well , they possibly take advantage of the helmet tracking and link their own calculation to the helmet position (after all they probably need this to recreate a full body IK).
So you can at regular interval reset foot position just under the head position (that is actually a best guess).
All 3 DOF IMU are easily turned into 6DOF with a bit of calculation , the only problem being the drift. The other problem is also common, since all the skeleton used in this field are usually proprietary (depending the position and number of trackers).
if you have a body tracking suit (like neuron or salto) you know that porting an IK from one model rig to another is not simple (axis reversed for example).
When i see in the demo the belt sensor being clipped on the side of the player (when logical position should be on the back) i think the IK is probably very approximativeā¦
basically they calculate a triangle with top being the belt tracker and the bottom being the 2 feet trackers. From that you can have an idea of the level of the user above the floor (if stand up or bending knees) and if feet are on the floor or not and approximate position. That is not rocket science but barely precise enough to be very useful.
the only valuable feature here is you can expect to have decoupled information for the direction you walk vs the direction you are looking at.
Could be interesting for games like āJus Danceā or āDance Centralā.
I asked them on Youtube and they told me they operate in 6dof.
Did they say how this happens from a technical standpoint? Do they have any real way to tell exactly where they are in space, or are we in fact talking about the sort of scenario I and nosys70 describe above?
My comment is in there
ā¦by which I suppose you mean not in the embedded video, but in the cesspool of abuse that tends be appear beneath any video on any Youtube video pageā¦ Okā¦ lookingā¦
I see nothing in their answer that reassures me.
if you look at 2:20" , the skeleton shows only 2 values , pitch and roll that is only 2DOF.
Since most IMU chip are 3DOF (angular) , i expect them to have this too, but we can believe that the 3 others DOFs are simply calculated (time and acceleration) that is good enough for gaming (if you can deal with the drift)
if the price is fair , this could be a nice add-on for some games.
Thanks for sharing. Backed that too. For 99 USD (didnāt make the extra low early bird at 79) itās worth a shot, even if Iāve already backed WalkOVRā¦ One of them has to work outā¦ 
EDIT: Backed THAT tooā¦ Not backed OUT tooā¦ DYAC!
Both projects have merits & uses.
Kat Loco is simply by the looks of it a locomotion solution. 3 sensors. (Motion detection?)
Walkovr is aiming much higher with a base if mem serves 5 nodes to allow Mocap capability & useage with expansion to upto 15 nodes plus a controller plugin.
Just backed Kat Loco as well. It is a decent cheap solution.
please allow this to have body tracking thatd just be the icing on the cake
For body tracking go for Walkovr; itās designed with expandable Mocap.
Walkovr
-
MOCAP (Motion Capture) -
5 sensor base (2ĆAngle 2ĆKnee 1ĆWaist)
-
Expandable to 15 sensors + 2ĆControllers
-
Sensors Wired to Waist sensor. Bluetooth -
Kicking could be supported
-
Expansion port
-
Mobile VR support
-
Applications outside VR Cheap MOCAP
-
Price
Kat Loco
-
Motion Detection (Based on readup)
-
3 Sensors (2ĆAnkle 1ĆWaist)
-
Locomotion System Solution only
-
No marketed Expansion
-
Wireless Sensors
-
Price
Walkovr has more pluses if your looking for more than a Locomotion system but Sensors are wired to base.
Kat Loco is good if you want wireless sensors with easier setup (3 sensors no wires). But has limited expandability & likely more latency. Just remember to take off angle bracelets to avoid being possibly treated like a criminal when going out. 
I dunno on these. I agree with Helio these are a much more simple solution, but I think for the price difference, the WalkOvr look to be better. That said, I suspect either will be much more physically exhausting to make extended use impractical. I backed the Cybershoes as something of a compromise here since you donāt have to support full weight, though they have their own issues as wellā¦including the fact their timetable keeps slipping.
I would probably have backed WalkOvr too if they had worked out a wireless solutionā¦I see the wires as problematic. It can already be a chore to get set up to play VR depending on your set up and having to strap on yet more sensors, particularly with wires, takes this beyond the Nth level of nerd we already are for being in VR to start with.
I backed both as each should be good for different uses.
Walkovr from past looking into MOCAP suits looks to be very inexpensive system. At Immersed 2017 there was one I think called Shadow Play(?) And it was very ecpensive back then if mem serves well over $1000.
Now to compare I think youād need 4 nodes more. Add Lighthouse tracking & Hand Tracking module you could likely play with out controllers. 
I meanā¦ couldnāt you just sit in a swivel chair and walk in place. 
With the WalkOvr it looks like the wires may be issue. On the Kat Iām not sure.
I already got the perception neuron and believe me, you would not wear that on and off for gaming. I think Kat has more chance because it is easier to plug an play.
Sure for Easy on off Locomotion Kat will have a benefit. Walkovr though you will have more than simple locomotion. Ie Possible kicks & mocap avatars.
i think Kat will be able to kick or other sort of leg movements.
Kicks are not precise, you just need to get an idea of some sort of fast movement happening. And most average human is not able to manage real kicks anyway.
The same for walk, you just need to get an idea if the user is walking or not, the exact place of his feet is not really important. That is also why the IK for legs is so easy and you do not need trackers for knees. if you know where are hips and ankles, you can get a pretty neat idea where are knees.
Am i pretty sure you could even do it by getting a sensor only on one foot.
It is all about gravity center.
Think about it like this toys made of an elastic passing trough straws.
Could be good enough for Soccer game also.
Well thatās why I said WalkOvr is more of a MOCAP system with Locomotion. Tracking the leg in WalkOvr will have benefits inside & outside of VR.
Kat Loco will be good for multiple users due to price & wireless. Where as suiting folks up in an Arcade Walkovr would not be a good choice.
It will be nice to explore both ideas. I am looking forward to extra nodes + Handtracking. 
I can see flick kicks being possible with Kat. But for those into aerobics & martial arts a full MOCAP can be good. Think of Guitar Hero vs Rocksmith.
for hand tracking, i imagine more a kind of wrist band with a camera like the leap motion that can track all fingers.
you woudl just need to attach them like you wear a wristwatch.
the advantage of doing this is the tracker is alway with your hand so it works in any postion , and you can track arm position at the same time.