Edge AI

Edge Computing Technologies

Our lab also specializes in deploying AI solutions across a range of edge-computing
platforms, enabling real-time intelligence in resource-constrained environments.

NVIDIA Jetson Nano AGX Orin

NVIDIA Jetson AGX Orin (64GB)

A powerful, high-performance embedded AI supercomputer designed for demanding robotics, autonomous machines, and edge AI applications.

Technical Specifications

CPU

AI Performance

Up to 275 TOPS
GPU

GPU

2048-core Ampere GPU with 64 Tensor Cores
CPU

Memory

32GB or 64GB LPDDR5
CPU

Power

15W–60W Power consumption

Use Cases

  • ● Real-Time Object Detection and MOT

  • ● Surveillance and Security

  • ● Industrial Automation

Hailo-8

Hailo-8 AI Accelerator

The Hailo-8 is a high-performance, power-efficient AI processor designed specifically for accelerating neural network inference on edge devices. In real-world applications, the Hailo-8 demonstrates impressive performance. For instance, it can process 672 frames per second on a 224x224 image resolution using ResNet-50, consuming only 1.67 watts, equating to approximately 2.8 TOPS per watt.

Technical Specifications

GPU

Performance

26 TOPS
GPU

Power

~3W typical power consumption
GPU

Interface

PCIe/MIPI/USB
GPU

AI Framework Support

PyTorch, ONNX, and TensorFlow

Use Cases

  • ● Smart Cameras and Surveillance

  • ● Edge AI for IoT

  • ● Smart Cities

Google Coral USB Accelerator

Google Coral USB Accelerator

This device is designed to bring accelerated Machine Learning (ML) inferencing to existing systems. It's a small, external coprocessor that connects to a host computer via USB. Its primary function is to speed up the execution of TensorFlow Lite models, particularly those used for tasks like image classification and object detection.
- Supported Host Operating Systems: Linux, Windows, macOS.
- Compatible with Raspberry Pi boards.

Technical Specifications

GPU

Processor

Edge TPU coprocessor
GPU

Performance

4 TOPS (INT8)
GPU

Power

2W power consumption
GPU

Connectivity

USB 3.0 Type-C socket

Use Cases

  • ● IoT Edge Devices

  • ● classification, pose estimation, object detection

Google Coral USB Accelerator

Raspberry Pi 5

Raspberry Pi 5 is a compact, powerful single-board computer used for diverse projects, embedded systems, and edge computing tasks like local AI processing, IoT control, and data collection.

Technical Specifications

GPU

Processor

64-bit 2.4GHz Quad-core Arm Cortex-A76
GPU

RAM

4GB, 8GB, or 16GB LPDDR4X memory options.
GPU

Power

Around 2.7W power consumption
GPU

Connectivity

Gigabit Ethernet, dual-band Wi-Fi, and Bluetooth 5.0

Use Cases

  • ● Mini Desktop Computing

  • ● Home Automation & IoT Hubs

  • ● Object Detection and Recognition