Hand based input list

Manu Machinae recently launched a kickstarter for a VR glove. It is more basic than yours but it is in the same price range: 200 euro for a pair.

https://www.kickstarter.com/projects/606955275/manus-machinae-be-part-of-the-machine

"serrarens" wrote:
Manu Machinae recently launched a kickstarter for a VR glove. It is more basic than yours but it is in the same price range: 200 euro for a pair.

https://www.kickstarter.com/projects/606955275/manus-machinae-be-part-of-the-machine

To be quite honest, as far as hardware is concerned, they're only missing a few components to be the same as my glove, though they are going after a different type of market, simulating already used input types. My glove could have a script made to be able to do the same thing fairly easy. It is nice to see others trying to make cheaper devices as well though. :D

2 3D mice would probably work.

See 3dconnexion.com

or even 2 Magic Trackpads.

It's no substitute for the gloves but it'll probably be the closest thing until then.

I have a question for the guys developing gloves for VR that implement bend-sensors for the finger motion:
1. What is the life of each bend-sensor like? When I was looking into it, they were notoriously unreliable, some breaking after a few hundred bends, some after a few thousand (but definitely not going to be lasting multiple years of moderate use)
2. What is the accuracy / calibration requirements for each bend sensor like?
3. Do you have the ability to receive input for finger adduction/abduction with bend-sensor controlled gloves? (i.e. can someone make a 'snip snip' motion with their middle and index fingers?

Ultimately, there's no replacement for hands in VR. You can have controllers pretending to be your hands, but until you can move each finger independently and watch those motions take place in VR, you're in a whole different ball-park.

Hello everyone,

Today I received an email that I should check out this topic and to give a reply.

Let me give you a small introduction; I'm Bob Vlemmix and I'm one of the Co-Founders of the Manus Machinae. As stated by Serrarens, we recently started our Kickstarter project in order to get funding's to realize a finalized product.

Since our kick-off we have met a lot of people who work with VR and they all gave us very positive feedback. I hope that by joining other communities like the Oculus forums that we will receive more positive feedback.

To answer mptp's question:
1. The life expectancy of our sensors is 1 million bends. We also know what the weakness of our bend sensor is and have found a way to remove it. Our experience so far is that we have a very durable bend sensor that can handle pressure.
2. The accuracy is high, making it easy to calibrate. Our finalized product we will come with easy to use software so it will be a simple task to calibrate. But as stated in our Kickstarter it will be completely plug and play; our finalized product will have a preset configuration that will be easy to use for everyone.
3. At this moment our product will not have that option. We don't measure this kind of movement because our focus is elsewhere.

On your last point: with the Manus Machinae it is possible to track the position of each finger. So we already have that option covert :)

Everyone feel free to ask me more questions!

Cheers,
Bob

@mptp: The flex sensors they seem to be using (based on the kickstarter) are the same ones I am using for testing purposes. You can find them (and the datasheet) on sparkfun. https://www.sparkfun.com/products/8606

The data sheet claims more then 1 million bends, but if you read through the comments you'll find many have them go bad faster. The issue isn't the actual flex part of the sensor, but where the connectors are will break due to flexing. They didn't put the plastic strip they have for the main flexing part all the way through, so if you don't mount it properly to your project, this section will snap rather quickly. If you solidify it and only let it flex where the actual flex sensor part is, then you get a very long life out of them.

Based on the video, it looks like it's pretty well mounted, so I'd say you'll get a pretty long life out of it. Personally, I'm probably going to try to make my own flex sensors before the kickstarter phase to try and lower cost. These home built sensors don't have the same inherent flaw of needing a rigid plastic, so should have an even longer life.

So worry not! They will probably last you years of life before they will go kaput.

@Bob Good luck with the kickstarter! I am curious to see how well it will go, since it's not full tracking and simulates other inputs.

Yeah I was going to build my own data-glove a while ago so I researched the crap out of those particular bend-sensors (and all other bend sensors I could find). I did read that there was that flaw, and that reinforcing the connector-section drastically improves the life of the sensors. However people were also often griping about how each bend sensor had slightly different voltage-response curves, and how these curves had the tendency to change slightly over time.
In general, I lost confidence in the whole 'conductive ink' technology that those sensors rely on. However, if you guys have solved that issue and are releasing a reliable consumer product using these sensors that don't require constant user-calibration, then I'll consider myself pleasantly surprised. :D

I do feel like only tracking the rotation of each finger about one-axis is a pretty severe limitation, however.

Oh, one other question - how does your system deal with hyperextension of finger joints? Are you using double-sided bend-sensors, or will hyperextensions be ignored by the software?