Parts ordered! Excited!

Zoombapup

Can I ask where you got the armbands etc?

I'm thinking that it'd be better to run the armbands through the outside of one of the boards there. My own plan is to create a board holder and then have a cover on top of that. So the board holder will have sort of a belt holder mechanism where you run the armbands through the board holder. I'll mock one up in 3dsmax over the weekend maybe.

My main goal is to make it so that my students don't kill them too often be treading on them while dismantling the system.

highmatergogo

Zoombapup I ordered these from Amazon:

Your idea sounds good. I'd be interested in seeing your model.

AxonSpark I saw your thread with the project boxes. Great idea and looks pretty robust. I did think about the glue gun idea to gain more grip in the elastic bands as they do seem to slip a bit in some quick tests.

I've never played with IMU type sensors but I can imagine that they need to fixed firmly to the object they're tracking. My goal is to avoid any 'slop' of the K-Ceptor. I think the best way to achieve this is to keep weight to a minimum and keep the pcb as flat as possible to the limb, thus reducing the fulcrum of which that weight can act with, creating 'slop'. This is where my idea for fairly thin plates that clamp the band material.

AxonSpark

I've found that these project boxes fit the K-Ceptors rather well, although they do require some drilling & cutting with a dremel: https://forum.chordata.cc/d/26-suit-materials/11

That said, I haven't found a project box that fits both the Hub & Raspberry Pi and is also fairly compact, so a 3D printed solution might work better for those parts.

I'll be interested to see how your elastic bands work out, as I just used regular velcro straps. I've been trying to find some sort of rubber solution that I could apply to prevent the slipping I'm getting in the thighs and upper arms (similar to Perception Neuron's straps). I'm not sure if this would be as simple as using a glue gun, or if something else would be required.

highmatergogo

So this is basically becoming a build thread 😂 Hope that's OK!?

Broke out the sewing kit to see if an idea I had will work; looks promising! Has anyone done anything similar? Not sure if the under arms would become uncomfotable after a long period.

Also cleaned my desk in preparation for the PCBs arrival next week. I have all the components now.

The IMU will easily be the smallest compnent I've ever soldered! Looking forward to the challenge!

daylanKifky

Hey!
I just found this thread. You are going so fast with your building process that I almost missed it!
It's great to see everything went smoothly, I can't wait to see the full body capture!

Going back to some of your questions and observations:

highmatergogo How does this system cope with magnetic disturbances? I know that's been an issue with the Neuron Perception kit.

The chordata system is based on inertial and magnetic sensors just as the perception neuron.
This type of capture is inherently susceptible to magnetic disturbances, but there seems to be spread concern about that particular suit being susceptible to magnetic disturbances. In our experience a strong magnetic disturbance at a short distance is required to influence the capture. For example you start sensing the influence of a big speaker at around 1 meter, and a laptop at around 70 cm.

More expensive suits implement algorithmic corrections in order to achieve better stability in the presence of magnetic disturbances. Chordata currently use the widely spread Madgwick's algorithm for the sensor fusion which is not aware of the magnetic disturbances. But there is plenty of scientific literature on this subject so we should be able to implement different algorithms once we end the prototypical stage.

highmatergogo How is foot to floor stability in the output animation?

Do you mean feet slip in translation?
translation in inertial capture is derived from the other measurements. Chordata currently doesn't implement an algorithm to obtain translation in space.
We are currently focusing our development efforts on usability, stability and documentation. The idea is: if the codebase is consolidated and well documented, and the overall system is easy to use improvements like these might start coming from the community. Our main goal is to provide a solid framework where to build a better mocap system all together.

highmatergogo So this is basically becoming a build thread 😂 Hope that's OK!?

That's great! and by the way: thanks for sharing your experience so meticulously. That's the kind of spirit that pushes projects like this one forward.

highmatergogo

Did a test print of the clamp idea. Seems pretty good. The hold is very solid. Think I'll try another that's slimmer.

highmatergogo

Had some fun last night!

Got a single k-ceptor setup and tracking in blender 👍

Couple more hours today and I should be able to do a full suit test.

highmatergogo

All sensors and hub finished and tested.

Will be doing a full body test tomorrow.

Braxton973

highmatergogo im not sure if this question has been covered or not but can you give me a rough estimate on how much was the cost of all the parts needed to make one full set ? There is a large parts list but. Not sure on the total cost.

And I'm assuming that u will need to double the parts list to make 2 full sets for 2 people right?

A kick starter campaign will start soon and its starts around $800 euros apparently. I'm hoping that if someone order the parts and build it from scratch it will cost way way less ?

And is there any instruction video on somebody Soderling the components together and setting it up on the computer? Not sure how difficult the build is.

I'd prefer to build on my own because my budget is super low and I could probably only afford to spend no more then $400 for the parts as of now.. but I may need some help to build this when I decide buy all the parts nesssessary to get started.

highmatergogo

I simplified the mount plate.

This new one prints much quicker.

https://www.thingiverse.com/thing:4019444

Guiltyfeet

Amazing job Dave!

highmatergogo

Sorry! Proper link to the new plate:

https://www.thingiverse.com/thing:4019444

highmatergogo

Hey daylanKifky thank you very much for your responses.

Yes, from my basic tests, magnetic disturbances won't be an issue for me.

In regards to the foot translation, I just wondered how 'planted' the feet look? One issue with the kinect setup was that the foot would look a bit 'floaty'.

So, if I were to walk around a room, would Chordata estimate the translation from its combined inertial sensor data?

I can't wait to give it a go tomorrow!

Thank you for your kind words; I've really enjoyed the build and I'm looking forward to being an official beta tester!

Oh, did you see my question about capturing more than one suit?

GlowBallVR

Hey, thanks for sharing your progress. I might try using your clamp solution too, how did it go mounting the K-Ceptors with it?
Does the simple plate grip the elastic bands enough? Do you just use a flat plate on the other side to screw it tight with? and then screwed the K-Ceptor on top?

highmatergogo

GlowBallVR

The simple plate seems to be working just fine. Few pics of real world use...

highmatergogo

Oh! Just remember to re-calibrate after you have fixed the kceptor to the plate if using screws! The screws do affect the readings but this is fine after a calibration with them in position.

highmatergogo

BTW @daylanKifky something I noticed on the BOM during my build is that item #8 on the Hub BOM should say "R4" not "R6" and it's a THT not SMD.

Then, item # 10 should be SMD not THT.

I think that's correct?

daylanKifky

highmatergogo
we will be releasing R2.3 soon and the BOM will be updated

highmatergogo

daylanKifky Great! Sorry, I'm a bit late to this party! 😀

GlowBallVR

@highmatergogo
Thanks for the pictures, I misunderstood how it worked.

How do you secure your hub, Raspberry and powerbank?

highmatergogo

GlowBallVR Currently just sewn onto an elastic waist band! This is very much an alpha build for me to get the system up and running to an acceptable level. My plan is to rebuild it once we're getting reliable captures from it.