Jump to content

Stationary tracker drift issue


Damon

Recommended Posts

Hi, we have been using the Vive Trackers and have recently noticed a data drifting issue. I'm not sure what could be causing it, but I can definitely say it's not the kind of drift you see when tracking is lost and then gained again. It happens when the tracker is just placed somewhere and not moved at all, with nothing in the way.

I've attached a graph of data we recorded showing what I'm talking about, which shows the X position value of a tracker over time, a minute or two. At the start, the tracker is just sitting on the floor, where it had been for a few minutes before the start of recording, and you can see there is little movement (just the usual noise, sub-millimeter). Then I pick up the tracker and move it around a bit, and that's the big jumps that go above and below the image. When I place the tracker back on the ground, you can see that X position start in one place and slowly drift down, eventually evening out at the end of the graph. That amounts to about 3 mm of drift, even though the tracker is on the floor, not moving at all. I have other graphs showing very similar "movement."

We are seeing this consistently. That is, after moving a tracker and then placing it somewhere solid where it doesn't move, there is about 10-15 seconds of data drift of about 2-3 mm before it basically figures out where it is. This is true for all three axes.

We've tried everything we can think of to remove this drift - using up to three lighthouses, shading the windows in the room, turning out the lights, and removing any shiny objects from the tracking area. We even tried turning a couple of the lighthouses upside-down (saw that solution somewhere on reddit).

Is this expected? Does the hardware just work this way, honing in on a true position over a number of seconds after a tracker stops moving? Is there anything we can possibly do about it?

rawXfromSteamVR-4.png

Link to comment
Share on other sites

@Damon - One piece of early internal feedback I received is that this could be related to thermals within the basestation as the trend your data shows is similar to trends we see around thermals. Positional data from stations can have more variance when the station is startup up and warming up. They're recommending that you test with the stations fully warmed up to rule that out as a potential source of the problem. Basically they're suggesting that you leave the stations running for 3 hours (you may need to disable BT power management) and then testing to see if your pose estimation is more accurate and if you still see the drift you're reporting above.

Link to comment
Share on other sites

Interesting. We never have thought about the base station temperature and its possible relationship to the tracker positioning, and I'll have to remember that. Why would BT power management matter on these if the base stations were just plugged into the wall anyway and aren't connected via bluetooth to the computer? Would turning on bluetooth on the computer and connecting to each base station possibly solve something like this?

I'm certain I can rule the thermal issue out in this case, though. I did do some tests with base stations that had just been turned on (actually when I used three, two were definitely warmed up already and one was cold), but some, like with the graph I posted above, were done with two base stations that had been on many hours, overnight.

Our idea was that this drift is some kind of pass filter for the data that steamvr automatically applies internally in order to increase accuracy, but we weren't sure.

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
×
×
  • Create New...