I invented a wearable to reduce motion sickness in VR.

I built it at Florida Gulf Coast University under HCI researcher Dr. Md Baharul Islam, designing a custom PCB, enclosure, and firmware. Scroll for a 3D visualization of the product's design!

fig. 1 · on the headset

It clips to the headset you already wear.

The clip mounts on the VR headset's strap, and the headset's own weight presses a bone-conduction transducer onto the bone behind your ear. No headband, no per-user calibration.

fig. 2 · timing

Bone conduction, rooted in vestibular physiology.

The unit mounts on the back of the ear, on a bone known as the mastoid. The design that I invented ensures a consistent fit on all users. Bone conduction delivered to the mastoid has already been shown to reduce VR motion sickness in clinical trials.

fig. 3 · drawn from the CAD

A novel timing mechanism.

I also invented a novel timing mechanism, where the bone conduction transducer delivers short bursts timed to the peak of linear jerk. According to sensory conflict theory, linear jerk is the timeframe where someone is most likely to suffer from VR motion sickness.

fig. 4 · status

Patent pending. The study is next.

I am registered as a co-inventor, with Dr. Md Baharul Islam, on FGCU's provisional patent, and first author of the IEEE ICIICE 2026 paper. The human-subjects study is being prepared now.

Invention · FGCU HCI research

A motion-sickness fix that rides on the headset.

A wearable that times bone-conducted vibration to the moments VR is most likely to make you sick. I designed and built it end to end: the mount, the circuit board, the firmware, the host software and the test tools.

The open prototype: a curved white housing holding a green circuit board with an ESP32-S3 module, a coin cell and a USB-C port
The working prototype, opened up.
The earhook clipped to a Meta Quest headset strap, resting behind a wearer's ear with its arm over the ear
Clipped to a headset strap.

The problem

Cybersickness keeps a lot of people out of VR. It comes from a conflict between what the eyes see and what the inner ear feels. Most fixes change the picture, by narrowing the field of view or blurring motion, and that costs presence.

I wanted a fix that leaves the picture alone and that people would actually wear. I ran customer research throughout development, and treated comfort and setup as requirements on the same level as effectiveness.

Design decisions

Nothing extra to strap on
Bone-conduction devices usually need a headband and per-user fitting. My clip mounts on the headset itself, angled about 45° so the headset's weight presses the transducer onto the mastoid.
One size, barely there
A flexible TPU hook rests over the ear and fits different ears. The whole assembly weighs 44 g, under 10% of a typical headset.
Works with what people already play
A Python host app reads head motion through OpenVR, so it runs next to any SteamVR app without changing it.
Timed to physiology
The otolith organs sense changes in acceleration. The device fires a brief 500 Hz burst just after each peak of linear jerk, where visual-vestibular conflict (the root of VR motion sickness) is strongest.

How the timing works

head acceleration size of jerk burst
Schematic. Top: head acceleration. Bottom: the size of jerk. Squares: bursts, just after each jerk peak.
Built and measured

From sketch to a 10-unit manufactured run.

A custom PCB (EasyEDA, JLCPCB), ESP32-S3 firmware in C++ with PlatformIO, a Bluetooth Low Energy link to the host app, and a Swift iOS app I wrote just to measure latency.

Cost per unit, 10-unit run
$49.06
Packet loss over 1,000 BLE packets
0%
Motion to vibration onset
158.8 ms
Full assembly
44 g
Earlier prototype with its custom PCB exposed inside the 3D-printed housing
Custom PCB, first build.
Host app window with BLE status, device picker, volume and threshold sliders, and a live trigger log
Host app: thresholds you tune live.
iOS test app showing BLE connection, current sound level, trigger threshold and an event log with timestamps
iOS app for the latency test.
Next · survey design

The paper proved the engineering. The study will test the effect.

I'm preparing the human-subjects study, with a lab reserved under my research assistant privileges. Each participant fills in the Simulator Sickness Questionnaire (SSQ) before and after every session, so every condition is compared on the same validated scale.

No efficacy results yet. The sham condition separates the effect of the timing from the effect of simply feeling a vibration.

  1. AControlVR as usual, device not active
  2. BShamDevice worn, vibration not timed to motion
  3. CActiveBursts timed to jerk peaks