Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Need Advice on a Small Robotic Design

  1. Mar 16, 2012 #1
    Hi. I'm building a small automated robot that will only work with 2 motor, 2 wheels, and 2 sensors. My goal here is for this little robot to make through a track without falling off. Here's a little sketch I drew on Paint to help understand what I'm saying.


    Each sensor will be connected to a motor. So the sensor on the left connects to the right wheel/motor, so when the sensor goes out of bound, the right wheel stops and the bot turns right. The sensor are just "clickers." The track is off ground so when the sensor is pressed down, the wheels move, when the sensor goes out of bound it click itself off and the wheel stops. But this is not the important part.

    My question is at what angle and how long should the arms holding sensor should be? The width of the entire base is just about 10 cm, so it is not very big. And the track is about 3 times the with, a little more.

    Now the track consists of some 90 degree turns. And a wavy zigzag portion where it goes from left turns to right turns very quickly but the turns aren't 90 degrees, it's a smooth turn.

    My theory is that if the arms are closer together (less angle) then it will be able to detect the corners quicker and turn before hand, but once you hit the wavy zigzag turns, one of the wheel falls behind. But if the sensors are very wide (right picture) then it will turn much more efficient but it won't be very good at those 90 degree turns. And then I have no idea of whether the arms should be very long or short. Probably a medium size is good. Too long and the wheels will follow too far behind, but too close and it might just overrun a corner.

    So if anyone can help me, i'd really appreciate it.

    What is a good angle to put the sensors apart?
    How long should the arms be?
  2. jcsd
  3. Mar 16, 2012 #2
    can you draw the robot while it's on the track?
  4. Mar 18, 2012 #3
    Take in consideration that it is not my best drawing, haha. I just made a quick sketch, but I believe the proportions are close.

  5. Mar 18, 2012 #4
    so, the black part, is that wire in the ground, or a wall as tall as the robot, or is it a chalk line, or something else?
  6. Mar 18, 2012 #5
    It's off the ground, about 5 cm off the ground. Not so big so if the bot falls it breaks apart, but enough so the sensor can react. It's basically a platform.
  7. Mar 18, 2012 #6
    have the sensor be spring loaded, so it drags along the ground
    use a potentiometer to measure the amount of extension.
    when the robot starts to run off the track, the sensor will extend (since it wants to be on the ground). when it extends, the potentiometer will give a different measurement.
  8. Mar 18, 2012 #7
    Yeah, that's basically what the sensor is. The sensor, motors, and circuit is all already made, I just have that physics question of how long and at what angle should the arms be. But you are absolutely right with the sensors, those are exactly what they are.
  9. Mar 18, 2012 #8
    that depends on the specific measurements of the course
    widest turn, narrowest turn, track width max/min, etc.
    if it's too wide, it'll never progress through narrow portions of track.
    if it's too narrow, it'll detect an edge too late
    if the arms are too long, they'll fall during turns
    too short, and they'll detect turns too late

    your best bet is to experiment with different configurations, and see what works best
  10. Mar 18, 2012 #9
    The track is the same width at all points and the narrowest turn is 90 degrees. The wide turns won't give a problem.

    My estimate of how long the arms should be was at least the width of the wheels. So when hitting that 90 degree turn it will detect it before it is too late to turn. If the arm is shorter then the right wheel is bound to fall off.
  11. Mar 20, 2012 #10
    Any other advice is greatly appreciated it. Thanks jehan though.
  12. Apr 3, 2012 #11
    It should be better to have a long and small angle arm. Using the potentiometer and spring mechanism suggested by Jehan, you can detect the curve ahead. When the arm go further off the ground (out of line), you will turn the wheel harder. then you can predict the curvy road in front and respond accordingly before it hit the edge.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook