Post by elmeunick9 on Apr 15, 2016 8:24:59 GMT -5
So the idea was something like this:
We have headphones for the ears and oculus for the eyes, we even have GVS, and ways to stimulate almost every other sense, so why it still doesn't feel like full immersion? Because we can't really touch anything! Headtraking devices, and in general movement tracking devices have this problem that inside the game you become a ghost. I could extend on that but basically there are only two solutions to this, we either paralyze the body at nerve level and intercept all signals (pretty hard to do and requires surgery, probably) or we paralyze the body with brute force (aka bondage) and capture the force the subject makes when tries to move.
So this project is about the second.
You can find the project source code and documentation here: gitlab.com/elmeunick9/Siyon/
Basically what I did was use weight sensors to both restrain the body (only the right arm actually) and at the same time capture the movements. So I bought all the material needed for that (some if it also for other projects), and started to build the planed device. I've encountered some problems along my journey:
In the first place the device must be huge, and strong, it must be able to restrain you completely when you're applying full force, and that was the first thing I was unable to do. The principal problem was the price of the material, I needed something cheap and strong, and I also needed to be able to shape it the way I needed. So at the end I choose a piece of wood and a piece of metal that most probably was aluminum (not really sure of it). It proved to not be sufficiently strong to restrain completely the movements during the tests. (It also hurt after extended wearing).
Second problem was the weight sensors, or rather, the chips used to interface with them. Basically to interface with each weight sensor you need a chip (so you need a lot of them), more accurately you need an analog to digital converser (ADC), typically you would mount that chip in a protoboard with some resistances to make it work properly, but doing that for every weight sensor would be too costly. Thankfully some company out there is selling a circuit called HX711 that does exactly that and that it is "relatively" cheap. So I bought those. Unfortunately the circuit limits the frequency of the signals to 10Hz, which is too low for real time. There were an alternative version of the chip that worked at 50Hz but was like 10 time more expensive.
Third problem was the game engine. They are just not designed for this kind of work. I realized that I needed to use ragdolls and interfere with the ragdolls in real time with the input from the sensors, that however wasn't that easy and I dropped/paused the project before I could even achieve that (notice that this was some of the last things I did on the project, so it's more a matter of lack of time rather than complexity itself). By the way, this is also needed for any other project out there involving full immersion VR.
And finally of course the money. The project would've shown better results if I used the right materials, but they are costly. Actually even using cheaper materials, just the indispensable for a whole suit would cost no less than 500€, a money that without excellent results on the tests I wasn't willing to invest.
So that's pretty much it. I hope someone may find this failed project useful, it least to know how others did things before them.
We have headphones for the ears and oculus for the eyes, we even have GVS, and ways to stimulate almost every other sense, so why it still doesn't feel like full immersion? Because we can't really touch anything! Headtraking devices, and in general movement tracking devices have this problem that inside the game you become a ghost. I could extend on that but basically there are only two solutions to this, we either paralyze the body at nerve level and intercept all signals (pretty hard to do and requires surgery, probably) or we paralyze the body with brute force (aka bondage) and capture the force the subject makes when tries to move.
So this project is about the second.
You can find the project source code and documentation here: gitlab.com/elmeunick9/Siyon/
Basically what I did was use weight sensors to both restrain the body (only the right arm actually) and at the same time capture the movements. So I bought all the material needed for that (some if it also for other projects), and started to build the planed device. I've encountered some problems along my journey:
In the first place the device must be huge, and strong, it must be able to restrain you completely when you're applying full force, and that was the first thing I was unable to do. The principal problem was the price of the material, I needed something cheap and strong, and I also needed to be able to shape it the way I needed. So at the end I choose a piece of wood and a piece of metal that most probably was aluminum (not really sure of it). It proved to not be sufficiently strong to restrain completely the movements during the tests. (It also hurt after extended wearing).
Second problem was the weight sensors, or rather, the chips used to interface with them. Basically to interface with each weight sensor you need a chip (so you need a lot of them), more accurately you need an analog to digital converser (ADC), typically you would mount that chip in a protoboard with some resistances to make it work properly, but doing that for every weight sensor would be too costly. Thankfully some company out there is selling a circuit called HX711 that does exactly that and that it is "relatively" cheap. So I bought those. Unfortunately the circuit limits the frequency of the signals to 10Hz, which is too low for real time. There were an alternative version of the chip that worked at 50Hz but was like 10 time more expensive.
Third problem was the game engine. They are just not designed for this kind of work. I realized that I needed to use ragdolls and interfere with the ragdolls in real time with the input from the sensors, that however wasn't that easy and I dropped/paused the project before I could even achieve that (notice that this was some of the last things I did on the project, so it's more a matter of lack of time rather than complexity itself). By the way, this is also needed for any other project out there involving full immersion VR.
And finally of course the money. The project would've shown better results if I used the right materials, but they are costly. Actually even using cheaper materials, just the indispensable for a whole suit would cost no less than 500€, a money that without excellent results on the tests I wasn't willing to invest.
So that's pretty much it. I hope someone may find this failed project useful, it least to know how others did things before them.
