Pathfinder2 VS EOCV-Sim

i proudly proclaim this “the pathfinder shuffle”

i don’t hate roadrunner, but i do have an alternative - pathfinder2, a library i’ve been working on for about a year now that’s pretty similar in functionality. my team decided we weren’t going to use roadrunner because we wanted to improve the accuracy of positional tracking over time and wanted to have more control over how we could control our movement. roadrunner isn’t a bad thing - it still works pretty well and i’ve seen a lot of teams make use of it - roadrunner doesn’t appear to be in as active of development as it was before and i personally wanted a solution that was easier to use and more robust, but using roadrunner won’t mean you can’t perform well

i feel you on that. i've spent nearly a year working on a library for odometry-based autonomous navigation - our robot was capable of navigating around a field entirely autonomously - and it's quite disappointing that all of that work has no use in my last year in the FTC program. i know the change is for the better, but i feel like i've wasted a lot of time and energy

hey, i just figured i’d put this out here in case it helps: meccanum wheels, by nature, have a LOT of slip - that’s how they’re able to move in any direction. i’ve had quite a bit of experience with the math and kinematic behind this (two libraries i’ve written/am writing, Pathfinder and Pathfinder2, as well as a library the other programmer on my team wrote, OdometryCore) and it’s VERY HARD to accurately track the position of a meccanum robot only using the encoders in the wheels. it’s not impossible, but it’s very, very challenging. the more your robot moves, the less accurate it becomes, and it will very quickly report positions that are nowhere near where your robot actually is. you can go for it, but i would strongly encourage you check out three wheel odometry instead - it’s been very accurate and allowed us to do really cool things, like accurately record a robot’s movement and play it back perfectly. you’ll have much more success with three wheel odometry - my team designed a custom odometry system and i’m sure they’d be happy to give you more information if you’d like. best of luck :)

Fixed GitHub link - https://github.com/deltacv/EOCV-Sim/releases/tag/v3.5.1

Hey, i have just made a pre release containing the fix i mentioned. You can find it here: https://github.com/deltacv/EOCV-Sim/releases/tag/Dev . Please let me know if it works so that I can make the full release !

EOCV is pretty much just a "glue layer" between OpenCV and the FTC SDK. So I'd recommend starting by gaining some familiarity with OpenCV in general. You should also look into the EOCV-Sim project (developed by Sebastian) which will let you iterate much more quickly as you develop your pipeline as compared to deploying to the phones every time.

https://github.com/deltacv/EOCV-Sim :) a simple software to test opencv algorithms on Linux, macos and windows

Also easy open cv sim (eocv-sim) is a really nice tool to develop pipelines without robot with you (really nice for us programmers when you mech people take too long :P) https://github.com/deltacv/EOCV-Sim

And if you wanna develop OpenCV pipelines fast and quick without needing s phone i recommend using (shameless plug) EOCV-Sim: https://github.com/deltacv/EOCV-Sim