Qualcomm brings augmented reality closer to reality with new chipset and development environment

Despite the technical challenges and lack of cost-effective platforms for augmented reality (AR), the concept still holds enormous potential for commercial and consumer applications. Some, including Tirias Research, even believe that AR devices could be one of the alternative platforms to the rectangular bricks we call the smartphones that everyone carries today. Like so many other electronic platforms, making AR a reality requires new technologies, new usage models, and new business models that will evolve over a few generations before the entire ecosystem matures sufficiently and that prices finally reach a point where AR makes sense to the masses. . Tirias Research estimates it could take five to ten years, but Qualcomm has just taken a major step towards making that future a reality with the introduction of the company’s first chipset platform designed specifically for performance constraints. , power and size of the AR glasses.

On the second day of the Snapdragon Tech Summit, Qualcomm announced the new Snapdragon AR2 Gen 1 platform. AR2 is important because it is based on realistic requirements for AR glasses, which include high levels of computing power and a high-speed connectivity, and it recognizes the need for a lightweight and practical, if not fashionable, design. The Snapdragon AR2 platform divides the different silicon blocks into three modules spaced around the bezels, allowing for sleeker and more comfortable designs. The three modules include an AR processor, an AR coprocessor, and a Wi-Fi connectivity module. The AR processor and coprocessor work with a host processor in a smartphone, PC, or even a network to provide a complete distributed computing architecture with the heterogeneous processing, sensor, and native AI capabilities of the Snapdragon platform. The connectivity module leverages Wi-Fi 7 and uses simultaneous high-bandwidth multi-link RF channels to provide high-speed/broadband connectivity up to 5.8 Gbps between the glasses and the processing host (as a smartphone).

Today, current AR/VR/MR (Augmented Reality, Virtual Reality, Mixed Reality – also known as XR) solutions typically use a single SoC for processing and connectivity. Although this single-chip solution is optimal for many platforms such as smartphones, it is difficult to fit large SoCs into small form factors such as glasses. Not only does splitting the platform into multiple components allows for better weight distribution and balance for the AR glasses, but the Snapdragon AR2 Gen 1 platform also reduces the wiring requirements of the distributed battery and distributed sensors by 45%, printed circuit board (PCB) area by 40%, CPU power consumption by 50%, and Wi-Fi power consumption by 40% compared to the Snapdragon XR2 platform at a single SoC

Built on the advanced TSMC 4nm process and optimized for AR processing demands, the Snapdragon AR2 Gen 1 platform delivers a 2.5x increase in artificial intelligence (AI) processing over the platform XR2 while operating at less than 1W. The AR2 platform also achieves sub-2ms latency over Wi-Fi and 9ms motion-to-photon latency (the lag between the user making a motion and their display on the screen).

MORE FORBESQualcomm Snapdragon 8 Gen 2 brings more AI to mobileBy Jim McGregor

The new platform will also leverage Qualcomm’s Snapdragon Spaces XR development environment, which is built on OpenXR standards and has been rapidly adopted by a broad ecosystem, including AR and VR system OEMs Honor, Motorola, Nubia , OnePlus, Oppo, RedMagic, Sharp, Vivo, Xiaomi and ZTE. During the Snapdragon Tech Summit, Qualcomm also announced partnerships with Adobe for Universal Scene Description (USD) support, Microsoft for its Mixed Reality Toolkit (MRTK) support, and Niantic for its Lightship toolset. Qualcomm and Niantic also announced AR glasses based on the Qualcomm AR2 Gen 1 platform that serve as development platforms.

One of the key points of the Snapdragon AR2 Gen 1 platform is that it will be coupled with a host processing solution. Eventually it will just be a cloud connection, but in the short term the host will likely be a smartphone, PC, or small, handheld processing solution, often referred to as a puck. Although neither Qualcomm nor any of its partners currently offers a small, handheld device for AR, other than a smartphone, Qualcomm has shown off the Snapdragon SDK, which could serve as a reference design for such a solution.

Tirias Research believes that advances in technology, combined with new uses and business models, will shift mobile personal computing from smartphones to other form factors, including augmented reality, over the next decade. Although there is still a long way to go, Qualcomm’s efforts to develop an AR-specific platform and an open development environment are key steps to enable the industry.