Xiaohan Jin (Sharon)

 

LiDAR SLAM

• We mounted a LiDAR scanner, IMU, and encoder on a mobile robot and utilized PKU-RobotSDK to control it.

• We realized LiDAR SLAM by sensor fusion

 

 

 

 

Monocular Visual SLAM - coolest part of my work!

• We calculated monocular visual odometry.

  • Fundamental matrix is estimated by 8-point algorithm.

• We estimated the global position by Extended Kalman Filter (EKF) with input of visual odometry, encoder and IMU.

• We mounted Xtion to perceive depth.  

• Finally, we reconstruct the whole 3D scene by OpenGL.