— Multi-rotor aerial autonomous vehicles (MAVs)primarily rely on vision for navigation purposes. However,visual localization and odometry techniques suffer from poorperformance in low or direct sunlight, a limited field of view,and vulnerability to occlusions. Acoustic sensing can serve asa complementary or even alternative modality for vision inmany situations, and it also has the added benefits of lowersystem cost and energy footprint, which is especially importantfor micro aircraft. This paper proposes actively controlling andshaping the aircraft propulsion noise generated by the rotors tobenefit localization tasks, rather than considering it a harmfulnuisance. We present a neural network architecture for selfnoise-based localization in a known environment. We showthat training it simultaneously with learning time-varying rotorphase modulation achieves accurate and robust localization.The proposed methods are evaluated using a computationallyaffordable simulation of MAV rotor noise in 2D acousticenvironments that is fitted to real recordings of rotor pressurefields.