains and sends their indices to the ER. In the second phase, the selected sub-bands are further trained by the ER, so that the ET obtains a refined estimate of the corresponding MISO channels to implement energy beamforming for WET. A training design problem is formulated and optimally solved, which takes into account the channel training overhead by maximizing the net harvested energy at the ER, defined as the average harvested energy offset by that consumed in the two-phase training. Moreover, asymptotic analysis is obtained for systems with a large number of antennas or a large number of sub-bands to gain useful insights on the optimal training design. Finally, numerical results are provided to corroborate our analysis and show the effectiveness of the proposed scheme that optimally balances the diversity and beamforming gains in MISO WET systems with limited-energy training.