Legacy Forum: Preserving Nearly 20 Years of Community History - A Time Capsule of Discussions, Memories, and Shared Experiences.

Too late for a Bot-Trot 4 Bottle Race?

Discussions regarding building a walking robot at home. Most of the robots participating at Robo-One competitions are custom fabricated.
6 postsPage 1 of 1
6 postsPage 1 of 1

Too late for a Bot-Trot 4 Bottle Race?

Post by hivemind » Thu May 29, 2008 12:30 pm

Post by hivemind
Thu May 29, 2008 12:30 pm

We've got Zog Doing the Bot - Trot with flying colours :)

phpBB [media]


-Chris
We've got Zog Doing the Bot - Trot with flying colours :)

phpBB [media]


-Chris
hivemind
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 211
Joined: Sat Jul 01, 2006 1:00 am
Location: between my computer and robot.

Post by NovaOne » Sun Jun 01, 2008 6:07 am

Post by NovaOne
Sun Jun 01, 2008 6:07 am

Good control, at quite a respectable pace.

Could you tell us more about how you integrate such smooth continuous turning into your control loop?
Good control, at quite a respectable pace.

Could you tell us more about how you integrate such smooth continuous turning into your control loop?
NovaOne
Savvy Roboteer
Savvy Roboteer
Posts: 405
Joined: Thu Jul 05, 2007 7:30 am

Post by hivemind » Sun Jun 01, 2008 8:00 am

Post by hivemind
Sun Jun 01, 2008 8:00 am

Thanks! We are pleased with their performance so far. In fact, we have gotten the bots both running on many different surfaces (melomine as shown here, wood, tiles, thin rugs, cement - with sand/debris) and retain this rate of stability.

We have a standard forward walking loop, and then we take analog input from a PS3 controller (one of the joysticks, or the IMU - both have been tested, but the stick was found to be more ergonomic) to "mix" in the turning. We use the mix to elongate the steps and shift the weight - thus turning the bots while they continue to walk forward. This works for both a forwards and backwards walk.
Thanks! We are pleased with their performance so far. In fact, we have gotten the bots both running on many different surfaces (melomine as shown here, wood, tiles, thin rugs, cement - with sand/debris) and retain this rate of stability.

We have a standard forward walking loop, and then we take analog input from a PS3 controller (one of the joysticks, or the IMU - both have been tested, but the stick was found to be more ergonomic) to "mix" in the turning. We use the mix to elongate the steps and shift the weight - thus turning the bots while they continue to walk forward. This works for both a forwards and backwards walk.
hivemind
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 211
Joined: Sat Jul 01, 2006 1:00 am
Location: between my computer and robot.

Post by hivemind » Mon Jun 02, 2008 1:13 am

Post by hivemind
Mon Jun 02, 2008 1:13 am

With Zog running circles around bottles, I thought I would let Oro take a shot at an incline plane:

phpBB [media]


He does pretty well, but is required to side-step in order to regain balance near the top.

-Chris
With Zog running circles around bottles, I thought I would let Oro take a shot at an incline plane:

phpBB [media]


He does pretty well, but is required to side-step in order to regain balance near the top.

-Chris
hivemind
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 211
Joined: Sat Jul 01, 2006 1:00 am
Location: between my computer and robot.

Post by StuartL » Mon Jun 02, 2008 7:38 am

Post by StuartL
Mon Jun 02, 2008 7:38 am

That's very cool. Presumably that's IMU driven? Are you doing full inverse kinematics to get that far or just using IMU to modify preset steps?
That's very cool. Presumably that's IMU driven? Are you doing full inverse kinematics to get that far or just using IMU to modify preset steps?
StuartL
Savvy Roboteer
Savvy Roboteer
Posts: 350
Joined: Mon Jun 04, 2007 3:46 pm
Location: Thatcham, Berkshire, UK

Post by hivemind » Wed Jun 04, 2008 12:23 am

Post by hivemind
Wed Jun 04, 2008 12:23 am

We are developing a full IK for our robot, but we found that for a lot of things this simply isn't necessary. For this specific move we generated point-to-point moves and enhanced them with dynamic mixing. And while our robot does have a 6-axis IMU, it was not utilized in this move-set, at least for this slope, it was entirely unnecessary - once again, we just used dynamic mixing to compensate.

-Chris
We are developing a full IK for our robot, but we found that for a lot of things this simply isn't necessary. For this specific move we generated point-to-point moves and enhanced them with dynamic mixing. And while our robot does have a 6-axis IMU, it was not utilized in this move-set, at least for this slope, it was entirely unnecessary - once again, we just used dynamic mixing to compensate.

-Chris
hivemind
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 211
Joined: Sat Jul 01, 2006 1:00 am
Location: between my computer and robot.


6 postsPage 1 of 1
6 postsPage 1 of 1