Let's Make Robots!

Rover kitten locator

Navigate around via remote control

Hello to all and thank you for looking at my ongoing project. This is my first LMR project, inspired by all the fine examples on this website. So far I have built my chasis from scratch, and modified 4 servos for continuous rotation, and use 4 servos to navigate. The tools used to build the chasis were a drill press, dremel tool, scroll saw. I would appreciate any feedback to help continue the development of the platform. I would like to add a basic arm, web cam (cell phone), navigation sensors, not sure which type to use yet. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Thanks again for looking, leave a comment.

jsh6061

 

 

4-18-2012 Update

 

Today I worked on tracking a straight line. I used resistors to convert the servos to continuous rotation. I was trying to use a multiplier to find the dead zone for the continuous servos. This caused them to be out of sync. I installed the potentiometers, and set the servowrite. to 90 and adjusted the potentiometers to get no movement. I am using an arduino uno and curently only have 6 pwm outputs. I was using 4 for the continous servos and 2 to steer. I switched to 4 for steering and 2 for continuous. I am able to align the wheels much better and currently with the continuous servos all set to the same 90 setting for neutral, the rover now tracks almost perfectly straight. I also added the top deck for future additions, ie sensors, arm, camera. 

Thanks for looking.

jsh6061

 

The potentiometer for adjusting neutral for the continuous servo, one on each wheel.

 

 

.

The upper deck, ready for the fun stuff !!!!

 

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Depending on the autonomous level you want to reach with this robot there are different ways to solve the "pulling to one side" problem.

To be able to correct, the rover must first detect the "pulling to one side". You go from an open loop to a closed loop system.

One detection-strategy could be that you measure the side pulling by using an X/Y plane. Consider the forward direction is X. If you detect an Y component in the movement, the robot needs to act inverse to the Y component to correct this unwanted movement. So an Y component must be compensated by a ( - q ) * Y component. q could be 1 but could also be greater or lower than 1.

Two tactics come into my mind to detect Y components in the movement:

  1. Use a hacked mousepointer. Let the mouse-ball touch the ground and read the X/Y data that comes from it. Correct the movement when you detect an Y component.
  2. Use a compass module. After a turn make a snapshot of the indicated compass direction. While further forward movement correct the movement by using the direction snapshot and the current direction.

Of course if you ask LMR there is several more tactics you can use to fulfill the correction strategy.

Your solution on how to turn is very neat. To use four servos here makes it possible to turn on one point - turn without travel. This is a powerful feature.

While the rotation seems very nice I ask myself about how the straight-drive goes. Can the rover drive straight without continuous steer corrections?

 ...and welcome to LMR.