Sensing Material Properties using Radio Frequency
Event details
| Date | 27.02.2024 |
| Hour | 13:00 › 15:00 |
| Speaker | Hailan Zhang Shanbhag |
| Location | |
| Category | Conferences - Seminars |
EDIC candidacy exam
Exam president: Prof. Andreas Burg
Thesis advisor: Prof. Haitham Al Hassanieh
Co-examiner: Prof. Yanina Shkel
Abstract
Sensing the world around us has become increasingly important in the age of automation and smart devices. Non-invasive contactless material sensing is a key primitive that can enhance a plethora of applications like robotic grasping and mapping, liquid and food quality monitoring, soil and plant sensing, quality control in warehouses, simpler security scanning as well as structural health monitoring of buildings, bridges, planes, trains, etc. Unfortunately, today material sensing requires bulky and specialized equipment like optical spectroscopy, X-Ray, and ultrasonography which typically cost tens of thousands of dollars. Past work in material sensing using radio frequency studied extremely limited scenarios using specialized equipment such as using ground penetrating radars (GPR) for measuring permittivity of materials under the soil, and thus is not translational for a more ubiquitous solution for material sensing.
As a result, there has been significant interest, in the wireless and mobile research community, to work towards enabling a scalable and practical solution. In particular, we will study recent work that leverages wireless signals from IoT radios to capture material properties which can enable a cheap, non-invasive, and ubiquitous alternative that can be used for everyday applications. Some of the techniques discussed require physically touching the material or attaching an RFID tag to the object a priori, which is sometimes not feasible if the object being sensed is unreachable or does not have a rigid surface. Other techniques we will discuss analyze the wireless signal properties after it has penetrated through the material to extract the material's permittivity. However, such techniques have only been demonstrated for liquids, require careful placement of the liquid container between two radios, and depend on the container the liquid is in. Finally another realm of wireless research looks at the reflection profiles from fruit to determine ripeness properties, however this requires THz frequencies which can be expensive. Thus, we ask the question: How can we push the performance of wireless sensing towards a contactless, cheap, and more generalizable solution?
Background papers
Food and Liquid Sensing in Practical Environments using RFIDs (https://www.usenix.org/system/files/nsdi20-paper-ha.pdf)
LiquID: A Wireless Liquid Identifier (https://dl.acm.org/doi/pdf/10.1145/3210240.3210345)
Permittivity Measurements of Multilayered Media with Monostatic Pulse Radar (https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=563284)
Exam president: Prof. Andreas Burg
Thesis advisor: Prof. Haitham Al Hassanieh
Co-examiner: Prof. Yanina Shkel
Abstract
Sensing the world around us has become increasingly important in the age of automation and smart devices. Non-invasive contactless material sensing is a key primitive that can enhance a plethora of applications like robotic grasping and mapping, liquid and food quality monitoring, soil and plant sensing, quality control in warehouses, simpler security scanning as well as structural health monitoring of buildings, bridges, planes, trains, etc. Unfortunately, today material sensing requires bulky and specialized equipment like optical spectroscopy, X-Ray, and ultrasonography which typically cost tens of thousands of dollars. Past work in material sensing using radio frequency studied extremely limited scenarios using specialized equipment such as using ground penetrating radars (GPR) for measuring permittivity of materials under the soil, and thus is not translational for a more ubiquitous solution for material sensing.
As a result, there has been significant interest, in the wireless and mobile research community, to work towards enabling a scalable and practical solution. In particular, we will study recent work that leverages wireless signals from IoT radios to capture material properties which can enable a cheap, non-invasive, and ubiquitous alternative that can be used for everyday applications. Some of the techniques discussed require physically touching the material or attaching an RFID tag to the object a priori, which is sometimes not feasible if the object being sensed is unreachable or does not have a rigid surface. Other techniques we will discuss analyze the wireless signal properties after it has penetrated through the material to extract the material's permittivity. However, such techniques have only been demonstrated for liquids, require careful placement of the liquid container between two radios, and depend on the container the liquid is in. Finally another realm of wireless research looks at the reflection profiles from fruit to determine ripeness properties, however this requires THz frequencies which can be expensive. Thus, we ask the question: How can we push the performance of wireless sensing towards a contactless, cheap, and more generalizable solution?
Background papers
Food and Liquid Sensing in Practical Environments using RFIDs (https://www.usenix.org/system/files/nsdi20-paper-ha.pdf)
LiquID: A Wireless Liquid Identifier (https://dl.acm.org/doi/pdf/10.1145/3210240.3210345)
Permittivity Measurements of Multilayered Media with Monostatic Pulse Radar (https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=563284)
Practical information
- General public
- Free