AWS JPL OSR Challenge #88

[Gizmotronn]

http://spacechallenge.tech
Documentation for the challenge - December2019-February2020

[Gizmotronn]

This is an ml challenge - find out more here: https://github.com/EXYNOS-999/AWS_JPL_DRL#workspaces/aws-challenges-5e09869a2eaf9046251f8fc4/board

[Gizmotronn]

#90 Gazebo sim is now installed. This means that my local env. is ready for testing with AWS!

[Gizmotronn]

So the big accomplishment for today: Gazebo & Linux (dual boot w/ Windows 10) #90 installed. That's kind of the big accomplishment for this week as well.
Also been doing some stuff in IrisDroidology/Manacaster - which has been forked to Github/Acord-Robotics.

I'm currently searching for Open Source science projects to do as well as having a look at the osr-code repo @nasa-jpl @acord-robotics/droid-building @acord-robotics

Contributors
@limenutt @EXYNOS-999 - contributors to this project

Dry Run
Dry Run on Sage/Robomaker with AWS & GazeboSim (sh) - initial dry run, rover snagged a rock :(. Got some more ideas after MY AWS DeepRacer challenge initial sim/run - #88 w/ the reward function:

#88 
params (for Deepracer): 
1. All wheels on track - important, as this improves stability (if all 4/6 wheels are on the track, that reduces the risk of crashing). This is important for both DeepRacer & JPL, however with the AWS JPL OSR challenge it will NOT be called all_wheels_on_track (unlike with DeepRacer AWS) because it's not a track - it's a sample of the Martian surface


2. Distance from center - more important on DeepRacer than on the JPL agent (this is because as long as the rover/agent stays within the boundaries, it doesn't matter. Since it isn't a race car, it doesn't need to worry about sharp turns. The rover will need to turn to avoid obstacles, however, but the speed will not be the same for the rover as for the DeepRacer car. There may be obstacles within the centre of the environment, which means that if this parameter is included (or prioritised) it will likely go into the obstacle). 


3. Speed (m/s) - important (the rover needs to complete the course fast, just like with the DeepRacer car/agent)

4. Fuel - important (there is no way to refuel/help out the real Mars rovers, and the same goes for this JPL challenge. It isn't necessary for the DeepRacer challenge, however it is one of the parameters that will be incredibly important for the rover agent. Efficient fuel consumption is a must.

5. More to come

Liam
@IrisDroidology

[Gizmotronn]

https://github.com/IrisDroidology

[Gizmotronn]

rishistyping/AWS_JPL_DRL@3a864f5

[Gizmotronn]

New Trello activity from card: Security Groups @ Board AWS_JPL_OSR_DRL

View here: 65971 https://trello.com/c/mC9wTutM/8-security-groups

[Gizmotronn]

A

[Gizmotronn]

New post from Trello Board: Reply To: A0-D2 Build Log – Siridium – July-September 2018 by Gizmotronn
Link: http://allianceofdroids.org.au/aod/jedi-order/droidologists/forum/topic/a0-d2-build-log-siridium-july-september-2018/#post-65960

[Gizmotronn]

rishistyping/AWS_JPL_OSR_DRL#1

We have moved over to the new repository now!

[Gizmotronn]

rishistyping/AWS_JPL_OSR_DRL#3

[Gizmotronn]

Thanks for an awesome challenge everyone
@wakeuplearn
@IrisDroidology
@limenutt
@EXYNOS-999
www.spacechallenge.tech