Friday, 15 April 2011

Line Follower - PID Method - Over corrected!



This post is a followup to the previous PID Line Follower post, of which I was using the PID method. As you would have noticed in the previous post, the Line follower robot was moving very smoothly over the black line, however it was extremely slow!

Please note that the line follower robot is only using 1 light sensor, the left one. The robot has been built with 2 sensors for future test algorithm.

In this followup post, you'll notice that the robot moves along at faster rate, but the robot seems to be constantly oscillating, with a few occasion that it manage to reach equilibrium. This is because the robot seems to be over compensating. The reason that this could happen is because is because the robot's feedback loop is too slow to react, and is NOT able to compensate quick enough, causing it to go into an oscillation mode. I would have to re-code the robot using NXC or RobotC to overcome the speed issue.

Another possibility is because of the variables that I have used to adjust the robot for a faster speed, rather than a smooth moving robot. Maybe the setup variables is inaccurate, thus the robot is unable to hit its equilibrium. However in the video above, the robot did reach an equilibrium state (moving straight) occasionally.

I will post the source codes in the future post after I have done investigation on how to make the robot move faster yet smoother.

If anyone has manage to complete the above task, please post in the comments how you manage to make the robot move fast yet smooth with 1 light sensor.

No comments: