Would be interesting to see a group of self-sustaining robots IRL

[project-turrim]

First step could be a simulation.

• http://gazebosim.org - Physics and robotics simulator
• https://github.com/nasa/astrobee - Small robot with a hand and no wheels in ISS
• "robotics hardware blueprint files"
• "robotics awesome github" -> https://github.com/kiloreux/awesome-robotics -> Gazebo
• "robot "metal" parts in gazebo" -> machines-07-00042.pdf :: Talks about practical usage of Gazebo in educational environment. Mentions "In general, the kinematic modelling of a rover depends on the physical characteristics of the robot and its components [13–15]"

[project-turrim]

.urdf files are modeling the physical parts of the robots

[project-turrim]

https://github.com/FlaSpaceInst/EZ-RASSOR - A simulation of a robot mining regolith. We could have it bring the regolith to a market in the simulation to sell the regolith for a currency or for energy. It would also be nice to have robots that can build other robots!

[project-turrim]

It would be pretty difficult for a robot to build another robot. However, maybe it's possible with the use a https://www.rethinkrobotics.com/sawyer . Or just a printing robot.

[project-turrim]

This is a printer that can print its own parts: https://en.wikipedia.org/wiki/RepRap_project . Coupled with another robot that can assemble it, we can achieve robots that build themselves.

[project-turrim]

robot1 = MinerRobot() robot2 = BuilderRobot() simulation = Simulation([robot1, robot2]) for i in range(100); simulation.run_an_hour() end_energy = Simulation.GetEnergy() assert end_energy > 100

[project-turrim]

Each robot could simply be one long heavy platform, with 4 wheels, and 1 or 2 arms similar to what we see here: https://github.com/UniversalRobots/Universal_Robots_ROS_Driver . You'd use one arm to hold a piece, and one arm to screw the piece onto the robot you're building. The challenge here is to build the heavy platform.

[project-turrim]

• The platform can be casted in sand fairly quickly.
• It will be difficult to produce the electronic parts of the robots.

[project-turrim]

• It seems to me like it's exponentially difficult to make a robot produce a modern industrial piece of a robot in an autonomous way. So, it would be good to first make robots autonomously produce simpler parts first.