Forum Discussion
Accelerometers, gyroscopes, and IMU's.. oh my
So I was looking at data gloves today, and the slim off chance of building my own. After my research, I have come to the conclusion that a data glove should implement:
- 2 accelerometers for each finger
- A 6DOF IMU (or accelerometer + gyroscope) on the top of the hand
- Control box near the wrist area
- Maybe integrate flex sensors as redundancy?
The bottom of the glove(s) should include some type of smart material which can provide haptic feedback.
For wireless capability, perhaps something like the ULP 2.4GHz RF Transceiver may do the trick? Depending on latency of course. This also depends on how much power the entire rig would consume.
Of course, the software driving all of this would be more of a hassle (for me), but it looks like there are already classes that are designed for the phone market that can be re-purposed for the VR environment.. so it would just be a matter of referencing those classes. And then there is the small matter of creating drivers, which I have NFC how to do.
- 2 accelerometers for each finger
- A 6DOF IMU (or accelerometer + gyroscope) on the top of the hand
- Control box near the wrist area
- Maybe integrate flex sensors as redundancy?
The bottom of the glove(s) should include some type of smart material which can provide haptic feedback.
For wireless capability, perhaps something like the ULP 2.4GHz RF Transceiver may do the trick? Depending on latency of course. This also depends on how much power the entire rig would consume.
Of course, the software driving all of this would be more of a hassle (for me), but it looks like there are already classes that are designed for the phone market that can be re-purposed for the VR environment.. so it would just be a matter of referencing those classes. And then there is the small matter of creating drivers, which I have NFC how to do.
4 Replies
- Give up on haptics - there's no haptics solution good enough to spend the time and money on for hackers right now. If you're a lab with time and money and manpower, then sure, but otherwise it's just too difficult a problem to solve without an amazing new idea.
Otherwise, I pretty much agree. However, for the purposes of VR, you need the absolute position of the hand, not just the configuration of the skeleton. This means that you need another IMU at each forearm and upper-arm, and one on the torso. You can solve for the torso one's position by knowing its vector from the rift's IMU (which can be done when the Rift is recentered), and then do forward kinematics using each known IMU position on the real skeleton to get the hand position in world space (allowing you to actually place your nice hand-skeletons).
Also, flex sensors aren't worth the expense. They're more expensive than 6DOF IMUs! - These people made some cool gloves, maybe it will help with ideas: http://theglovesproject.com/glove-versions/ Good luck with your experiment!
- These wouldn't have integrated haptics. I suspect "good" haptics won't come into play until we get affordable meta-materials that are low cost.
The tech that sounds promising is a fiber that fluctuates it's elasticity based on electrical input. The material gets stiffer or more flexible depending on how much electrical current is passed through. - You can make your glove using the components you mentioned here. But a good haptics require much efficient and good material made components. You must test it in a lab to get the desired accuracy. But you can also improve your circuit by using testing over these components.
circuit assembly
