Go direct to the page content

Artificial Intelligence of Things Laboratory(AIoT Lab.) at POSTECH

aims to develop and expand the use of wireless networks. We are currently interested in IoT networks, applications of AI in mobile networks, next generation networks*, and human activity recognition using wireless signals.

* : Wi-Fi 7, Bluetooth 6 and 5G NR & B5G

Research Topics

Wireless Communication and Artificial Intelligence

  • read more
  • read more
  • read more
  • read more

Recent Papers

more
  • Wireless LAN
  • Medium Access Control
  • Full Duplex

FDCR: A Full-Duplex Collision Resolution Scheme for Next-Generation Wireless LANs

Hyeongwoo Jo, Hyeongtae Ahn, Eunhyun Kim, Young-Joo Suh

November 2021 Published on IEEE Communication Letters

Full-duplex capable wireless stations can detect incoming signals during transmission. Therefore, they can reduce channel time wasted in collisions by detecting a collision and stopping the ongoing transmission immediately. However, in collisions between full-duplex and legacy stations, even if the full-duplex station stops transmitting, the collision duration does not decrease since legacy stations cannot detect the collision. This letter proposes a novel medium access control scheme, called FDCR, which resolves a collision between full-duplex and legacy stations. FDCR provides collision resolution without modifying the 802.11 standard of legacy stations. When a collision occurs, the full-duplex station reports information on the colliding stations to the access point. Then transmission opportunity is granted to them. The analytical model and simulation results show that the throughput of FDCR outperforms existing studies.

read more
  • Multiple ToF Estimation
  • STO
  • OFDM

Recycling Sampling Timing Offset of Wi-Fi for Estimating Multiple ToFs of Superimposed Signal

Jaegook Lee, Jio Gim, Young Deok Park, Young-Joo Suh

November 2021 Published on IEEE Communication Letters

Many Wi-Fi based device free localization (DFL) methods have been proposed for indoor location based services. Unfortunately, the received signal is superimposed with Line-of-Sight signal and reflections so that multi target DFL is only possible by estimating the time of flight (ToF) of each signal. To estimate multiple ToFs, we utilize the sampling timing offset (STO) that inherently occurs by asynchronous sampling timing between TX-RX. By utilizing STO, we can generate signals mimicking oversampled signals. We put the signal to our correlation based ToF estimation algorithm. We achieved 0.75-4.65 ns median error when 2-5 signals are superimposed.

read more

Lab News

more

If you're interested,

feel free to contact our lab manager!

j91postech.ac