Looking for help from a GUI Programmer, working with C++ and is interested in working on a game changing public project

[onthegrid007]

Looking for help from a GUI Programmer, working with C++ and is interested in working on a game changing public project.

DM for more info if you are interested. The back-end code is in written C++, must know basic to advanced GUI programming skills. Must be able to develop and work collaboratively to do updates and upload a new version of the GUI every FRC game.

This is not paid. Purely volunteer work, if you want your name on something monumental in FIRST history, this is the way forward. The entire project will be have a public GNU license on GitHub.

[connor.worley]

Jest aside, can you share any more information?

[onthegrid007]

The back-end consists of Machine Learning Neural Networks, code that with a proper GUI could easily in the next few years take over the autonomous programming environment and easily exceed PathFinder. It was attempted publicly only once back in 2010, and I haven’t seen anyone making noise of a new and upcoming one.

Believe me I am not naive. There could quite possibly be a worlds winning team that has this same type of code, and keeping it private is their way of winning. I am also aware of rule R15 is which I want this to eventually become a public project where anyone can help fix bugs and suggest changes.

I suck at any GUI QT type programming, I am however amazing at back-end coding. If there someone out there that wants sole recognition for creating the first GUI for this project, that is what I am offering.

Regarding the meme, my code may be possibly worth money, but that’s not what I am after. If I can get my code onto thousands of teams robots like PathFinder, and it work better and have more functionality… That’s worth more than any money I could ever want.

[marshall]

Might I suggest making this an open source project and making it publicly available on GitHub or GitLab so that others can contribute to it as a project. No need to be secretive about it.

[virtuald]

You might try dear imgui, which the WPILib GUI simulator uses. It’s pretty simple to grok compared to other GUI frameworks.

But if you’re going to make it public eventually, I don’t see why you aren’t starting public.

[calcmogul]

@onthegrid007

The back-end consists of Machine Learning Neural Networks, code that with a proper GUI could easily in the next few years take over the autonomous programming environment and easily exceed PathFinder.

What is it though? Could you go into more detail on what problem you’re trying to solve exactly, and how machine learning, from a technical point of view, solves that problem?

@onthegrid007

There could quite possibly be a worlds winning team that has this same type of code, and keeping it private is their way of winning.

Excuse me while I go grab my tinfoil hat. /s R15 also requires that they open-source anything they reuse from year to year. Also, there’s far simpler, more effective means of robot control than ML in FRC at this point. That’s what the top-tier teams are using. They want to find the Pareto optimal solution to the reward-effort tradeoff (best reward for a given amount of effort, or the least amount of effort for a given reward).

@onthegrid007

I want this to eventually become a public project where anyone can help fix bugs and suggest changes.

We’ve had community projects get widespread official adoption before. Pathfinder was one (it got replaced by WPILib’s trajectory generation in 2020). There was also https://github.com/wpilibsuite/frc-characterization. As @marshall said, open source it. If it turns out to be useful, we might adopt it as an official tool.

[Nate_Laverdure]

@mods suggest merge with existing thread

FRC Neural Network - HELP

I created a neural network. 3x8x16x8x2. Activation is tanh. It took in delta encoder values one for each dt side, a millisecond clock value scaled for a 30 second auto between 0 and 1. The outputs were two values between -1 and 1 I scaled them up to -12 and 12 voltage-current values to control each dt side. For back-propagation, I wrote a recorder to capture the voltages of the motors at every watchdog interval using a customized mutex timing lock (thread-safe), then scaled the values and bac…

[hamac2003]

This will probably get flagged, but I don’t care