Microchip releases compact PolarFire FPGA Ethernet Sensor Bridge for edge AI

Microchip Technology Inc. has announced the release of the second-generation PolarFire FPGA Ethernet Sensor Bridge (Revision 2.0). This production-ready platform is designed to simplify sensor connectivity for compact edge AI systems, such as humanoid robotics, medical AI, and industrial automation.

The new board is 60% smaller than its predecessor and doubles camera support, accommodating up to four cameras. It is compatible with NVIDIA Jetson and utilizes NVIDIA Holoscan Sensor Bridge technology. The platform uses 10Gb Ethernet to connect sensors to AI computing systems, aiming to reduce bill-of-materials complexity and overall system costs by replacing proprietary interfaces with a standardized architecture.

Key technical features include on-board optical latency measurement circuitry, which allows developers to use the NVIDIA Latency Display Analysis Tool to verify end-to-end latency from sensor capture to AI inference. The board also features USB-C power support, an FPGA Mezzanine Card (FMC) connector, and support for various interfaces including MIPI CSI-2, I2C, UART, GPIO, SLVS-EC 2.0, 12G-SDI, HDMI, and DisplayPort.