The research paper referenced above indicates a 0.5/sec sampling rate. Would a battery-backed inverter be able to sufficiently attenuate the power signature on this time-scale?