By The Orbital Index: https://github.com/orbitalindex/awesome-space

RL related:

https://github.com/Kirkados/D4PG-Phase-3

A D4PG Implementation for manipulator-enabled spacecraft, tasked with capturing and stabilizing a piece of space debris. Trained entirely in simulation, it will be deployed to the Spacecraft Proximity Operations Testbed experimental facility at Carleton University.

https://github.com/wp6726454/waveglider_RL

https://github.com/DLR-RM/stable-baselines3

Stable Baselines3 (SB3) is a set of reliable implementations of reinforcement learning algorithms in PyTorch. It is the next major version of Stable Baselines.

https://github.com/pytroll

PATRUIS (FOSS): a very complete library including as well basic classes and methods (as frame changes or orbital parameters conversions) as much higher level ones (orbital propagator, attitude laws or maneuvers sequence …).

Relation to Orekit???

Other softwares from CNES: A full list at: https://logiciels.cnes.fr/en/cat%C3%A9gorie-de-logiciel/dynamique-du-vol

42 is a comprehensive general-purpose simulation of spacecraft attitude and orbit dynamics. Its primary purpose is to support design and validation of attitude control systems, from concept studies through integration and test. 42 accurately models multi-body spacecraft attitude dynamics (with rigid and/or flexible bodies), and both two-body and three-body orbital flight regimes, modelling environments from low Earth orbit to throughout the solar system. 42 simulates multiple spacecraft concurrently, facilitating studies of rendezvous, proximity operations, and precision formation flying. It also features visualization of spacecraft attitude.

https://sourceforge.net/projects/fortytwospacecraftsimulation/

https://github.com/IBM/spacetech-kubesat

KubeSat is an open-source project for building a cognitive, autonomous framework for satellite constellations and swarms. It provides the framework needed to develop and operate tasks to be performed on Satellite. Also, it allows for the simulation and optimization of multi-satellite communications.

Too complicated for me now, as of 20210706.

News reports:

• Two new open source projects, the Space Situational Awareness project and the Kubesat project, give more people access to space tech and democratize access to…
• Open source gravitates to outer space

Based on

1. NATS (What is it?)
   • https://toutiao.io/posts/p883vaw/preview NATS是一个开源、轻量级、高性能的分布式消息中间件，实现了高可伸缩性和优雅的Publish/Subscribe模型，使用Golang语言开发。
   • https://zhuanlan.zhihu.com/p/40871363 核心基于EventMachine开发，原理是基于消息发布订阅机制，每台服务器上的每个模块会根据自己的消息类别向MessageBus发布多个消息主题，而同时也向自己需要交互的模块，按照需要的主题订阅消息。能够达到每秒8-11百万个消息，整个程序很小只有3M Docker image，它不支持持久化消息，如果你离线，你就不能获得消息。使用nats streaming可以做到持久化，缓存等功能。
   • Similar
     • Lightweight Communications and Marshalling (LCM)
2. Kubernetes (What is this?)
   • https://zhuanlan.zhihu.com/p/29232090 是一个开源的 Linux 容器自动化运维平台，它消除了容器化应用程序在部署、伸缩时涉及到的许多手动操作。换句话说，你可以将多台主机组合成集群来运行 Linux 容器，而 Kubernetes 可以帮助你简单高效地管理那些集群。构成这些集群的主机还可以跨越公有云、私有云以及混合云。
   • https://www.redhat.com/zh/topics/containers/kubernetes-architecture
   • https://www.redhat.com/zh/topics/containers/what-is-kubernetes
3. Redis (What is this?)
   • https://www.jianshu.com/p/01b37cdb3f33 Redis 是一个 Key-Value 存储系统。Key-Value Store 更加注重对海量数据存取的性能、分布式、扩展性支持上，并不需要传统关系数据库的一些特征，例如：Schema、事务、完整 SQL 查询支持等等，因此在分布式环境下的性能相对于传统的关系数据库有较大的提升。

https://github.com/IBM/spacetech-ssa

Low earth orbit (LEO) is a crowded place and the number of anthropogenic space objects (ASOs) entering LEO is rapidly escalating. With this population boom also comes the inevitable increase in close encounters (conjunctions) between objects. The heart of the space situational awareness (SSA) problem is to predict where ASOs are and where they are going; everything from a spec of paint to the International Space Station.

https://github.com/Kirkados/D4PG-SPOT

Code used to train a chaser spacecraft to capture a spinning target with a manipulator

https://github.com/Kirkados/AIAA_GNC_2021

A chaser quadrotor is tasked with learning how to track a target quadrotor on its own using deep reinforcement learning. In order to have the technique transfer well from training in simulation to a real experimental facility, the deep reinforcement learning is restricted to guidance-only. Velocity-based and acceleration-based guidance approaches are compared in the paper: Hovell, K., Ulrich, S., and Bronz, M., “Acceleration-based Quadrotor Guidance Under Time Delays Using Deep Reinforcement Learning,” AIAA Guidance, Navigation, and Control Conference, Nashville, TN, 11-21 Jan, 2021.

These two cannot run under Windows because they requries the package pyvirtualdisplay for rendering.

Cosmographia is a visualization program rendering the solar system and its bodies in 3D to create a freely navigable map of the solar system. The program allows manipulation of time and camera (observer) position. It can use SPICE data to visualize trajectory, orientation, and sensors flown on and observations taken by interplanetary spacecraft, to support scientific or engineering analysis, and perhaps even public outreach.

NAIF offitial site: https://naif.jpl.nasa.gov/naif/cosmographia.html

User's Guide: https://cosmoguide.org/