Energy-efficient miniature autonomous robots have a huge potential for enriching and saving our lives. For instance, these robots’ ability to navigate and explore autonomously in remote, crowded and dangerous environments make them very ideal for 3D mapping, monitoring, space exploration and search / rescue. However, current state-of-art miniature robots have limited capabilities because their limited on-board power does not support the amount of computation required for performing basic tasks that enable autonomous navigation and exploration, such as localization (for determining the position of the robot), 3D sensing (for mapping / object avoidance), and path planning (for environment exploration). Currently, these three tasks are very challenging to perform on conventional CPUs and GPUs that consume 10W to 100W of power, which are orders of magnitude higher than what is available on the miniature robots (~200mW). Thus, the goal of my research is to co-design both the algorithm and hardware so that all three tasks can be performed on-board the miniature robots with higher performance while consuming much less power compared with CPUs / GPUs.

Under the supervision of Prof. Vivienne Sze and Prof. Sertac Karaman, I am grateful for having the opportunity to design several energy-efficient hardware accelerators that enable autonomous navigation and exploration. One of such accelerators contains a novel memory architecture and data delivery method that support multiple energy-efficient, high-performance cores for computing Shannon mutual information (MI) at any arbitrary scan location on a map. Since the MI computation speed often limits the robot’s ability to autonomously explore an unknown environment, my proposed accelerator greatly reduces exploration time by performing two orders of magnitude faster while consuming an order of magnitude less power than a server grade CPU. Seeing the impact of this accelerator not only inspires me to work towards an energy-efficient, high-integrated chip that enables all required functionalities for autonomous navigation and exploration on miniature robots, but it also makes me more determined to become a leader in the autonomous robotics field so that I can make a positive impact on human lives.

In my free time, I enjoy pursuing several hobbies. I love music and have practiced piano for 15 years and percussion instruments for 7 years. Currently, I am a percussionist at MIT Concert Band. I also love to give back to the community. Every Sunday since my high school years, I volunteer at Advent Valleyview Retirement Centre by helping seniors with computer tasks and music-making.

Education

Awards & Honors

  • 2018
    EECS Grad Alumni Fellowship
    Awarded by MIT EECS Graduate Fellowship Committee
  • 2016
    Kenneth Carless Smith Award
    Awarded for academic achievement in 3rd year
  • 2015
    Andrew Alexander Kinghorn Scholarship
    Awarded for academic achievement in 2nd year
  • 2015
    Russell Reynolds Memorial Scholarship
    Awarded for academic achievement in 2nd year
  • 2015
    NSERC Undergraduate Student Research Award
    Awarded of U of T's ECE Department for summer research opportunity at Professor Genov's lab.
  • 2014
    Anthony A. Brait Scholarship
    Awarded for academic achievement in first year
  • 2014
    University of Toronto Scholar
    Awarded for academic achievement in first year