Augmented Reality, Cloud Development, Image Recognition, Mobile Development
The global augmented and virtual reality market is predicted to be worth $215b by 2021, with $108b attributed to mobile AR. However, while mobile platforms have gained significantly in performance, making AR an everyday consumer experience still has some challenges to overcome. Just think whatever happened to Google Glass? These challenges include the need for better recognition and rendering performance, plus quick access to a large database of recognition targets. Mobile devices are not suited to meet these challenges, and cloud solutions have too much latency to make the experience user friendly. Could low-latency, high-bandwidth mobile 5G be the answer?
Although consumers may be aware of mobile AR through gaming, such as Pokémon GO, other potential mobile use cases include education, cultural attractions and medicine, to name but a few examples. However, for mass market adoption of mobile AR to be achieved, improvements in mobile performance and connectivity are needed. While there are a number of cross-platform AR software development kits (SDKs) such as Vuforia, and more recently Apple’s ARKit, mobile connectivity has yet to provide the required level of support to overcome on-device performance limitations of AR.
Current mobile AR solutions rely mostly on the handset to perform AR; performing image recognition on a high-definition camera image and rendering the augmented scene typically at 30 frames per second. However, detecting images and rendering a scene in 30ms is a challenge for mobile chipsets, and SDKs correspondingly limit the number and type of scenes that can be recognised.
For example, the market leading Vuforia SDK, limits on-device recognition to just 100 target images. SDK providers overcome this limitation by off-loading the recognition to a cloud service which can index millions of targets and perform recognition much faster, but at the cost of latency. For example, Vuforia’s cloud service can respond within 1.5s – 45 frames after the camera recorded the target. This latency significantly reduces the effectiveness of AR, perhaps preventing wide-scale adoption.
Performing recognition and rendering only on the device also has one other drawback — it means that it is more difficult to share the AR experience with other users, but this sharing could have huge benefit not only to the consumer experience, but also, for example, for medical applications, where, for example, a shared AR experience could enable junior doctors to share a real-time view of a complex case with off-site consultants.
5G has the potential to offload the compute-intensive elements of AR to be serviced by edge and cloud-based services through a reduced latency network. This potentially means that more complex camera scenes could be assessed which are rendered in the cloud, enabling even entry-level mobile devices to offer rich AR experiences. And because the recognition and rendering is done in real-time on a network device, and not on a mobile, these scenes can be shared more easily.
So, what is 5G and how can it achieve this? In simple terms, 5G is the next generation mobile communications standard, one up from 4G. This conjures up the idea of faster mobile data access speeds, which is true, but does not do it justice.
5G is set to offer mobile data speeds from 1Gb/s to 10Gb/s, network latency of 1ms and greater capacity. To put these into perspective, the fastest UK broadband is 300Mb/s from fixed fibre. 5G could therefore offer from 3 to 30 times faster data speeds from a mobile device compare to what you could get in your own home. Current 4G network latency is 50ms, so 5G could offer 50 times faster network access — as fast as Wi-Fi, if not faster. And all this while offering this experience to more users in the same mobile cell area compared to 4G.
These far higher data speeds and low latency mean that AR processing can be done on mobile edge or cloud-based servers, unlocking the potential for AR to become a far richer, everyday, shared experience.
When might this happen? 5G is currently still being developed, but is now moving into a more stable development phase which is set to allow the first deployments to go public in 2019, with more wide-scale adoption set for 2020. Of course, even with the infrastructure in place, it will take the likes of Apple and Google to kick-start the mobile AR revolution and take AR to the limit.
We have expertise in bringing leading-edge technologies to life. With experience in developing bespoke AR solutions, and working with evolving hardware and software standards, Pervasive Intelligence have the expertise to cut through the complexity to give you a solution you can use in the real world.