Research On Acoustic Weak Signal Processing And Application Technology For Intelligent System

Sensing technology empowers intelligent systems to interact with the physical world.With the ability to sense the environment,intelligent systems can perceive information from the physical world,allowing them to better understand and interact with their surroundings.Acoustic sensing aims to utilize acoustic signals in the environment to perceive the physical world.By establishing mapping models from acoustic signals to physical quantities,we can infer certain physical quantities based on specific parameters of the measured signals.Acoustic signals have gained widespread attention in the field of Internet of Things(Io T)due to their ubiquitous presence in the physical world,abundant hardware support,and excellent sensing capabilities.In practice,acoustic sensing encounters significant challenges in detecting weak signals due to interference and noise.This problem manifests in the following ways:(1)Low sensing capacity: Since humans and machines share the same acoustic field,their sounds can interfere with one another,resulting in difficulty detecting and sensing any object within the acoustic channel.(2)Short sensing range: The low transmission power,high noise floor,and severe signal attenuation intrinsic to acoustic signals lead to a low signal-to-noise ratio,which limits the coverage of acoustic sensing.(3)Few sensing objects: Acoustic sensing requires the target to emit or reflect sound.Therefore,objects that cannot produce sound and are not subject to external sound cannot be sensed by current acoustic sensing systems.This thesis aims to address the research problems mentioned above by expanding the sensing capacity,extending the sensing range,and developing more sensing objects.The main research objectives are as follows:(1)To investigate acoustic passive sensing technology and enable concurrent localization of multiple sound sources under signal interference,thereby expanding the acoustic sensing capacity.(2)To study long-distance acoustic active sensing and achieve real-time detection of signals with nonlinear distortion under a low signal-to-noise ratio,thereby extending the acoustic sensing range.(3)To explore acoustic sensing technology based on vibration side channels and enable acoustic sensing systems to cover small appliances that are often overlooked,thereby developing more acoustic sensing objects.Experimental results show that our proposed solutions can:(1)Achieve concurrent localization of more than three sound sources,with a median error of 0.66 m.(2)Achieve localization of moving targets up to 120 m away,with a relative error of 0.5% and a refresh rate of 20 Hz.(3)Achieve acoustic sensing and communication for small appliances,with a sensing distance of 10 m and a throughput of up to 195.8 bps.