Kyuin Lee

RESEARCH EXPERIENCE

• Present

  • ​Wisconsin Embedded Systems and Computing Lab  UW-Madison  ​
    Graduate Research Assistant, Prof. Younghyun Kim  [Website]

    • Investigating and developing series of secure and usable device authentication (or pairing) methods leveraging various environmental contexts to prove coexistence of mobile and IoT devices.
    • Proposed SyncVibe, fast and convenient device pairing protocol to transmit and receive pairing information utilizing vibration motor and accelerometer.
      • ​Designed and implemented clock recovery technique to maximize data transmission accuracy and throughput under timing jitter of Android OS.
      • Developed automated testing scripts to capture and analyze vibration waveform representing various bit patterns.
      • Implemented the prototype using commercial off-the-shelf smartphone and microcontroller (MCU) to evaluate it under various environments and transmission media.
    • Proposed ivPair, usable in-vehicle device pairing protocol, to derive secure pairing pin using simultaneously measured vibration within the vehicle.
      • ​Designed and implemented signal alignment technique to solve sampling frequency mismatch between commercial mobile devices.
      • Implemented pin extraction method to extract identical pins from varying locations within the vehicle.
      • Evaluated the prototype built with accelerometer and MCU on different vehicle and road types.
    • ​Proposed IoT device authentication method named VoltKey, which leverages spatiotemporal randomness in the 120 V power line to authenticate devices connected to identical power line.
      • ​Designed and implemented key extraction and sampling rate estimation algorithm to extract identical random bits from predictable power line waveform.
      • Implemented custom  hardware prototype capable of noise measurement, key extraction, and supplying power to existing IoT devices.
      • Evaluated the prototype on real-world environment within home, office, and lab under different attack scenarios.
    • ​Proposed mobile and IoT device authentication method named AeroKey, which uses randomness in the ambient electromagnetic radiation (EMR) to authenticate proximate devices.
      • ​Designed and implemented key extraction algorithm to extract identical keys from noisy EMR measurements using commercial MCU. 
      • Evaluated the performance of the prototype within home, and lab environment under varying attack scenarios.
    • ​Proposed pedestrian mobility verification mechanism named PEDRO, where only moving pedestrians can be admitted to the vehicular ad hoc network.
      • ​Implemented simulation framework to derive optimal round-trip time of wireless signal that results in robust security against different attack scenarios.
      • Evaluated the performance of PEDRO in real-world road settings to verify the simulation results.
    • ​Provided solutions to address current limitations of zero-interaction authentication works such as proximity control and predictability of the generated keys.

• 2016 - 2017​

  • Wireless Sensing and Embedded Systems Lab  CMU 
    ​Masters Researcher, Prof. Anthony Rowe  [Website]

    • Designed and implemented battery-operated IoT hardware prototype capable of transmitting object distance measurement through LoRa wide area network.

    • Interfaced laser ranging breakout board with ARM Cortex-M3 processor through custom developed printed circuit board.

    • Implemented energy efficient firmware on TI-RTOS to periodically measure distance between any outdoor objects.

    • Utilized MQTT and protocol buffers to publish messages on LoRa network.

  • ​System Level Design Group  CMU  
    ​Masters Researcher, Prof. Radu Marculescu  [Website]

    • Connected Raspberry Pi to LoRa network for large-scale mosquito sensing across campus.

    • Implemented energy efficient k-nearest neighbors algorithm in C to classify mosquito species using measured wingbeat sound.

    • Implemented 3D simulation framework in Python to visualize and analyze mosquito population distribution on campus.

    • Connected millions of Twitter posts to Mongo database and deployed interactive web application for scalable social network analysis.

• 2015​​​

  • Computational Biology Lab  CMU 
    Undergraduate Researcher​, Prof. Natasa Miskov-Zivanov

    • ​Implemented an algorithm in C to compute probability of cancer cell signaling.

    • Built interactive cell signal simulating web application in Python based Django framework.