Research On Wireless Signal Based Quantitative Sensing Technologies For Industrial Environment

As the bridge connecting the physical world and cyberspace,Internet of Things(Io T)sensing technology has always been a hot research topic.It can be classified into two cat-egories,the traditional sensor-based sensing and the emerging sensorless sensing.Com-pared with the former,the latter adopts more pervasive and ubiquitous wireless signals,e.g.audio,radio and optical signals,to perceive the complicated physical world.The radio frequency(RF)signal based wireless sensing technology studied in this dissertation refers to the approach that captures the effects of the changes in the target's physical state on the RF signal's properties.In recent years,wireless sensing technology has received extensive attentions due to its advantages like non-contact measurement,wide monitoring range,and easy deployment and maintenance.Among all potential application scenarios,wireless sensing applications for indus-trial environments are undoubtedly one of the most valuable directions for exploration.First,there are indeed a large number and variety of real monitoring demands in the in-dustrial environment.Second,the empowerment of smart sensing in the industrial envi-ronment needs to minimize its impact on existing production lines and devices.Wireless sensing can satisfy the sensing demands while not requiring the transformation and up-grading of existing production lines and devices,and not interfering with their operating conditions.Nevertheless,the existing wireless sensing technologies either only can per-form qualitative detections or can hardly meet the high-accuracy requirements of indus-trial monitoring.Therefore,how to improve quantitative sensing capabilities of wireless sensing technologies in industrial monitoring applications needs to be explored urgently.In response to this crucial issue,this dissertation aims to improve the timeliness of quantitative sensing,reduce the measurement error of quantitative sensing and increase the measurement dimension of quantitative sensing.Two types of RF signals,radio fre-quency identification(RFID)and millimeter wave(mm Wave),are studied.The main research contents include the following three aspects:(i)Quantitative sensing of the liq-uid leakage based on the mutual coupling effect of multiple battery-free tags: we leverage the fact that the asymmetric near-field mutual coupling effect of two adjacent RFID tags is only sensitive to the liquid leakage within a limited range near these tags,build a multi-source differential signal variation pattern to improve the efficiency and effectiveness?(ii)Quantitative sensing of the mechanical vibration based on the spatiotemporal geometric features of mm Wave reflection signals: we exploit the relationship between the spatiotem-poral geometric features of the complex mm Wave signals and the tiny-amplitude vibra-tions,construct a quantitative model for the error analysis,and propose 8)-level vibration measurement error control approaches?(iii)Quantitative sensing of the two-dimensional(2D)rotor orbit based on the exploitation of mm Wave ghost multipath reflections: we explore mm Wave's ability to recognize and separate the multipath reflections,and use the additional observations carried in them to quantitatively measure the sub-8)8)-level two-dimensional rotor orbit of rotating machinery.The evaluation results show that the approaches proposed in this dissertation can respectively(i)improves the timeliness of the liquid leakage monitoring and has a high detection accuracy of 97.2% and around 10% relative quantitative sensing error of the leakage parameters?(ii)reduces the measurement errors of 8)-level mechanical vibra-tions by achieving less than 5% and less than 0.03% relative quantitative sensing errors of the vibration amplitude and frequency,respectively?(iii)extends the sensing dimension of micro-displacements by achieving less than 10 um absolute quantitative sensing error of the 2D rotor orbit measurement? The last part of this dissertation summarizes the core research ideas and discusses future research directions.