T & R Visual Navigation

Teach & Repeat Visual Navigation using a Jaguar 4 wheels Robot

Background

During the short internship with Dr Li Sun from 2019.12-2020.01, and under the help of two PhD students from the Czech Technical University in Prague, I deployed the following Teach & Repeat algorithms on a Jaguar robot 2x2 wheels(differential base), with a ZED stereo camera mounted on it. All codes can be accessed from github

Left: The Jaguar robot's outlook Right: The Testing place is near a church

The final video demo is attached below, you can click the picture to watch it.

Teach & Repeat visual navigation

System

Hardware

  1. a ZED camera(1st/2nd/mini series are all nice)
  2. a PS4 joystick (for the first teach round, not used in repeat round)
  3. a Ackerman chassis or robots with differential base (control commands will be different)
  4. a Jetson Xavier/Orin Developer Kit 32G/Laptop. (Computing Unit)
  5. a ac8265 wifi card is necessary (if you use Jetson Dev kit)
  6. a USB Hub
  7. a USB to TTL serial port Line

Software

  1. PS4 ROS driver
  2. Ackerman chassis ROS driver/Robot Velocity Control ROS driver
  3. ZED camera ROS driver
  4. ROS1-melodic

Framework

Relevant Links

Day-to-night navigation using a single experience map

Visual Teach & Repeat 3

References

[1] Krajník, T., Majer, F., Halodová, L. and Vintr, T., 2018, October. Navigation without localisation: reliable teach and repeat based on the convergence theorem. In 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (pp. 1657-1664). IEEE.