Airborne Ultrasound

Airborne ultrasound is an emerging field exploring the use of ultrasonic waves for three-dimensional imaging and object detection in air, primarily for robotic applications and privacy-preserving human sensing. Current research focuses on developing high-resolution imaging systems using large microphone arrays and advanced signal processing techniques like beamforming and novel triangulation methods based on time-of-flight and parallax. These advancements leverage machine learning models, including variations of U-Nets and variational autoencoders, to improve image quality and enable tasks such as human segmentation and instance segmentation from ultrasound data, offering a potential alternative to camera-based systems where privacy is a concern.