Author Topic: How viable is this idea  (Read 770 times)

0 Members and 1 Guest are viewing this topic.

Offline VeteranRobotBeginnerTopic starter

  • Beginner
  • *
  • Posts: 2
  • Helpful? 0
How viable is this idea
« on: October 12, 2012, 11:39:52 PM »
I am building a robot that needs to carry a small wooden block in a straight line to a square target area upon which it will drop the block and continue on. It needs to be small enough to fit inside of a 12x12x12 inch box.

This is the idea I have for the drive system:

Those polygonal scribble shapes are supposed to be gears.

So the design involves one motor controlling all 3 wheels. The 2 front wheels are attached by a rod that has a gear around it. The back wheel is attached to a gear that is belted together with the motor gear and wheel rod gear.

How feasible is this idea? I have no mechanical background and this is my first idea, so feel free to offer me advice on how to improve it.

If I used this as my drive system, how strong of a DC motor would I need to make this work? The payload it will be carrying is a 2x2x2 inch block of wood.

Edit: Would I be better off just using two rear wheels instead of one? Should the motor perhaps be positioned directly in line with both motors so that it is connected with the belt on both sides? Should I forget the belt and go with a differential drive (2 motors controlling each of the front wheels separately)?
« Last Edit: October 12, 2012, 11:46:04 PM by VeteranRobotBeginner »

Offline Soeren

  • Supreme Robot
  • *****
  • Posts: 4,672
  • Helpful? 227
  • Mind Reading: 0.0
Re: How viable is this idea
« Reply #1 on: October 13, 2012, 02:11:14 PM »
Hi,

How feasible is this idea? I have no mechanical background and this is my first idea, so feel free to offer me advice on how to improve it.
It will only allow it to go in a straight line and if that's what you want, at least use a timing belt and the pulleys to match - front and back wheels should rotate at the same speed of course.


Edit: Would I be better off just using two rear wheels instead of one? Should the motor perhaps be positioned directly in line with both motors so that it is connected with the belt on both sides? Should I forget the belt and go with a differential drive (2 motors controlling each of the front wheels separately)?
Two gearhead motors, differential drive and a single passive aster wheel will make it so much easier - and you can expand on it later on, or use the parts for a more advanced 'bot.

You'll probably need a servo for picking up and dumping the block.
Regards,
Søren

A rather fast and fairly heavy robot with quite large wheels needs what? A lot of power?
Please remember...
Engineering is based on numbers - not adjectives

Offline VeteranRobotBeginnerTopic starter

  • Beginner
  • *
  • Posts: 2
  • Helpful? 0
Re: How viable is this idea
« Reply #2 on: October 14, 2012, 12:15:01 PM »
I am now thinking I'll go with differential drive for the two front wheels. There will be no back wheel, it will just drag the back of the bot on a low friction castor. The two wheels will be extra large in order to make the slope of the top surface of the robot great enough for the payload to slide down it. The payload will rest on the inclined top surface with enough potential energy to slide off onto the ground. There will be a servo keeping something in the way of the payload that will move out of the way at the appropriate time to allow the payload to slide off.

The robot must drop the payload within a 6x6 inch square BLACK target area (a poster board on the floor), that is within a larger 18x18 inch square WHITE area. Both squares have 1 inch thick black outlines.

So how will it know when and where to drop the payload? One idea is to use a visual sensor. I do not know much about how these work. There are many tutorials online for obstacle avoidance, but my bot will just need to sense the target area, drive forward a small amount, and release the payload from the rear. Here is a basic idea of pseudocode for this:
Code: [Select]
     IF visual sensor senses sudden change from white to black THEN
          stop motor after 0.5 seconds (the bot drops off the rear, so it needs to move forward a bit)
          tell Servo to rotate out of the way of payload
          start motor
     END IF

The 1 inch thick black outlines of the target areas scare me though. Will this throw off my robot's sensor? Should I instead program it to ignore the first line it sees and stop at the next?

This is my first robot btw, and this is for a school project.

EDIT: Another option is to just program the robot to drive 10 feet, stop, release the payload, drive forward another 10 feet. I think I forgot to mention, but the target area is 10 feet from the start, and the finish is 10 feet after the target area.
« Last Edit: October 14, 2012, 12:18:36 PM by VeteranRobotBeginner »

Offline waltr

  • Supreme Robot
  • *****
  • Posts: 1,942
  • Helpful? 98
Re: How viable is this idea
« Reply #3 on: October 14, 2012, 05:47:53 PM »
Quote
The robot must drop the payload within a 6x6 inch square BLACK target area (a poster board on the floor), that is within a larger 18x18 inch square WHITE area. Both squares have 1 inch thick black outlines.

So how will it know when and where to drop the payload? One idea is to use a visual sensor. I do not know much about how these work. There are many tutorials online for obstacle avoidance, but my bot will just need to sense the target area

Look up (google and search these forums) for 'Line following' as you can use the exact same sensor scheme which is what a line following robot does, detects the edge of black/white areas.
The only difference is the code you write that uses the sensor info.

Offline Soeren

  • Supreme Robot
  • *****
  • Posts: 4,672
  • Helpful? 227
  • Mind Reading: 0.0
Re: How viable is this idea
« Reply #4 on: October 14, 2012, 06:33:00 PM »
Hi,

The robot must drop the payload within a 6x6 inch square BLACK target area (a poster board on the floor), that is within a larger 18x18 inch square WHITE area. Both squares have 1 inch thick black outlines.

Do you have a link to a sketch or a photo of this?

As I understand it, if driven in a straight line through the course you'll drive on:
~8½' on a (here) undefined surface color
1" of black outline
5" of white
another 1" of black outline
6" of black dumping site
yet another 1" of black outline
5" of white
last 1" of black outline
~8½' on a (here) undefined surface color

I don't understand why there has to be a border around the already black dump site, and if it is so, I'd just define it as an 8" black - can you confirm if it is indeed what it looks like?


An IR reflectance sensor would work well in combination with odometry.
With an encoder disc on each wheel, you can make it go straight and count the approximate distance to the first black border, then let the IR detect whether you're over black or white and when the block is in place, revert to odometry to get to the goal line.

If you load the block manually, it should be unloaded at the back of the 'bot, or you'll have to drive around it, which will likely take you off the straight and narrow.

Is there any markers/borders/whatewer that you can use to keep the straight line?
If not, I doubt that you can get it to run as straight as needed. Aiming for the middle of the 6", the distance of 120" means that if one wheel is marginally faster (like travelling 0.04" more than the other on the 10'), it will not hit the dump site and your 'bot will have to search for it - much more involved to make software for.


Here is a basic idea of pseudocode for this:
Code: [Select]
     IF visual sensor senses sudden change from white to black THEN
          stop motor after 0.5 seconds (the bot drops off the rear, so it needs to move forward a bit)
          tell Servo to rotate out of the way of payload
          start motor
     END IF

I'd mount the sensor in the front or somewhere under its "belly" and look for the black-to-white transition when it is on its way out of the dump site and then let it slide the block off and continue - I take it that "anywhere in the black square" is OK, as long as the block doesn't touch white?


The 1 inch thick black outlines of the target areas scare me though. Will this throw off my robot's sensor? Should I instead program it to ignore the first line it sees and stop at the next?

You know that the "wrong" black line/border is 1" and that will be easy to measure, by counting clicks from the encoders.

With optimum placement of the sensor, you can dump the block when you see the black/white transition - you may even be able to dump it without stopping and land it right in the middle.
Only one sensor means that you cannot be absolutely sure that the block will land at dead center - you decide how importans that is.

Since you are allowed up to a 12" square footprint, 3 to 4 sensors could be used to make sure you are right over the 6" square.

If you have any links to this, do post them, as it will help us help you :)
Regards,
Søren

A rather fast and fairly heavy robot with quite large wheels needs what? A lot of power?
Please remember...
Engineering is based on numbers - not adjectives

 


Get Your Ad Here

data_list