LG aims to make VR nausea a thing of the past with a little help from artificial intelligence

Alan Martin May 29, 2018

For some, VR is a transformational experience, allowing them to see things in a way they simply couldn’t without the technology. The 360-degree view of the world tricks the brain, and for those few minutes, they really are somewhere else. For others, however, the experience is a deeply unpleasant one. Any sense of wonder is quickly replaced with a growing feeling of nausea, as motion sickness takes grip. Some (pretty informal) polls have suggested the numbers could be as high as 80% of first-time users feeling a bit queasy, and given first impressions are very important that’s a problem that needs to be fixed.

LG aims to make VR nausea a thing of the past with a little help from artificial intelligence

LG to the rescue. The South Korean company is currently working with Sogang University to develop AI technology which cuts down on motion-to-photon-latency (the delay between your movement being reflected on screen) and motion blur, according to a report in Business Korea.

The solution is quite a clever one, and at its best has reduced photon latency and motion blur to “one fifth or less” of current levels. It tackles the tricky issue of screen resolution which is widely recognised as a likely cause of motion sickness, but a tough one to fix with current technological bottlenecks. In short, because the eyes are so close to the screen for VR to work, resolutions need to be very high to fix the problem – and with one screen per eye, most consumer computer hardware isn’t up to outputting more than 1080p each. So if you try, you’ll likely get more latency and blur and a very full sick bucket. oculus_go_headset_front_close

The researchers’ answer was to develop a deep learning algorithm that can upscale low-resolution images in real time. Presumably, because it’s quite hard to work with subjects who constantly need to throw off the headset, LG and the researchers developed a precision motor which simulates neck muscle movements and an optical system modeled on the human visual cortex to measure photon latency and motion blur of VR headsets.

“This study by LG Display and Sogang University is quite meaningful in that [it] developed a [technology] which accelerates with low power realised through AI without an expensive GPU in a VR device,” said Professor Kang Seok-ju from the department of electronics engineering, who led the study.

It wouldn’t instantly fix VR’s mass adoption problem – the cost of hardware, fiddly setup and lack of compelling software is a bigger issue – but making life better for people who want to enjoy VR but can’t without feeling physically sick would definitely be a good start.