Author Topic: Exoskeleton design  (Read 2183 times)

0 Members and 1 Guest are viewing this topic.

Offline PomacsTopic starter

  • Beginner
  • *
  • Posts: 5
  • Helpful? 0
Exoskeleton design
« on: July 24, 2015, 03:56:53 PM »
Hello buddies  :)

I am designing a lower extremity exoskeleton. For simulation, I found that I need a walking human in a simulation environment to attach the exoskeleton to it.
I tried OpenSim but wasn't satisfactory, and finally I will use MSC Adams for dynamic simulation.
It's actually a long time from I began to make a lower body of a human walk stably. I am facing a lot of problems regarding, the body segment lengths and the stable waking gait that fits it and the type of contact between the feet and the ground.

Can anyone help me with more guidance?

Offline mklrobo

  • Supreme Robot
  • *****
  • Posts: 558
  • Helpful? 15
  • From Dream to Design at the speed of Imagination!
Re: Exoskeleton design
« Reply #1 on: July 27, 2015, 03:38:04 PM »
 ;) Hello!
I have not played with the software your are describing. I just draw a free body diagram, and go experiment
from there, one the intranet is exhausted. There are other posts that may help you. I am planning a
evolution of walking members, so I can think over time, which one is the best. What parameters are you
facing to work with?

Offline cyberjeff

  • Full Member
  • ***
  • Posts: 114
  • Helpful? 7
Re: Exoskeleton design
« Reply #2 on: July 29, 2015, 07:13:22 AM »
Humans make ground contact on the balls of their feet. They belong to the plantigrads:

You happen to have a working model that you can take for a test drive whenever you wish.

Walking is not stable it requires momentum and dynamic balance. Balance can be done shifting arms, hips, torso or applying forces on the foot in the proper direction. The corrections are minute but necessary.

Humans are capable of an enormous number of gaits, all requiring micro corrections.

I've been working on a 4 legged digitigrade. I'm just  in software at the moment waiting for more servos. 4 legs is easier than two as the corrections can be done by one of the feet, but it  is still involved.

I'm curious as to how you program the gaits, how do you do  that?

I've broken this into limbs, and the limbs are broken into the significant joints with the rear  being different than the front. Unlike the way animals process their muscles my robot must work on the servos setting angles rather than forces. There will be a feedback process from an accelerometer/ compass and pressure sensors on the feet which I will write later. The gait is broken into forward and back strokes with a cubic bezier describing the motion of the servos for each servo. A cubic bezier requires 4 points to be set, so, for one step, with each limb having 3 servos that is a total of 96 points for all 4 limbs. Fortunately this is relatively easy to get a starting configuration and the left mirrors the right with a time offset.  How others do this, I have no clue.