This project enables streaming Oculus VR headset pose information to an FRC robot using the Network Tables 4 (NT4) protocol. This pose information can be used by the robot control system to accurately map its surroundings and navigate around a competition field, practice space, or any other location. QuestNav produces a more stable and reliable tracking pose than any other FRC vision solution (LimeLight, Photon Vision, etc.)
Check out the full video here!
Ask questions and get support in the Chief Delphi thread!
Using a VR headset for robot localization has several advantages:
- Lower cost than most FRC vision solutions
- Multiple SLAM cameras enable redundant, calibrated VIO
- Camera tracking is fused with an onboard IMU to provide accurate position and velocity estimates
- Up to 120Hz robot odometry refresh rate (linked to the Oculus headset framerate)
- The Qualcomm XR2G2 platform offers several CPU/GPU cores, 8GB RAM, dedicated video encode/decode hardware, and at least 128GB onboard storage
- Well-supported ecosystem and API that enables off-the-shelf libraries for mapping a 3D environment
- Self-contained, rechargeable battery
QuestNav requires the following to get started:
- FRC robot and/or control system
- This project was tested using REV swerve modules assembled with REV NEOs and Spark Max motor controllers.
- Quest 3S headset
- A Quest 3S headset is recommended due to its lower cost and excellent tracking performance. It's currently not clear whether the depth projector on the Quest 3 benefits FRC applications.
- Supported USB-C to Ethernet + power passthrough adapter
- 3D printed mount that attaches the headset to your robot
TL;DR: An external 5V supply is REQUIRED for the headset to operate properly on an FRC robot. 5V can be supplied from any stable source - including the RoboRIO's USB ports.
As per the FRC robot rules (R707 in the 2024 ruels), wireless communication is not allowed within the robot. In order to comply with this rule, all communication with the Quest headset must be done over a wired link at competition events. USB to Ethernet adapters offer a convenient way to communicate with the headset, however, the correct style of adapter must be chosen to avoid connectivity issues.
The USB port on Quest headsets was never designed to constantly supply 5V at "high power" (5V @ 500mA as per the USB spec), so supplemental power must be provided to ensure a reliable connection. Only USB to Ethernet adapters that support power passthrough should be used on a robot! These products are often sold as "USB C port replicators", "USB C docks", "USB C charging adapters", etc. A list of recommended, tested adapters can be found here.
There are several ways to power the Quest headset + USB to Ethernet adapter on an FRC robot:
1. Power the headset using a USB port on the RoboRIO
The USB ports on the RoboRIO provide a stable 5V source that the headset can use to power the Ethernet adapter. This port will not supply enough power to charge the Quest headset, so you'll need to keep an eye on charge levels throughout the event. This is the quickest and easiest option.
2. Power the headset using a 5V USB battery bank
This approach has several benefits:
- It provides the Quest headset and USB to Ethernet adapter with clean, reliable power
- The battery should supply enough power to sustain the Quest headset's internal battery indefinitely
- Charge state can be monitored externally using the power meter on the battery bank
- Battery banks can be replaced without interrupting (powering off) the headset
- No soldering required
NOTE: It's important to make sure only 5V is supplied to the robot. You should avoid power banks that support USB C Power Delivery (PD). In our testing, we've noticed that some USB to Ethernet adapters will boot loop when a voltage greater than 5V is applied. Alternatively, you can use a USB A to USB C cable to foce a USB PD power bank to deliver 5V to the headset.
3. Power the headset using a good quality 5V regulator
This approach is likely the most convenient. However, if implemented incorrectly, it may lead to several issues that might be difficult to debug.
Recommended USB-compliant 5V regulators:
Recommended 5V regulators:
For non-compliant regulators, you'll need to use a USB breakout board like this one that includes onboard sense resistors so that the headset knows that it's connected to a 5V power source.
A high-level overview of the software architecture is shown below.
At it's heart, QuestNav is merely a VR app designed to push data to Network Tables. However, some one-time setup is required before we're able to push a custom app to a VR headset.
- Follow this guide to set up "Developer Mode"
- Sign into your Meta developer account and download the Meta Quest Developer Hub (MQDH)
- Configure the following settings on your Quest headset:
- Enable travel mode (link to instructions)
- Set the display timeout to 4 hours in
Settings > General > Power > Display off
- Enable battery saver mode in
Settings > General > Power > Battery saver mode
- Disable WiFi in
Settings > WiFi
- NOTE: Be sure to completely turn off WiFi, otherwise the headset will constantly disconnect from the robot network as it tries to look for the internet
- Disable Bluetooth in
Settings > Bluetooth
- NOTE: Disabling Bluetooth will break the companion app functionality
- Disable the guardian for development purposes
Settings > Advanced > Experimental Settings > Enable Custom Settings
and TURN OFFPhysical Space Features
,MTP Notification
, andLink Auto Connect
- These settings might also be located in
Settings > Developer > Experimental Settings > Enable Custom Settings
on some older OS builds
- These settings might also be located in
- Plug the headset into your PC and install the example .apk using MQDH or adb (
adb install QuestNav.apk
)- NOTE! The example app team number is hard-coded to 9999
These setup steps are required once per boot. This setup process can be prevented by ensuring your headset remains powered on using an external battery.
- Plug the USB-Ethernet adapter into the USB port on the Quest headset
- Start the QuestNav app using MQDH or by selecting the app icon in the launcher
- Check that the Quest headset has connected to your robot and is writing pose data
- Download and install Unity Hub from the official website (link)
- Open Unity Hub, sign in, and install Unity 6 (6000.0.25f1) LTS
- Select the following:
- "Microsoft Visual Studio Community 2022"
- Android Build Support"
- "OpenJDK"
- "Android SDK & NDK Tools"
- Select the following:
- Click "Install" and wait for the installation to finish
The Unity installer may fail to install "Unity Support for Visual Studio" if you already have Visual Studio 2022 installed. In this case, you'll need to install it separately.
- Download the Visual Studio 2022 Community installer (link)
- Run the installer and select
Modify
- Scroll down to the
Gaming
subsection and selectGame development with Unity
- Click
Install
and let the installer do its thing
Installing Quest Link is not necessary, but may make
- Install the Quest Link software (link)
- Install Git using whatever method you prefer. You can download Git for Windows here.
The main editing window will only open if a project is active.
- Click
Add > Add project from disk
and select the unity subfolder in this repository - Click on the newly imported project
- Wait for Unity to compile assets and open the main interface
This package is required by the C# Network Tables library.
- Download the latest release with a
.unitypackage
extension from here - In the main unity window, select
Assets > Import Package > Custom Package
- Browse to the package you downloaded and click
Import
Be sure to resolve any Project Setup Tool warnings that appear in the Console
tab! You need to look for warnings in both the PC
and Android
tabs to successfully build your Unity project.
- The build tool is located in
Meta > OVR Build > OVR Build APK...
- Set your desired OVR build path and click
Build
- This tool prioritizes build speed above all else, so it will consume all CPU resources for a few minutes!
- If everything works, then upload the .apk to your headset!
The installer may not automatically detect Visual Studio as the default editor for Unity. This may lead to missing Unity namespaces and missing debugging options.
- Follow this guide so that Unity knows to use Visual Studio for debugging
- The main scene is located in
Assets > Scenes > QuestNav
- The pose streaming code is located in
Assets > Robot > MotionStreamer.cs
.
Unity will need to download and install several Android packages during the first build, so it might be helpful to connect to the headset over USB the first time to make sure it can download everything from the internet. Subsequent builds should be faster and will not require additional downloads.
Q: Are you doing anything to initialize its location? Or do you have an idea how you'd recommend teams to initialize its location?
A: QuestNav currently does not do anything to initialize the robot position beyond reseting it so the "field oriented drive" is correct. I'm sure someone will come up with a way to use April tags for initialization. There's a Quest development API that enables placing locations on a 3D map and sharing that map between headsets.
Q: How does the Quest headset do with hard hits from other robots? What happens if it loses its FOV by sitting in front of a wall for awhile (aligning or whatever to score)?
A: It does okay from what I've found. The test robot don't have bumpers so robot-to-robot testing has been limited. The headset has a very wide FOV, so some occlusion is okay. We've even gone as far as covering the two front cameras and the robot localized successfully.
Q: Does the Quest headset need to see one side of your robot or does it just need to see out of one side of your robot?
A: The headset just needs to see out one side of the robot.
A: QuestNav uses a USB-Ethernet cable connected to the Oculus headset's USB port to communicate with the RIO.
A: Make sure the headset is powered on and the app is running. You can also check that the headset is shipping data to the robot by checking the variables in the Driver Station.
A: Currently, there are two levels of "zeroing". There's the local "zero" that sets the robot heading for field-oriented drive. There's also a pose "zero" that's equivalent to long-pressing the quest button for a few seconds that is computed on the headset. The latter is currently triggered using network tables to indicate to the Quest headset that the user has requested a pose "zero".
A: Pull requests adding this feature are always welcome!
Q: Can I push this around on the cart during camera calibration, or do I need to push with wheels on carpet or wait until a practice match?
A: The headset doesn't necessarily require that you show it the field before using it in a match. Much in the same way that you don't need to calibrate anything before playing a VR game.
Q: Does running it in the pits or practice field mess with its understanding of a competition field's spaces? I don't know how long it holds on to remembering a space, especially without an internet connection.
A: Nope, it doesn't mess with anything. In fact, it will probably infinitely expand the map as you travel from the field to the pits with the headset powered on. I've seen my headset accurately map our entire house!
A: QuestNav is designed to publish the pose information to network tables on every VR display frame update. This means that the update rate is tied to the display frame rate, which by default is 90Hz. You can push it to 120Hz without any issues.
A: QuestNav currently ships back a frame count and headset timestamp along with each pose estimate packet. The headset latency has to be very low, otherwise, we'd make people very sick very quickly. Robot code will probably be the larger latency contributor.
A: I recommend using the Quest 3S on robots.