CEVA, Inc., the leading licensor of signal processing platforms and artificial intelligence processors for smarter, connected devices, has announced that Artosyn Microelectronics, a leading supplier of drone and robots system-on-chip (SoCs) in China has licensed and deployed the CEVA-XM4 intelligent vision platform to power the computer vision and deep learning workloads of its upcoming AR9X01 artificial intelligence (AI) SoC.
“At Artosyn, we are constantly pushing the boundaries of drone innovation, adding new features and technologies that improve performance, flight time and autonomy,” said Bo Shen, CTO of Artosyn. “The CEVA-XM4 intelligent vision platform provides us with the processing capabilities and the development suite to truly harness the power of AI in our drone SoC designs.”
Artosyn has been a leading silicon vendor for the drone market since the company was founded in 2011, and its controller chips have powered millions of drones to date with the world’s biggest drone manufacturers. The AR9X01 is Artosyn’s most complex chip ever designed, and uses multiple CEVA-XM4 cores to analyse the real-time flight environment and make use of artificial intelligence for a range of tasks including object detection, classification and tracking. Due to the inherent low power nature of running computer vision algorithms and deep learning inference on the CEVA-XM4 platform compared to CPU- or GPU-based alternatives, the AR9X01 allows drone manufacturers to maximise flight time and improve overall drone performance.
“Our CEVA-XM family of intelligent vision processors, and artificial intelligence co-processors continue to lead the industry for vision-based machine learning in embedded devices,” said Ilan Yona, vice president and general manager of CEVA’s Vision Business Unit. “We’re pleased to announce Artosyn, a proven leader in the drone space, as the latest in a long line of customers who have chosen to adopt our CEVA-XM4 platform, enabling them to leverage the capabilities of AI for their AR9X01 SoC.”
CEVA’s latest generation imaging and vision platforms address the extreme processing requirements and low power constraints of the most sophisticated machine learning and machine vision applications used in smartphones, surveillance, AR & VR, drones and self-driving cars. These DSP-based platforms include a hybrid architecture composed of scalar and vector DSP processors and hardware accelerators, coupled with a comprehensive Application Development Kit (ADK) to streamline software deployment. The CEVA ADK includes: CEVA-Link for seamless software level integration with a host processor; a range of widely used and optimized software algorithms; the CEVA Deep Neural Network (CDNN) real-time Neural Network software framework which streamlines machine learning deployment at a fraction of the power consumption of the leading GPU-based systems, and; state-of-the-art development and debugging tools.