How to use docker to run PAL's REEM-C Gazebo sims?
I want to run Gazebo PAL's REEM-C sims. I started out with ROS Kenetic on Ubuntu Xenial system resources a couple years ago. Completed “boot camp” using Kenetic/Xenial. To check my reinstall, I re-trained with every tutorial, except by using ROS Melodic/Ubuntu 18.04.2 LTS. Gazebo basic training (3 wheeled bot, Velodyne) required moving meshes between Inkscape, FreeCAD, Blender. However, my quad core tutorial box (over time and a few ROS and Ubuntu re-installs later) now runs Bionic Beaver. It was suggested I use a docker to run REEM-friendly Gazebo environment so I don't have to redo my training box. If I don't see otherwise I will install docker and try to get docker to emulate the ROS Kenetic/Xenial environment required by ROS/Gazebo REEM-C sims. This may blow my training enviroment away, require wipe-and-reload, so I'm hoping somebody who has worked with docker can advise me if what I am planning on doing is feasible. I'll have plenty of lab resources to accelerate my goal of using REEM-C in non-industrial environments if I can just show C in custom progressive training modes to my non-technical, affluent associates, but not before. Any help is appreciated in advance.
Edit:
REEM-C
The goal I want to achieve is not to simply sell PAL robots or show showcase REEM-C's bipedal prowess, or kicking a ball thru a goal post, or charging up an uneven hill with a KRISS Vector, but to simulate REEM-C's carefully engineered quality and superior agility in Gazebo's fixed settings. I know my targeted user settings and their stated requirements thru 15 years of well-earned trust and experience.
Full requirements of what I want to achieve are available, all doable in my opinion, in ordinary PAL/ROS/Gazebo simulations.
REEM-C's very first hypothetical performance is included below to answer your question about me explaining more about what I want to do with REEM-C in Gazebo simulation. PAL's REEM involvement in my Gazebo world would begin and finish in 7 fixed-process settings that, if done correctly may not change much. These real-world, existing packaged processes have changed very little in the last 50 years.
Once I/we get Gazebo simulations correct, we advance to the next six of seven total possible settings. I am already convinced (beyond doubt) a bipedal ROS robot can execute correctly in the seven categorical requirements. I've designed robotic operating system software for manufacturing, performed AI research for Honeywell Aerospace, been a System Analyst, IT Manager, Director, and completed basic ROS/Gazebo training over two ROS/Ubuntu upgrades with correct simulations.
This is where I need PAL's help: taking my system design (and coding) expertise to PAL (inside Gazebo) simulations with a Kenetic docker in my desired situation.
Once we have successfully completed a few Gazebo videos of one or two of the six categories, I'll be reaching out to my affluent associates with our newly constructed, PAL-approved Gazebo videos ...