Microchip FPGA board sends real time sensor data to edge AI

Microchip Technology has developed an FPGA card that can collect and process data ready for edge AI.

The PolarFire FPGA Ethernet Sensor Bridge collects data from a range of sources, particularly cameras, and sends it to the Nvidia’s Holoscan software running on Jetson GPUs at the edge.

he PolarFire FPGAs enable multi-protocol support, and this first board works with MIPI CSI-2-based camera sensors and the MIPI D-PHY physical layer. Future versions will support sensors with interfaces such as SLVS-EC 2.0, 12G SDI, CoaXPress 2.0 and JESD204B.

Nvidia Holoscan helps streamline the development and deployment of AI and high-performance computing (HPC) applications at the edge for real-time insights. It brings into a single platform the necessary hardware and software systems for low-latency sensor streaming and network connectivity with optimized libraries for data processing, sample AI models for jump-starting AI inference pipeline development, template applications to facilitate rapid prototyping and core microservices to run streaming, imaging and other applications.

The PolarFire bridge card converts the high bandwidth sensor data into an Ethernet stream for the Nvidia IGX and Jetson platforms running the Holoscan software for edge AI and robotics as well as enabling AI/ML inferencing in medical, industrial and automotive applications.

Details are on the PolarFire FPGA Ethernet Sensor Bridge web page.