Research On Eavesdropping Technology Based On Millimeter Wave Radar

Sound eavesdropping refers to a third party covertly listening in on private conversations that are not intended for public disclosure.In contemporary society,eavesdropping is considered to be an important technical means for obtaining military,political,economic,scientific,technological,and industrial intelligence from other countries.Eavesdropping technology is also widely used in many special scenarios,such as an effective criminal investigation tool that plays an important role in cracking down on major crimes.However,the widespread use of soundproof materials in modern buildings and the development of anti-eavesdropping detection technology have significantly reduced the possibility of traditional sound eavesdropping.Therefore,advanced eavesdropping technology has been continuously developed in recent years,attempting to break through existing eavesdropping defense mechanisms from multiple directions.With the widespread application and deployment of smart devices,smart device-based eavesdropping technology has been extensively researched in recent years.Its core idea is to capture sound vibrations using different types of sensors.Some experts eavesdrop on users’ speech information through built-in sensors.Research shows that sound can cause slight vibrations in the gyroscope,which can be used to reconstruct simple vocabulary.Sound also has a slight impact on the accelerometer embedded in the device,and some speech information can be recovered using accelerometer data.However,the eavesdropping method based on built-in sensors requires the pre-deployment of software and hardware on the monitored device,and requires access permission to the sensor.It also requires certain prior knowledge of the monitored device,such as device type,sensor type,and sensor access permissions.Other experts believe that weak vibrations caused by eavesdropping on target voices can be observed through wireless signals,thus achieving remote eavesdropping.This type of attack reduces the difficulty of eavesdropping deployment and is not easy to detect,thus significantly improving practicality.However,current eavesdropping research based on wireless signals requires prior knowledge and can only target a limited vocabulary set,unable to cover the entire human speech spectrum.Millimeter-wave radar has the advantages of low cost,small size,light weight,and high resolution,and can work continuously.In recent years,they have been widely used in obstacle avoidance systems in vehicle radar.Therefore,we plan to use frequencymodulated continuous wave signals to detect weak vibrations with high precision,and combine the powerful information restoration capability of the current popular generative adversarial networks to reconstruct sound based on vibration information,in order to achieve eavesdropping on targets in private spaces.This platform can be divided into four functional modules:sound source vibration positioning module,sound source vibration extraction module,spectrum information enhancement module,and high-quality audio reconstruction module.In order to guarantee the quality of audio reconstruction and achieve the eavesdropping task without prior knowledge,we incorporate the data augmentation function of generative adversarial networks into a mmWave radar-based sound restoration system.This integrated approach can effectively improve the quality of audio reproduction while ensuring comprehensive coverage of eavesdropping targets.The close cooperation of various functional modules not only improves the accuracy of eavesdropping,but also greatly improves the efficiency of eavesdropping.