Microchip accelerates real-time edge AI deployments with NVIDIA Holoscan platform

     To help developers build artificial intelligence (AI) -driven sensor processing systems, Microchip Technology has released the PolarFire FPGA Ethernet sensor Bridge that supports NVIDIA's Holoscan sensor processing platform. The PolarFire FPGA supports multiple protocols and is an integral part of Microchip's platform and is the first solution compatible with MIPI® CSI-2® based sensors and MIPI D-PHY℠ physical layer.

     Subsequent solutions will support sensors with a variety of interfaces, including SLVS-EC™ 2.0, 12G SDI, CoaXPress® 2.0 and JESD204B. The platform enables designers to leverage the power of NVIDIA's Holoscan ecosystem while leveraging PolarFire FPGas' energy-efficient technology, low-latency communications, and multi-protocol sensor support.

     NVIDIA Holoscan helps simplify the development and deployment of edge AI and high performance computing (HPC) applications, enabling real-time insights. It integrates the hardware and software systems needed for low-latency sensor streaming and network connectivity into one platform. The platform includes optimization libraries for data processing, AI model examples to launch AI inference pipeline development, template applications to facilitate rapid prototyping, and core microservices to run streaming, imaging, and other applications.

     The PolarFire FPGA Ethernet sensor bridge Bridges real-time sensor data to NVIDIA Holoscan as well as NVIDIA IGX and NVIDIA Jetson platforms for edge AI and robotics, opening up new edge-to-cloud application deployment possibilities. It also supports AI /ML reasoning to facilitate AI applications in the medical, industrial, and automotive markets.

     Bruce Weyer, Microchip's vice president of FPGA business unit, said: "The Ethernet sensor bridge is based on Microchip's energy-efficient, secure and reliable PolarFire FPGA platform, and by combining our flexible FPGA architecture with NVIDIA's advanced AI platform and multiprotocol support, we are helping developers create innovative real-time solutions. Revolutionize the sensor interface model in a variety of powerful AI-driven edge applications."

     By leveraging low-power Microchip PolarFire FPGA technology, the NVIDIA Holoscan sensor bridge effectively manages high-bandwidth data from different sensors over Ethernet, enabling real-time, high-performance edge AI processing on the NVIDIA AI platform. Energy-efficient designs also benefit small, energy - or cost-sensitive applications.

     PolarFire FPgas provide embedded security and security features that help protect against potential cyber threats and support physical, device, design and data integrity to address sensor application security. The PolarFire FPGA also has single event immunity (SEU) capability, making it highly reliable in radiation-affected environments such as space or high-altitude applications and medical environments. SEU immunity also helps reduce the risk of data corruption and system failure.