ROBOTS DREAMS - Robot news and robot hacks from a true robot geek

« MINIROBOTICA - National Robotics Contest of Mexico - Report & Video | Main | Who Needs Animal Crackers? »

Building One Of The ROBO-ONEs: A Handy Shortcut

We installed the larger feet on Gulliver, our KHR-1 robot, and added some make-shift soles to give him more traction and less slip. Of course, changes like that require adjusting all the motion sequences since we’ve effectively changed the robot’s basic geometry. In the process of adjusting the motions we came up with a very useful method to test the results.

 

Motions usually contain repetitive sequences. For example, to walk forward you might create a motion like this:

1. Home position
2. Natural position
3. Shift weight to the right leg
4. Lift the left leg
5. Move the left leg forward
6. Place the left leg and shift the weight onto it
7. Lift the right leg
8. Move the right leg forward
9. Place the right leg and shift the weight onto it
10. Repeat steps 4 through 9 as many times as you need
11. . . .
12. . . .
13. Natural position
14. Home position

At first we tried creating the whole motion sequence with all of it’s steps using the motion editor in the HeartToHeart application. That worked well enough, but we found that we frequently wanted to make small adjustments to a servo position. Keeping all the data in one motion file created a situation where we had to repeat the same changes in several places for each adjustment we made.

 We happen to be using the HeartToHeart Ver1.3 software, but
 the techniques used here should work with other versions.

To get around that problem we broke the motion sequence down into separate parts. For example, we created a motion labeled “HOME” whose sole contents were the servo positions for HOME. We stored that motion as M23.

A second motion, named “NATURAL” was created with the servo positions for a semi-crouched posture, and was stored as M24.

Then we stored the repetitive walk moves in a motion sequenced labeled “WALKFWD”.

Then we used the scenario editor to create the total sequence we wanted to test.

For example, in this particular sequence we had the robot start in the home position, move to the natural position, take two complete steps forward, then end in the natural position.

We could have extended the scenario to include multiple repetitions of the forward step– say an additional five or ten steps, or added other motions.

The advantage of this method is that it allows us to quickly make changes to the basic motions, yet test the changes through multiple repetitions. The scenario data just references each set of motions. If we make a change to a pose in the WalkFWD motion the change is automatically reflected in the corresponding scenario.

 

Posted by Lem on May 20, 2006 | Permalink | Email this post

---

TrackBack:

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d83451da2069e200d8352cda7a53ef

Note: To avoid spam and malicious content, comments and trackbacks are screened by a carbon lifeform prior to appearing on this weblog.

---

Listed below are links to weblogs that reference Building One Of The ROBO-ONEs: A Handy Shortcut:

---

Comments:

---

Post a comment:

Note: To avoid spam and malicious content, comments and trackbacks are screened by a carbon lifeform prior to appearing on this weblog.

Name:

Email Address:

URL:

Remember personal info?

Comments:

Advertise on this site