Localization and Navigation for a Differential Drive Robot


Autonomous Robot Navigation is a technical elective course. It covers robot platforms & modeling, control structures, sensing & estimation, localization, and motion planning. Most labs in this course required implementing algorithms on Jaguar Lite Robots. The course concludes with a final navigation competition for which the team implemented and evaluated probabilistic motion planning algorithms.

The labs were done with Kunal Menda and the final competition team also includes Aishvarya Korde and Mo Zhao. See the Github repository here.

The labs in the course are as follows:

  • Odometry Localization: Modeled differential drive robot and error from state-estimation through odometry.
  • Point Tracking: Developed PID motor controller and state-space point tracking controller.
  • Particle Filter Localization: Developed and tuned a particle filter to localize the robot in the known start location, wake-up robot problem, and kidnapped robot problem.
  • Motion Planning: Develop an implementation of a Probabilistic Road Map to navigate the robot through a complex physical environment.

Here is a selected lab report:

Back to Projects