2 line trackers vs. 3

I have always used 2 line trackers to follow a line, but I always see people using 3. What advantages/disadvantages do 3 line sensors provide?

3 sensors allow you to more accurately correct your movement. All 3 should see the line at all times. If the left side loses the line, then you need to drive a little right to correct and vice versa for the other side

How can all three see the same line at the time? There is a fair amount of space between each sensor, and the lines are pretty thin.

I dont understand how this would work. I normally just use 2, where is the left one hits the line, the left wheels stop, if the right one hits the line, the right wheels stop. Its simple and works very easily.

I do not think this is right, at least, not the way that VEX currently has you set up the sensors (with the spacing of the holes and such). I believe the idea is that the middle sensor sees the line, while the sensors to the left/right are there to let you know when you’ve gone too far to the left/right.

In 99% of cases, I can think of no clear advantage of three sensors over two, that is assuming that the two sensors straddle the line. However, this leaves some possible scenarios where your route mine consider it’s still following a line, but instead you’ve actually drifted out into the open (when you get hit or something).

  • Sunny G.

The language on this page might be a bit confusing, but this is the best explanation/guide I can find for how certain numbers of line sensors work:
http://www.wrighthobbies.net/guides/linefollower.htm

For most of our robots we have three light sensors mounted, but often just use two of them. For the last couple of games, line following has been much less important than line detecting (which is much easier), but that might not be the case for Sack Attack (just like it wasn’t for Elevation).

With the Vex Line Following Course, ( provided in a Foldout ) when On Course, the Middle Sensor should see the Line and the Left and Right Should Not…

So as far as I see it there are two types of line followers. I like to call them analog and digital. Digital line followers either see the line our they don’t nothing in between, digital followers usually have at least 2 sensors. Analog followers can tell how far it us from being centered on the edge if the line, usually these sensors are controlled by a PID.

The analog followers are usually pretty good because they have a better resolution than the digital ones but most have the draw back of having difficulty dealing with changes in direction. Generally for vex this doesn’t matter because the tapped lines don’t curve.

But this got me thinking for a non-vex competition with a more variable line if there is a solution that is the best of both worlds. I was thinking that the follower would be a combination of a digital and analog, where one sensor is following the edge of a line with a PID controller while another is diagonally in front so that if the first sensor is in the edge of the line and the second is on the center of the line. If the second sensor senses that it has left the line the program will tune the PID to account for the coming turn. What do you guys think of this? Can you think of anything better?

Third type of sensor is where you oscillate about the edge of the line. When you see the line, go off the line. When you don’t see the line, try to get onto the line. All of that while going forward.

I’ve seen teams use the “third type” with great success.

  • Sunny G.

Will this is a kind of digital follower, it only sees two states, on the line or off the line. My experience with these kinds of followers is that they are either too slow or they loose the line easily