Qualcomm Technologies, Inc has officially announced its new Qualcomm Robotics RB3 platform, the first fully integrated robotics platform in the company’s history. The platform is based on a customized variant of the Qualcomm Snapdragon 845 processor at the heart of most 2018 flagship smartphones, and boasts a number of relevant features. What makes this platform so ambitious isn’t just how far it could reach with its being fully integrated and easy to use, but rather its goal; AI technology driven by 5G.
The RB3 platform features LTE connectivity, a built-in mapping and spatial awareness solution, hi-fidelity sensor support, and a specialized Qualcomm AI engine on board the Snapdragon 845’s AI coprocessor. Naturally, there’s also Wi-Fi connectivity. The whole package is protected by “vault-like” security. All of this would be quite impressive on its own, but word that the platform will be getting an update with 5G support in the near future is what really drives home that this is a future-forward platform.
The RB3 will be the very first fully integrated platform to feature both a separate AI coprocessor capable of offline and independent machine learning, and 5G connectivity, for both Qualcomm and the industry at large. What this means is that RB3 devices and robots are going to be able to perform a sort of hybrid AI trick; the Snapdragon 845 does a nice job of supporting onboard machine learning and even a low level of simulated neural networking, but that can now be combined with outside smarts, or fed outside data, via a 5G network.
The platform supports many Linux-based operating systems, as well as the specialized Robot OS. Support for Ubuntu is planned for the future. Developers will be pleased to know that it’s compatible with Amazon AWS RoboMaker, along with Qualcomm’s own Neural Processing SDK, Computer Vision Suite, and the Hexagon DSP featured in some other Qualcomm products. What this means is that all of Qualcomm’s robotics-related tools are ready to go out of the box, and support for a far wider range of robotics and AI tool sets is coming.
That means nigh-instant transmission, essentially tucking a vast neural network, a powerful machine learning PC, a massive data set, or whatever else a company may have in the cloud right into their RB3 products. The full integration and ease with which this platform can be deployed suddenly takes on an entirely new meaning.
The RB3, as the name implies, is not Qualcomm’s first foray into robotics. The company has considerable experience in this field, and it shows in the full-suite product being presented here. On top of that, the Snapdragon 845, the chip that the RB3 is based on, is not the company’s first to feature onboard machine learning capabilities. That honor goes to 2017’s Snapdragon 835, found in phones like the Samsung Galaxy S8.
What all of this means for consumers is that we will soon begin to see robotics products that have extreme power efficiency, speedy processing, and most meaningful of all, both onboard and 5G-powered offsite AI capabilities. While AI with the capacity for emotion isn’t a thing just yet, when it does become a thing, this platform will theoretically be able to run it. More realistically and immediately speaking, use cases like smart mapping for robotic vacuums, advanced computer vision, and logic-based decision making for AI in positions that call for it should all be coming soon.