Assembly-Oriented Design: Difference between revisions

From Open Source Ecology
Jump to navigation Jump to search
No edit summary
 
(16 intermediate revisions by the same user not shown)
Line 1: Line 1:
Robossembler:
Robossembler - https://gitlab.com/robosphere


<html><iframe width="560" height="315" src="https://www.youtube.com/embed/AFROcGW73j0" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe></html>
<html><iframe width="560" height="315" src="https://www.youtube.com/embed/AFROcGW73j0" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe></html>


*This speaks OSE language: modular, designed for automated assembly. That is what OSE does, except we are working on a slightly different angle: producing for design-for-assembly, as we are interested in large complex devices (houses, tractors, etc). Our second step will be designing for gradual automation of build. In Robossembler, the approach is to design an automated assembler, then develop applications. At OSE, we design applications, then design the automated assembler.
=OSE Assessment=
 
*This speaks OSE language: modular, designed for automated assembly. That is what OSE does, except we are working on a slightly different angle: designing products for design-for-assembly, as we are interested in large complex devices (houses, tractors, etc). Our second step will be designing for gradual automation of build. In Robossembler, the approach is to design an automated assembler, then develop applications. At OSE, we design applications, then design the automated assembler.
*Absolutely inspiring: builds upon [[ROS2]], [[Gazebo]], 3D printed parts, FreeCAD modules for robotics including gripping geometry!
*For example, [[Seed Home v2]] starts as [1] a complete digital model. We then design the [2] 3D printer that prints the parts, based on an early prototype of [[D3D Giga]] 3D printer. At a later stage, [3] we design the automated assembler - such as a scaled version of Robossembler.
*We are at stage [1] now as a means to generate cash flow, near term goal of [[$1.2M Net Revenue Model]]
 
=Other Notes=
*'Skills market' for trained models.
*Meta-Reinforcement Learning - robot learns to learn
*Machine learning for robotics as such is recent work, 2021 - meta actor-critic with advantage weighting (MACAW) [https://xbpeng.github.io/projects/MACAW/2021_ICML_MACAW.pdf]
 
=Files=
 
[[File:griptool.png|200px]]
 
from https://gitlab.com/robosphere/arm-tools/grip-tool/-/blob/main/grip-tool.STEP
 
==Is motor design open source?==
 
[[File:openstepper.png|600px]]
 
=Links=
*Gitlab docs - [https://robosphere.gitlab.io/robossembler-docs/docs/]
*Robossembler Telegram - [https://t.me/robossembler_ru]

Latest revision as of 18:05, 19 May 2022

Robossembler - https://gitlab.com/robosphere

OSE Assessment

  • This speaks OSE language: modular, designed for automated assembly. That is what OSE does, except we are working on a slightly different angle: designing products for design-for-assembly, as we are interested in large complex devices (houses, tractors, etc). Our second step will be designing for gradual automation of build. In Robossembler, the approach is to design an automated assembler, then develop applications. At OSE, we design applications, then design the automated assembler.
  • Absolutely inspiring: builds upon ROS2, Gazebo, 3D printed parts, FreeCAD modules for robotics including gripping geometry!
  • For example, Seed Home v2 starts as [1] a complete digital model. We then design the [2] 3D printer that prints the parts, based on an early prototype of D3D Giga 3D printer. At a later stage, [3] we design the automated assembler - such as a scaled version of Robossembler.
  • We are at stage [1] now as a means to generate cash flow, near term goal of $1.2M Net Revenue Model

Other Notes

  • 'Skills market' for trained models.
  • Meta-Reinforcement Learning - robot learns to learn
  • Machine learning for robotics as such is recent work, 2021 - meta actor-critic with advantage weighting (MACAW) [1]

Files

Griptool.png

from https://gitlab.com/robosphere/arm-tools/grip-tool/-/blob/main/grip-tool.STEP

Is motor design open source?

Openstepper.png

Links

  • Gitlab docs - [2]
  • Robossembler Telegram - [3]