Edge TPU Coprocessor

• Provides high-performance ML inferencing at the edge with up to 4 TOPS (Trillion Operations Per Second).

• Optimized for TensorFlow Lite models, enabling fast processing with minimal power consumption (approx. 2 TOPS per watt).

Integrated System-on-Module (SoM)

• Features an NXP i.MX 8M SoC with a quad-core Cortex-A53 CPU and a Cortex-M4F microcontroller, ideal for handling both high-level OS tasks and real-time operations.

• Includes 1 GB LPDDR4 RAM and 8 GB eMMC storage for onboard data and program memory.

• Built-in Wi-Fi (802.11ac) and Bluetooth 4.1 for wireless connectivity.

Multiple I/O Interfaces

• USB 3.0 Type-A and Type-C, HDMI 2.0a, 40-pin GPIO, and MIPI CSI/DSI interfaces support cameras, displays, and external modules.

• Gigabit Ethernet and microSD slot for expanded storage and networking.

Audio and Video Features

• Integrated dual PDM microphones, 3.5 mm audio jack, and speaker terminals make it suitable for voice and media applications.

Compact and Expandable Design

• Small form factor with extensive GPIO headers makes it suitable for embedded and production-ready designs.

• The SoM can be removed and used directly in custom hardware, allowing easy transition from prototype to product.

Power Efficiency

• Operates on 5V input via USB Type-C, with efficient power management for mobile or embedded use.

Developer-Friendly Environment

• Runs Mendel Linux, a Debian-based OS tailored for Coral, ensuring compatibility with standard Linux development tools and TensorFlow Lite.