Manipulation of Acoustic Focusing for Multi-target Sensing with Distributed Microphones in Smart Car Cabin

Smart car cabins have drawn great attention in the past few years to enable a wide range of applications, including in-car entertainment, road noise cancellation, and voice control. These applications have been underpinned by the addition of a very large number of acoustic modules in the cabin, which has further opened up new opportunities for acoustic contactless sensing in smart cabins. Different from the traditional smartphone or smart speaker for acoustic sensing, in this paper, we envision to manipulate the inherently embedded speakers and microphones in car cabin to produce 3D spatial signal focusing for sensing. We propose the new concept of in-car occupancy grids, which enables target identification in 3D cabin space. In order to accomplish signal focusing in a small spot region, we establish the distributed acoustic 3D spatial beamforming model. To further address the phase ambiguity induced by large microphone spacing, we design dual frequency transmitted signals and "virtually"create a low-frequency signal to sense targets, which characterizes the highly sensitive and noise-free properties. We build a distributed acoustic sensing system in car and conduct both benchmark experiments and real-world application studies. The results demonstrate that the proposed system can be applied to various sensing tasks, which has superior performance in terms of accurate occupancy detection, multi-target sensing, very subtle motion detection, and interference resistance.