ftc-dashboard VS PhotonFTC

I would highly recommend this library. You can tune variables quicker using it. https://github.com/acmerobotics/ftc-dashboard

Noted, thanks! Usually when we connect to the web interface, it's to use FTC Dashboard, which uses 8000 for the websocket.

FTC Dashboard lets you graph telemetry and download it as a csv file.

FTC Dashboard shows a live feed of the camera if u set it up right.

It could be a memory leak somewhere. If you use FTC Dashboard, make sure to update to the latest release (0.4.7) as it fixes a memory leak when the dashboard is installed but disabled, which was affecting some teams at competition. If you are using vision code, make sure you are not creating new objects every loop cycle, which will fill all system memory until it crashes. (We have had this issue multiple times in the past when we are using OpenCV-based detection, it's an easy mistake to make.)

Yeah - just run it from the dashboard. https://acmerobotics.github.io/ftc-dashboard/

The SDK has a built in camera stream, which isn't great, but it can help. If you are willing to go a little more complex, look at ftc dashboard It has a few really great features including the ability to live view the camera stream and allowing you to set up variables so you can update them on the fly.

This is the change that was needed for FTC Dashboard: https://github.com/acmerobotics/ftc-dashboard/pull/88/commits/e85c0e7134035650d2a6d3d90c2bbf2a882580f4

FTCDashboard

Photon FTC implements some of these changes (https://github.com/Eeshwar-Krishnan/PhotonFTC) specifically check out the new driver (VL53L0X the name of the chip used in the 2m sensor) (https://github.com/Eeshwar-Krishnan/PhotonFTC/blob/main/PhotonCore/src/main/java/com/outoftheboxrobotics/photoncore/Neutrino/Rev2MSensor/Rev2mDistanceSensorEx.java)

It's been a while since I've dived into the SDK. If you have specific questions on the thread lock, there's several people on the discord that know how the thread locking works very well. I haven't tested it myself, cause it came out after I graduated, but Eeshwar developed software to overcome the lock and speed stuff up by doing calls asynchronously. It's still better to limit to a single thread though. https://github.com/Eeshwar-Krishnan/PhotonFTC

Yeah saw it a lot when control hub first came out. Now see people use it for the acceleration from photon - https://github.com/Eeshwar-Krishnan/PhotonFTC. I believe it is legal.

We tested the PhotonFTC (v2.0.1 as we're still on SDK 8.0) software. After adding the recommended repository and dependency statements to the teamcode build.gradle, and changing minimum android SDK to 24, we reran the testing opmode, in two ways: 1) No PhotonCore.enable() statement, and 2) with a PhotonCore.enable() statement. The result was: (1) loop time, reading all four sensors, was 30-35ms; and (2) also 30-35ms.

PhotonFTC (my personal favorite)

Photon?

For autonomous in addition to drive encoders or dead wheels for position estimation, teams have found that using distance sensors can be used to relocalize. Distance sensors however often require you to make a ton of very time expensive calls. Team 7244 developed a library called photon that helps remove a ton of this delay, as well as enabling a higher precision mode on the rev 2m distance sensors. https://github.com/Eeshwar-Krishnan/PhotonFTC