2022 Fiscal Year Annual Research Report
Experimental and theoretical study on physical layer authentication for IoT systems
| Project/Area Number |
20K19801
|
|---|
| Research Institution | Toyo University |
|---|
Principal Investigator |
朱 金暁 東洋大学, 情報連携学部, 助教 (30754329)
|
|---|
| Project Period (FY) |
2020-04-01 – 2023-03-31
|
| Keywords | authentication / radiometric feature / device identification / wireless communication / physical layer security |
|---|
| Outline of Annual Research Achievements |
Physical layer authentication, which identifies a device by exploring some "born" physical features of radio waves transmitted by wireless devices, is a promising technology to solve authentication issue in IoT networks. In this project, we planed to study physical layer authentication schemes from both experimental and theoretical aspects. In the AY 2022, we continued to explore identification schemes to improve the performance in authenticating wireless devices. Firstly, we explored the performance supremum of CFO (carrier frequency offset) based physical layer identification. Here, the supremum means the least upper bound and CFO is a widely used physical layer feature. Such study can provide a new sight for evaluating the quality of features, and thus improve the performance of the identification scheme. Secondly, unlike the previous methods which extract features manually by carefully designed feature extracting algorithms, we applied deep-learning algorithms on raw radiometric data to extract hidden features automatically for the device identification. The performance of our scheme was validated by experiments using 50 off-the-shelf Wi-Fi devices, and our experiment results showed that the deep-learning algorithm based identification schemes can achieve comparable or superior identification performances to the state-of-the-art schemes in the entire SNR (signal-to-noise ratio) regime.
|
