Jump to content

Inconsistence and wrong accuracy sometimes for some trackers and controllers with 4 base stations


pnuzhdin
 Share

Recommended Posts

Hi,

 

I am using Vive Pro with 4 Lighthouse ver 2 base stations. I am using stable SteamVR 1.1.4.

 

I've found that sometimes some trackers and controllers become inconsistent and have wrong accuracy relative to other trackers. I am not sure that it's reproducible using 2 base stations only but I can reproduce with 4 base stations for sure. Also while it's the intermittent issue it's reproducible quite often in a case you restart SteamVR or power off and power on trackers or controllers. Sometimes it can be fixed by itself but more often it's not fixed until you power off all base stations and power on them again.

 

The error between a consistent and an inconsistent tracker or controller can be very large ~0.5 meters. Also, the error depends on the distance and angle from the tracker to the base stations. It looks like trackers have different coordinate systems for a while.

 

I have other 4 Lighthouse ver 2 base stations but still can reproduce the issue so I don't think that I have 8 broken base stations actually. It also doesn't affect only some tracker or controller it's reproducible for any controller and tracker. It's reproducible on another Vive Pro with 4 other trackers and 2 other controllers on different PCs. It's reproducible at least in 2 different environments/rooms.

 

Here the video where I've started rotating 2 trackers fixed by my hand in the single point side by side. The left tracker had 0.4 meters error before I started to record the video actually. In the middle of the video, I've powered off all base stations and powered on them again to show how the issue can be fixed and how consistent trackers should actually be.
https://www.dropbox.com/s/l1p3vxqbqgir0cn/Headset%20Mirror%2012_18_2018%2010_39_16%20PM.mp4?dl=0

Here the raw and vrserver logs https://www.dropbox.com/s/nsyxg7t8f0t7109/logs.zip?dl=0.

Link to comment
Share on other sites

, I'll send you an email with some additional info about the 2.0 stations.

 

A look at your logfile shows a great number of reflection issues in some of your sessions which would could certainly cause the types of issues you described. In one instance, a specific lighthouse position  

(ending in 138F) resulted in 312329 back-hits (aka reflections - the beam is bouncing off a surface before hitting the tracked component) which is really high - another one had over 100K. Using the info I'm sending you, I would suggest adjusting your mounting locations and doing some reflection mitigation and see if it reduces your issues. It's a little hard to tell what's happening in the video without full context but it seems like the issue is dependent on which direction you're facing which supports the reflection theory. 

 

Other causes of this could be BT interference. The trackers ship with cradles for the dongles - you're supposed to space the USB dongles at least 6 inches apart because they contain powerful BT radios that will interfere with one another if they're placed too closely together. 2.4Ghz WiFi can also cause interference and cause a spotty connection. 

Link to comment
Share on other sites

, just ctrl-F "back-facing" and it will pull up the correct lines. They're usually found in VRserver but they're also scattered around the consolidated system report. I've copied a line from the log above that I was referring so you can see the line I was referring to in better context. 

  

Tue Dec 18 2018 22:40:44.590 - lighthouse: LHR-4A99138F H: Dropped 312329 back-facing hits, 41 non-clustered hits during the previous tracking session

 

Link to comment
Share on other sites

I need to do more testing, but it seems I have a high number of back-facing hits in my logs, and after a thorough search through my room, the only truly reflective thing left in the room is the tempered glass window on my wife's PC case. If I can see the reflection of the lighthouse with my bare eyes in the glass, I assume that the sensors can see it in the reflection as well?

Link to comment
Share on other sites


This is one of those troubleshooting cases where our human perception actually makes it more difficult to visualize what's happening. While a surface may be non-reflective to visible light, it can totally be reflective to IR light (which is used by the base stations). Also, objects that may be opaque to visible light can be totally transparent to IR light (

).

 

 The unfortunate answer is that unless you have an IR camera, it can be pretty difficult to troubleshoot IR reflections. One good thing to keep in mind is that objects that are a different temperature than the surround environment (such as a cold plate) can becomes IR mirrors due to the temp differences. 

 

. You can use something like your cellphone LED and just feather your finger over the LED to get the flashing effect. The only way to really dubug the window is to cover it up with something that's opaque to IR. 
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...