I Am Basically Very Very Unknown About Line Tracking Robots Plllllllllllllllllllllllllllllllllllllllllllllllll llllllllllllllllllllllllllllllllllllllllllllllllll llllzzzzzzzzzzzzzzzz Give Ideas To Complete My R&d Project...................

My Title Is: Line Tracking Robots Using Pid Controllers....

 

This is the first result on Google for "pid line follower:"

http://www.seattlerobotics.org/encod...cle2/index.php

It has a lot of useful information. Good luck.

__________________
-Ian

 

heres all you need to know about line following robots, well the fundamental stuff like PID, shaft encodes, photo - transistor line follower circuits.

its quite long but ther're first year notes so not too tricky.
hope it helps - good luck!!

Attached Files

Practical Robotics - A first course ... all.zip (1,015.8 KB, 815 views)

 

hi,
i am building a line tracing robot to trace white line on dark surface. i have used two arrays of ultra bright blue LED and IR sensors for sensing white line. one array is in front and one at rear.
what method can i use to trace the line using the outputs of both these arrays. tracing on one array is very simple. i think readings of 2 arrays will give better resolution. please help.
i tried adding, subtracting the 2 output but the result is non-linear and hence cant use a PID control logic on it. please suggest what method shud i use to obtain a near-linear result using both the arrays.

 

Does anybody here know how to implement a PID controller(not for line tracer but for Motor control) in PIC C18. I did not get lucky when I googled it.

 

Thread hi-jacking side...

PID controller - Wikipedia, the free encyclopedia

You literally implment that equation in your PIC. The main difference is you have to have a integration and derivative that are constantly being calculated for every sample (and making use of past data to do so). THis is called numerical analysis- calculating things in real time. It's always an approximation because it's "real" and not theoretical. More complicated methods get a better approximation.

So for the simplest kind of derivative, you only need the current sample and the previous sample. You subtract the two and divide by the time interval between them. This gives you the slope of the line between the previous and current samples which is a derivative. More complicated methods that form better results might require use of the previous N samples in addition to the current sample.
Numerical differentiation - Wikipedia, the free encyclopedia

Integration is done in a similar way. Integration gets the area under the line so the simplest (and least accurate) form would be to multiply the time interval by the difference between the previous and current sample.
Numerical integration - Wikipedia, the free encyclopedia

 

Thankyou. I love the prompt repies iam getting here.

 

Does anybody here know how to do he tuning of PID controllers ie to find the value of Kp Ki etc.