Discussing F´ integration with JPL-developed planning and analysis tools

[barugonman]

Hi all -

I'm working with other JPLers to continue developing planning and analysis software, that was largely originated from JPL projects such as M20 and Clipper. The plan is to complete the migration of these tools into open source for continued development, and complete other valuable functionality that had been identified during the engineering phases (but never funded). Me personally, I happen to have retired from JPL recently but intend to continue working on these tools for the time being. Tools include the Aerie/PlanDev suite that has been open source for some time, and also the spacecraft analysis tools for which the open source process was completed.

Anyhow, these sorts of tools are best developed with the support of a flight system simulation environment, such as the Workstation Test Set/WSTS flight software development frameworks used at JPL. And further, we needed something fairly generic and extendable. I had been considering ROS as I had worked in that environment before, however it was recommended I review F' as an alternative. I did start reviewing and so far out of the box F´ looks to fit my needs pretty well. The command-and-telemetry GDS system is familiar enough for me, and I was able to hook up the hello-world demo to one of our Python plotting libraries without much trouble, so my current plan is to continue down the F´ path. (Furthermore I've worked with at least a couple of the developers on this project, that give me a warm fuzzy feeling about the whole thing :-) )

I'll be skimming through the repos to see what sorts of things have already been built on/around F', but would be interested in any pointers people have to similar sorts of efforts to review. (Or anything else I should keep in mind).

Additionally, if working with these new tools sounds interesting, people are welcome to check out what's been built and give things a try once we get a bit farther with the tool integrations.

Let me know what you think -

Thank you -

Dan Allard

[vincewoo]

Hi Dan! I didn't realize that you retired, congratulations! I also didn't realize that these tools had been open-sourced, really appreciate your continued work on them even in retirement. I'm working on the Clipper Ops team these days and having a tool like Aerie/Merlin is a huge improvement to the many Ops planning activities we've had to do.

Our section is definitely investing a lot into F' for the foreseeable feature so it would be a great option for using as a basis for your development. It is also gaining popularity outside of JPL as well with many universities and commercial applications.

Totally support this effort!

[barugonman]

Awesome great to hear from you Vince, you were one of the folk I was happy to see onboard :-) ! Yea we got the Clipper built set of tools out, plus others like the memory analysis tools and the rest of the EAS tools are expected! Will update in this thread as I go, take care -

Dan a

[mrmuszynski]

I've been thinking about what I want to contribute to this effort, and one thing that keeps coming to mind is that in order to appropraitely develop, test, and demonstrate planning and analysis tools effectively, it would really help to have some kind of reference mission defined. @vincewoo do you know if anyone in the FPrime community is doing something like that?

Right now my vision would be to define a simple LEO mission (some colleagues have recommended using FireSat as inspiration)... just enough to define a realistic orbit (probaby using a real mission as reference and just propagating out TLEs or something), and defining a relatively simple set of commands in an F-Prime deployment that don't really do anything under the hood other than make some engineering data that can be used to drive analysis development and demonstration. The point would be to make something that you could command via products built in Aerie, that would react to simulated orbit events, and that could produce some nominal and off-nominal demo telemetry so potential users of analysis tools can see how they behave in various scenarios.

[barugonman]

@mrmuszynski FYI most of what I've seen along those lines is here: https://github.com/orgs/fprime-community/repositories?type=all but haven't dug that deep into it.

[barugonman]

A bit more about ideas I've had about robotics demonstrations with fprime and the tools:

• Find a simple robot toy that has an interface I could plug fprime into, control via. commands and sequences. Figure out some sort of 'monitors' to provide telemetry
• Animatronic halloween figure that runs off a master sequence, but also reacts to commands and the environment in some fashion. Reacts with different movements and voice clips according to reasoning in the FSW
• ...

[barugonman]

Here's the overall architecture of the first prototype I was looking to complete, which I consider an MVP set of tools for demonstrating the analysis 'close the u' capability:

And here is a 'plan' of sorts for my progress. The top row is for the fprime/dtat-library channel integration I finished up (but still need to get checked in properly), and the bottom section shows the rest of the tools being integrated and the order I'll probably follow to finish integration:

Noting while I did make progress on 4, I decided I wanted to get more of the overall infrastructure in place before I had to dig in and actually figure out Aerie command expansions.

(Also noting I'm trying to use Mac Keynote for diagrams, after years of powerpoint, haven't quite got the hang of it yet)