# Directional haptic bracelet

I'm trying to build a bracelet (fastened such that it does not rotate with respect to the wrist) with small vibrating motors at regular intervals. The goal is to pulse these motors to indicate the direction of magnetic north relative to the bracelet, with the motor closest to magnetic north being pulsed. This is inspired by similar projects using belts or anklets instead of bracelets. In my previous attempts, I have had difficulty telling which motor was vibrating, because activating any motor caused the entire bracelet to vibrate.

Are there any existing projects or products with directional haptic input via bracelet? If not, can you recommend any techniques for isolating vibrations between segments?

(I'm using an arduino, a HMC5883L 3 axis magnetometer, and small eccentric rotating mass motors, but I'm open to suggestion)

• Cool project. I've seen similar things in the past. How much force do you think you could get out of piezos mounted normal to the bracelet? Is it cost-effective? – Jason_L_Bens May 20 '14 at 19:03
• Suggested experiment: Attach 4 (or 6) small eccentric mass motors around the wrist, with no bracelet or other attachment between them: Use band-aids to attach each, for instance. Then pulse various motors to see if the wrist can sense which one is vibrating. I believe you will discover that at anything greater than a very mild vibration, the wrist cannot identify specific vibration sources, as the bones of the wrist will vibrate along, fudging up the key purpose. – Anindo Ghosh May 21 '14 at 3:35
• That's a great experiment idea, Andino. Can you repost it as an answer so I can mark this question solved? – pkinsky May 21 '14 at 14:54

Interesting idea. I take it that the idea is that the bracelet needs to be parallel to the ground (i.e., the forearm needs to be vertical).

How about a different usage model: Have the forearm be horizontal (as if the user is pointing somewhere), and pulse the (single) vibrator using Morse Code to indicate the direction the user is pointing?

Obviously, I'm assuming the magnetometer is mounted on the bracelet itself. If it's mounted elsewhere, other usage models may be possible.

• The magnetometer is mounted on the bracelet, but it should work as long as the plane of the bracelet is not perpendicular to magnetic north. As long as some part of the bracelet's circle is closer to magnetic north than the rest of the circle, it should be able to provide useful information. – pkinsky May 20 '14 at 17:36
• In response to your suggested usage model: interesting, are you suggesting using the magnetometer to detect the direction a user is pointing in, and pulsing a single motor to provide information about that direction? (ie: hot, cold, on target, etc) – pkinsky May 20 '14 at 17:39
• Well, I'm actually suggesting that you name the direction the user is pointing in Morse Code; i.e., N, NE, E, SE, etc. It seems to me that it would be more generally useful, and it wouldn't be difficult for users to learn the four letters required. – Dave Tweed May 20 '14 at 20:59

My idea would be to use encapsulated pager motors, and 'mount' them on the inside of the bracelet with vibration-absorbing foam, with the edge towards the wrist.

• Those are the type of motors I'm using. I'll rig something up to test how well foam isolates the vibrations, can you recommend a type of foam to use? – pkinsky May 20 '14 at 17:33
• No, if the project was mine I'd start experimenting. Does it need to be a rigid bracelet, can't it be (at least partially) elastic? – Wouter van Ooijen May 20 '14 at 18:00
• It will probably be mostly elastic, perhaps rigid motor unit 'beads' strung on an elastic ribbon. Looks like it's off to the lab! – pkinsky May 20 '14 at 18:11
• In your previous experiment, was the bracelet all rigid? Otherwise I can't imagine why you couldn't tell which motor was vibrating. But keep in mind that the resolution you can achieve is limited by the distribution and amount of nerve cells. My guess would be that 8 directions is about the maximum. – Wouter van Ooijen May 20 '14 at 18:28
• The last attempt was a non-elastic cloth bracelet with sewn-in patches. Ridgid might be the wrong word for it, but the vibrations weren't well isolated. – pkinsky May 20 '14 at 18:33

In the vein of Dave Tweed's use model, I'd actually try to do it with only two vibe motors - one on either side of the arm. You swing your arm around and the motor closest to north vibrates. If you're pointing the arm due north, then both vibrate. If you point the arm due south, then neither does. It'd be a simpler use model, though it requires user intervention (putting the arm into the proper position) to trigger a 'seek north' cycle.