felipenmartins / robotics-simulation-labs Goto Github PK

Hi,

Some great tutorials you've done here!

I've been using lab 3 (localisation / odometry readings) for a project of my own but am struggling to get my head around something.

Using the accurate values for the robot, I would have thought that the pose info would be rather accurate, however I'm finding that I get a compounding error and after a 50 laps of the track, my values are off significantly:

• phi: -1.1137 vs. 0.0323
• x: -0.75 vs. 0.0172
• y: 0.03 vs. 0.439

I was wondering if you had any insight on what the cause of this might be? I get this result even if I just copy your provided solutions, and have tried with your values for the wheel radius / wheel distances and those provided on the webots site, and both are wildly off what the simulation suggests is correct.

I am wondering if the ePuck proto has some manufactured margin of error to simulate a real-world scenario?