For this quarter, I will work with Tsangkai addressing on EM-SLAM problem. First half of quarter:

  1. Investigate EM-SLAM problem in detail. Validate the result from "EM-SLAM with Visual/Inertial Application" using Python.
  2. Try to achieve convergence of this algorithm. If unable to achieve, figure out the reason why it fails and try potential methods
  3. Give a demo of robot in 2D case and analyze the performances.

Second half of quarter:

  1. Convergence with EM-SLAM is usually not guaranteed. I will focus on developing new ways to make this algorithm converge.
  2. EM-SLAM is an offline algorithm. I will try to develop an online version of EM-SLAM algorithm and give a demo

Next Post Previous Post