UK AI pioneer demonstrates ‘Full-Stack’ SLAM on Arm-based processors at Embedded Vision Summit
London, UK. 15th September 2020. SLAMcore, a leader in Visual SLAM and Spatial AI systems essential for autonomous robot navigation, today announced that it has joined the Arm AI Partner Programme. As a key player behind the explosion in on-device AI, Arm plays a crucial role in the development of capable, commercially viable robots as solutions to many of today’s challenges. Working as part of the Arm® ecosystem will further accelerate development and use of SLAMcore’s Spatial AI algorithms across this exciting growth market.
Simultaneous Localization and Mapping (SLAM) is both essential for autonomous mobile robots and extremely difficult to get right. It is a computationally challenging workload that not only needs to be well-designed but optimized to the specific on-board processors used by each robot. Arm-based System-on-Chip (SoC) architectures dominate the embedded processor market and are increasingly the first choice for robotics firms due to their high-performance, low-energy use, and compact form factors. The success of Arm designs also provides robot developers with a wide choice of vendors.
Few robot developers have the specialist SLAM expertise, or the time and resources needed, to develop bespoke SLAM solutions for every robot and combination of processors and sensors they require. By supporting and optimizing for Arm architectures, SLAMcore ensures developers can easily integrate SLAMcore algorithms perfectly matched to the embedded processors they select, achieving lower costs and faster time to market.
SLAMcore already supports Arm-based systems including Raspberry Pi 4 and NVIDIA Jetson TX2 processors – both of which will be demonstrated at the Embedded Vision Summit. In addition, Qualcomm recently announced SLAMcore as the development partner on its new RB5 Robotics Platform. Joining the Arm AI Partner Programme will help to extend this range to some of the most powerful and commonly used processors in the industry. Working with Arm will also help SLAMcore to further enhance its Spatial AI SDK to capitalise on the capabilities of Arm’s cutting-edge embedded processors.
“SLAM and Spatial AI are exciting technologies helping to enable the next wave of experiences and autonomy in devices. These are complex workloads which require efficient and trusted processors across many devices and performance needs,” said Dennis Laudick, vice president of marketing, Machine Learning Group, Arm. “The addition of SLAMcore to the Arm AI Partner Program highlights our commitment to enabling a range of developers with the innovation and performance needed to deliver the next-generation of AI-powered experiences.”
“Optimization of SLAM algorithms to the processors and hardware they use is essential to deliver accuracy and speed in a computationally efficient way. Rather than settle for weaker solutions that work with generic hardware, we have focused on tailoring our algorithms to work brilliantly with the best available processors and sensors,” said Paul Brasnett, vice president of strategy at SLAMcore. “Robot developers can be confident that the combination of the SLAMcore Spatial AI algorithms and hardware from our partners will deliver highly accurate, robust and reliable SLAM for their robot designs.”